wwwelseviercomlocatejep

Journal of Environmental Psychology 24 (2004) 31ndash42

Implicit connections with nature$

P Wesley Schultz Chris Shriver Jennifer J Tabanico Azar M Khazian

Department of Psychology California State University San Marcos CA 92078 USA

Received 10 May 2002 received in revised form 21 February 2003 accepted 27 March 2003

Abstract

Previous research has suggested that attitudes about environmental issues are rooted in the degree to which people believe that

they are part of the natural environment Researchers have distinguished between egoistic concerns which focus on self and

biospheric concerns which focus on all living things In the current paper we argue that the type of concerns a person develops

about environmental issues is associated with the extent to which the individual believes that she is part of nature We argue that

this connection is implicit and exists outside of conscious awareness Two studies are reported on the relationship between implicit

connections with nature and explicit environmental concerns and on the cognitive strategies associated with egoistic and biospheric

attitudes Study 1 reports the results from a modified Implicit Association Test (IAT) designed to measure the degree to which

people associate themselves with nature Results showed a moderate positive relationship between biospheric concerns and implicit

connections with nature and a negative relationship between implicit connections with nature and egoistic concerns Study 2

replicated this basic effect and also examined the testndashretest (immediate 1 week and 4 weeks) reliability of the explicit and implicit

measures Results are interpreted within a broad model of environmental inclusion

r 2003 Elsevier Ltd All rights reserved

1 Introduction

With each passing year we are presented with moreand more evidence that human behavior is adverselyaffecting the natural environment Newspapers maga-zines television news shows and every other form ofmedia routinely convey information on a range of topicsfrom pollution to energy conservation to global warm-ing Survey data indicate that a high percentage of peoplein the United States and many other countries through-out the world know about and express concern for theseissues (Dunlap 1991 Dunlap Gallup amp Gallup 1993Dunlap Van Liere Mertig amp Jones 2000 Ray ampAnderson 2000) In this paper we propose that the typesof environmental attitudes a person develops areassociated with the extent to which an individual believesthat she is part of the natural environment Data fromtwo studies are presented on implicit connections withnature using the Implicit Association Test (IAT)

Before summarizing the previous research on thistopic it is useful to offer some definitions for the termsused in the research literature In the sections thatfollow we will use the terms lsquolsquoconcernsrsquorsquo lsquolsquoattitudesrsquorsquolsquolsquovaluesrsquorsquo and lsquolsquoworldviewrsquorsquo and the distinctions be-tween these concepts may not be readily apparentIndeed these terms are often used interchangeably in theresearch literature We use the term environmental

concern to refer to the affect (ie worry) associatedwith beliefs about environmental problems For exam-ple a person may be concerned about the harmfulconsequences of air pollution for his or her health orconcerned about the long-term consequences of im-proper disposal of hazardous household waste Attitude

refers to a personrsquos evaluative judgment about aparticular entity (Eagly amp Chaiken 1993) Attitudesare typically expressed in degrees of favorability as in lsquolsquoIam in favor of establishing a curbside recyclingprogramrsquorsquo or lsquolsquoI support deposits on beverage contain-ersrsquorsquo The construct of environmental attitudes refers tothe collection of beliefs affect and behavioral intentionsa person holds regarding environmentally relatedactivities or issues

We use the term worldview to refer to a personrsquos beliefabout humanityrsquos relationship with nature A personrsquos

ARTICLE IN PRESS

$Our appreciation goes to Maribel Peraza for her work on Study 1

Portions of this paper were presented at the meeting of the Western

Psychological Association April 2002 Irvine CA

Corresponding author

E-mail address wschultzcsusmedu (PW Schultz)

0272-494403$ - see front matter r 2003 Elsevier Ltd All rights reserved

doi101016S0272-4944(03)00022-7

worldview serves as a cognitive paradigm or beliefsystem Dunlap and Van Liere (1978) and Dunlap et al(2000) have argued that a personrsquos environmentalparadigm constitutes a fundamental part of a personrsquosbelief system it is a lsquolsquoprimitive beliefrsquorsquo and influences awide range of concerns and attitudes At the broadestlevel are values which are conceptualized as importantlife goals or principles (Rokeach 1973 Olson amp Zanna1993) Values function as an organizing system forattitudes and beliefs and they are viewed as determi-nants of attitudes Examples of values would includelsquolsquoequalityrsquorsquo lsquolsquowisdomrsquorsquo lsquolsquoambitionrsquorsquo and lsquolsquofreedomrsquorsquo (cfSchwartz 1994) The term environmental values refers tothose values that are specifically related to nature or thathave been found to correlate with specific environmentalattitudes or concerns

2 Conceptualizing environmental concernmdashwhy do

people care

For the last 30 years researchers have examined theunderlying factors that influence peoplersquos attitudes andconcerns about the environment and environmentalissues A host of demographic variables experiencespersonality dimensions beliefs about control efficacyand responsibility have all received considerable re-search attention (Schultz Oskamp amp Mainieri 1995Hwang Kim amp Jeng 2000 Kals amp Maes 2002) Mostof this research has been based on traditional attitudetheory in which attitudes about environmental issuesare measured on cognitive affective and behavioraldimensions Although fruitful this approach to thestudy of environmental attitudes lacks a clear theoreticalfoundation for why a person develops the types ofattitudes she does

One notably different approach was that taken byDunlap and his colleagues in the New EnvironmentalParadigm (NEP Dunlap amp Van Liere 1978 Dunlapet al 2000) Rather than measuring specific attitudesDunlap and his colleagues developed a series of items toassess lsquolsquoecological worldviewrsquorsquo In examining the coremessage of the environmental movement during the1970s Dunlap and Van Liere (1978) noted that therewas more than just attitudes and concerns aboutenvironmental issuesmdashthe movement was challengingfundamental views about the relationship betweenpeople and nature The NEP scale was designed tomeasure this new emerging worldview which includedbeliefs about the limits to growth humanityrsquos ability toupset the balance of nature and humanityrsquos right to ruleover nature

More recently research has begun to differentiatebetween different types of environmental attitudes andto develop a theoretical model for the relationshipbetween worldview and specific attitudes Stern and his

colleagues have developed a Value-Belief-Norm (VBN)model to explain environmental attitudes and behaviors(Stern Dietz amp Guagnano 1995 Stern Dietz AbelGuagnano amp Kalof 1999 Stern 2000) The VBNtheory postulates a causal chain of variables that leadsto behavior values worldview awareness of adverseconsequences for valued objects perceived ability toreduce the threat and personal norms for proenviron-mental behavior In essence the theory predicts that anindividualrsquos values interact with specific perceptions of agiven situation (perceived adverse consequences to avalued object and perceived ability to do somethingabout it) to yield behavior

Within the VBN theory values provide the source ofconcern for environmental issues and for proenviron-mental behavior Stern and his colleagues (cf Stern ampDietz 1994 Stern et al 1995) have identified three setsof values associated with environmental attitudes whichthey labeled egoistic altruistic and biospheric Egoistic

values are focused on self and self-oriented goals (egsocial power wealth personal success) altruistic valuesfocus on other people (eg family community human-ity friends) biospheric values focus on the well-being ofliving things (eg plants animals trees) Conceptuallyeach of these sets of values can lead to attitudes ofconcern for environmental issues and ultimately tobehavior when activated

There is considerable evidence for the existence ofvalue-based environmental concerns (Thompson ampBarton 1994 Stern et al 1995a b Schultz amp Zelezny1999 Schultz 2000 2001) For example Schultz (2001)asked participants in 14 countries to rate their concernfor the harm caused by environmental problemsto a number of objects Analyses showed a clearstructure corresponding to egoistic (me my lifestylemy health and my future) altruistic (people in mycommunity all people children future generations)and biospheric (plants marine life birds animals)concerns The structure of these concerns was largelyconsistent across the 14 countries sampled See alsoSchultz (2000 2002b)

These findings are consistent with the VBN theorybut we have provided a slightly different interpretationWe have argued that an individualrsquos belief about theextent to which she is part of the natural environmentprovides the foundation for the types of concerns aperson develops and the types of situations that willmotivate them to act At one extreme is the individualwho believes that she is separate from naturemdashthatpeople (and specifically him or her) are exempt from thelaws of nature and superior to plants and animals (seealso Opotow 1994 Opotow amp Weiss 2000) At theother end of the continuum is the individual whobelieves that she is just as much a part of nature as areother animals and (taken to the extreme) that the samerights that apply to humans should apply to plants and

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4232

animals We have referred to this core belief asconnectedness with nature (Schultz 2002a)

The notion of connectedness has a rich history in theenvironmental literature Philosophers have long arguedfor the importance of the belief that lsquolsquoIrsquorsquo am part ofnature (cf Leopold 1949 Bateson amp Bateson 1987Callicott 1999 2002) These philosophical underpin-nings have found their way into sociology (Weigert1997 Dunlap et al 2000) and psychology (Kahn 1999Kidner 2001) Researchers and scholars writing aboutthis topic use terms like lsquolsquoecological identityrsquorsquo lsquolsquorelation-shiprsquorsquo lsquolsquoidentificationrsquorsquo or lsquolsquoonenessrsquorsquo to refer to this corebelief At the individual level connectedness is apsychological variable and yet it has not been carefullystudied or operationalized by psychologists In my ownwork (Schultz 2001 2002a) I have used a single-itemmeasure modified from Aronrsquos (Aron Aron Tudor ampNelson 1991 Aron Aron amp Smollanrsquos 1992) work oninterpersonal relationships consisting of a series ofoverlapping circles labeled lsquolsquoselfrsquorsquo and lsquolsquonaturersquorsquo (see alsoDutcher 2000) We refer to this measure as theInclusion of Nature in Self scale (INS) The resultsfrom studies using this measure (Schultz 2001) havefound connectedness to correlate with biospheric con-cerns (r frac14 031) and with self-reported environmentalbehavior (r frac14 041)

Despite the encouraging results with the INS scalethere is reason to believe that it might not be the bestway to assess connectedness First it is only a singleitem and therefore not subject to tests of internalreliability Second and more importantly it relies onself-report This requires that the participant have anexplicit belief about his or her relationship with natureIn debriefing sessions with participants it became clearthat many of our respondents had not thought aboutthis issue and did not have an existing belief to express(although they were able to create one in response to ourquestion) Dunlap et al (2000) has suggested that apersonrsquos worldview is lsquolsquoprimitiversquorsquo Indeed it seemspossible that a personrsquos sense of connectedness is not aconscious one or at least not a belief that is thoughtabout on a regular basis or readily available for retrieval

Recent developments in social cognition have pro-vided an alternative measurement technique for asses-sing the degree to which people associate themselveswith nature Importantly this relatively new measuredoes not require conscious awareness of the associationGreenwald McGhee and Schwartz (1998) have devel-oped the Implicit Association Test (IAT) to measureautomatic concept-attribute associations The IATmeasures associations through reaction time to pairsof concepts presented on a computer screen For eachtrial participants are instructed to match an item (egDaisy or Moth) with the appropriate concept (egFlower or Insect) as quickly as possible Two conceptsare then combined (Flower and Good Insect and Bad)

The degree of association between the two concepts ismeasured by the difference in response time tocompatible (Flower and Good Insect and Bad) andincompatible (Flower and Bad Insect and good) trialsResearch using the IAT has found it to correlatemoderately with explicit measures (ie self report) ofattitudes (Greenwald amp Banaji 1995 Greenwald et al1998 Farnham Greenwald amp Banaji 1999) Greenwaldand Banaji (1995) suggest that implicit attitudes areautomatic and influence decisions and actions withoutawareness

The IAT has been utilized in various types ofresearch Greenwald and Farnham (2000) describedhow the IAT can be used as an indirect measure of self-esteem and how it can further the understanding of anindividualrsquos social identity The study suggests thatexplicit measures can be biased but that implicitmeasures like the IAT are not affected by such biasesAnother study using the IAT found that racial prejudicewas more accurately measured through implicit mea-sures such as the IAT (Greenwald et al 1998) ratherthan with self report measures Indeed the IAT does notappear to be subject to the same desireability effectsroutinely found in self-report measures and it is difficultto lsquolsquofakersquorsquo a good score on the IAT (see Greenwald et al(2002) for a review of IAT research)

The current set of studies were designed to develop anIAT procedure to measure connectedness with natureOur primary goal was to test hypotheses about therelationship between implicit associations with thenatural environment and explicit environmental con-cerns as well as to examine the stability of implicit andexplicit measures across time Specifically for thepurpose of this paper we predicted that low connected-ness would be associated with egoistic concerns aboutenvironmental issues while high connectedness wouldbe associated with biospheric concerns In the first studyreported below we developed an Implicit AssociationTest to measure the extent to which an individualassociated lsquolsquoselfrsquorsquo with lsquolsquonaturersquorsquo We expected to findthat differences on this IAT task would correlate withexplicit measures of environmental concerns Given theprevious research using the IAT to measure self concept(cf Greenwald amp Farnham 2000) we expected thecorrelations between explicit measures of environmentalattitudes and the IAT to be small (rB020)

3 Study 1

31 Method

311 Participants

Participants in the study were 160 undergraduatesfrom California State University San Marcos Partici-pants were recruited from the Psychology Departmentrsquos

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 33

Human Participant Pool A sample size of 160 wasselected in order to provide 80 power for a correlationcoefficient of 020 (Cohen 1988 p 87)

312 Materials

A questionnaire and computerized test were devel-oped to measure environmental attitudes implicitattitudes toward natural and built environments valuesand demographics Measures included the revisedversion of Dunlaprsquos New Environmental ParadigmScale (NEP) Schultzrsquos Environmental Motives Scaleand a revised version of Greenwaldsrsquo Implicit Associa-tion Test (IAT)

New Environmental Paradigm Scale The revisedversion of the NEP scale was used containing 15 itemsItems in the scale were rated on a 5-point Likert scaleranging from 1 (strongly disagree) to 5 (strongly agree)The instrument is designed to measure the degree towhich people view humans as a part of nature ratherthan as consumers or protectors of nature (Dunlap et al2000)

Environmental Motives Scale This instrument mea-sures concern about environmental problems causedby human behavior (Schultz 2000 2001) Concern forenvironmental issues is divided into three categoriesegoistic altruistic and biospheric Participants rated12 items about which they were concerned from 1 (notimportant) to 7 (supreme importance) Egoisticitems were me my future my prosperity and myhealth altruistic items were future generationshumanity people in the community children bio-spheric items were plants animals marine life andbirds

Implicit Association Test A modified version ofGreenwaldrsquos IAT was used to assess automatic con-ceptndashattribute associations (cf Farnham et al 1999) Acomputerized test was created to measure response time(in ms) needed to classify words associated with naturaland built environments The procedure consisted ofseven blocks of trials Prior to beginning the testparticipants were informed that the task involvedmatching words with categories They were then givenexample items and allowed to ask questions Partici-pants were instructed to go as quickly as possible and tokeep their fingers on the keyboard at all times The IATwas administered using Superlab 174 running on aPowermac 7300 computer with a 1500 color monitor

Following the introduction participants were pre-sented with seven blocks of 10 trials Each trial consistedof a word (selected at random from the relevantcategories shown in Table 1) and a set of categoriesThere were four categories of words used Me Not meNature and Built The five words and their correctcategory classification are shown in Table 1 The Meand Not me words were taken from previous studiesusing the IAT to measure self-concept The Nature and

Built words were selected on the basis of face validityThe blocks were presented as follows

Block 1 NaturendashBuilt Block 2 MendashNot me Block 3 NatureMendashBuiltNot me Block 4 NatureMendashBuiltNot me Block 5 BuiltndashNature Block 6 BuiltMendashNatureNot me Block 7 BuiltMendashNatureNot me

An instruction screen along with examples of Blocks1 3 and 6 are presented in Fig 1 The words werepresented in random order within each of the blocksBlocks 1 and 2 were considered practice as was Block 5Blocks 3 and 4 were considered lsquolsquocompatiblersquorsquo pairingsand Blocks 6 and 7 were considered lsquolsquoincompatiblersquorsquoThe strength of the association between lsquolsquoselfrsquorsquo andlsquolsquonaturersquorsquo is reflected in the ease with which a participantcould complete the compatible trials relative to theincompatible trials That is if a person has a cognitiveassociation between self and nature then it should beeasier for them to complete those trials than when thecategories are incompatible The reverse would be truefor a person with a strong association with builtenvironments

Two versions of the IAT procedure were developed inorder to counter-balance for order of NatureMe(Blocks 3 and 4 above) and BuiltMe (Blocks 6 and 7above) trials In the first rotation (shown above) NatureMe was presented first in the second rotation BuiltMe(shown as Blocks 6 and 7 above) was presented first (asBlocks 3 and 4) Participants were randomly assigned toone of two rotations The average error rate across the70 trials was 413 or 59 One subject had a high errorrate and was dropped yielding a working sample of 159Response latencies were averaged within each of theblocks Prior to averaging the data were screened foroutliers and errors Reaction time for trials that wereanswered incorrectly were removed The data were thenscreened for outliers on an individual basis such thatreaction times that were more than 3 standard devia-tions above or below the participantrsquos mean score wereremoved

ARTICLE IN PRESS

Table 1

Categories and words used to measure connectedness with nature

Categories

Nature Built Me Not me

Animals Building I It

Birds Car Me Other

Plants City Mine Their

Whales Factory Myself Them

Trees Street Self They

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4234

The IAT effect was then produced by subtracting theaverage response latency for the compatible trails(Blocks 3 and 4 above) from the average responselatency for the incompatible trials (Blocks 6 and 7above) Higher scores indicate faster responses forNatureMe trials which we interpret as a greaterassociation between self and nature (ie connectedness)

313 Procedure

Upon arrival participants provided informed con-sent Each participant was then randomly assigned torotation (1 or 2) and order (1 or 2) Rotation refers tothe sequence of IAT blocks described above Orderrefers to whether the participants completed the IAT orthe questionnaire first Once the researcher determinedthe order and rotation they proceeded to administer thetest Participants were then debriefed

32 Results

The first set of analyses examined the psychometricproperties of each of the measures The 12 items fromthe Environmental Motives scale were factor analyzedusing a Principal Axis analysis with an obliminrotation The results revealed a three-factor structure

that explained 66 of the total variance with a clearpattern of factor loadings The results are shown inTable 2 The factors were consistent with those found inprevious research and represent biospheric egostic andaltruistic concerns Correlations between the rotated

ARTICLE IN PRESS

NATUREOR

ME

Instructions

Press d FAST for words described by the left concept name Press k FAST for words described by the right concept name Nothing will happen if you press the wrong key press the other key immediately When you are ready press the spacebar to proceed GO FAST Some mistakes are OK

NATURE BUILT

ANIMALS

NATUREOR

ME

CAR

BUILTOR

ME

MINE

Sample Instructions

Block 3 Sample Compatible Trials The correct response here is ldquoBUILT OR NOT MErdquo

Block 6 Sample Incompatible Trials The correct response here is ldquoBUILT OR MErdquo

Block 1 Sample Practice Trials The correct response here is ldquoNATURErdquo

BUILTOR

NOT ME

BUILTOR

NOT ME

NATUREOR

NOT ME

Fig 1 Sample screenshots from the Implicit Association Test developed to measure connectedness with nature

Table 2

Factor loadings from the Environmental Motives Scale with oblimin

rotation

Item Factor 1

(egoistic)

Factor 2

(biospheric)

Factor 3

(altruistic)

Plants 076

Marine life 088

Birds 088

Animals 082

Me 081

My future 078

My prosperity 065

My health 093

Future generations 072

Humanity 070

People in the

community

060

Children 077

Note Factor loadings smaller than 020 are not shown The results

reported in this table are from Study 1 Factor loadings from Study 2

were nearly identical

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 35

factors were r frac14 002 (egoistic and biospheric) r frac14 028(biospheric and altruistic) and r frac14 051 (egoistic andaltruistic) Scores for each were created by averaging theitems Biospheric (M frac14 554 sd=139) Egoistic(M frac14 549 sd=140) Altruistic (M frac14 620 sd=096) To create scores that controlled for differencesin response tendency a mean correction procedure wasused The average response to all 12 of the environ-mental motives items was computed for each respon-dent This score was subtracted from each of the threescale scores to produce mean corrected biosperic(M frac14 020) egoistic (M frac14 025) and altruistic(M frac14 046) scores

Alpha reliability for the NEP scale was 078 Themean score was 357 (sd=046)

Scores for the IAT-Nature scale were produced byaveraging the 10 trials within each block The IAT effectproduced by subtracting the mean score for the twoblocks of NatureMe items (M frac14 1049 sd=350) fromthe mean score for two blocks of BuiltMe trials(M frac14 1335 sd=418) was 286 (sd=286 N frac14 159)An IAT effect of 0 would indicate no association and apositive IAT effect in this context indicates a generaltendency among our participants to associate self morewith nature than with built environments We alsoanalyzed the IAT data separately by rotation The scoreswere similar for the two rotations with the IAT effect forthe rotation with NatureMe BuiltNot me (ie compa-tible trials) presented first of 264 (N frac14 83) The secondrotation with BuiltMe NatureNot me (ie incompa-tible trials) first had an IAT effect of 309 (N frac14 76)

The second set of analyses examined the relationshipbetween implicit associations with nature (connected-ness) and explicit measures of environmental concernWe hypothesized that participants with a greaterimplicit connection with nature (measured with theIAT) would tend to have higher biospheric concernsand lower egoistic concerns Correlations were calcu-lated using the mean corrected egoistic altruistic andbiospheric scale scores NEP and IAT-Nature scoresAs seen in past research all three environmentalconcerns correlated significantly with the NEP egoistic(reth158THORN frac14 038 po0001) and altruistic (reth158THORN frac14018 po001) negatively and biospheric (reth158THORN frac14049 po001) positively

Biospheric and egoistic concerns had similar relation-ships to the IAT-Nature scores Biospheric concernswere positively correlated with the IAT reth158THORN frac14 021p frac14 0009 while egoistic concerns were negativelycorrelated reth158THORN frac14 016 p frac14 004

4 Study 2

The results from Study 1 showed a small butinterpretable pattern of relationships between implicit

associations with nature and specific types of environ-mental attitudes In order to more fully understandthese relationships a second study was conducted usinga similar procedure Our goals in this second study wereto replicate the findings from Study 1 (particularly therelationship between connectedness biospheric andegoistic concerns) to examine the relationship betweenglobalndashlocal processing styles and environmental atti-tudes (both explicit and implicit) and to examine thestability of IAT-Nature scores across time using a testndashretest procedure

41 Method

411 Participants

Data were obtained from 100 undergraduate students(40 males 60 females mean age=2362 sd=568)Participants were recruited from the Psychology De-partmentrsquos Human Participant Pool at California StateUniversity San Marcos A sample size of 99 was selectedin order to allow for 33 participants in each of threeretest conditions providing 80 power to detect a largeeffect (estimated r frac14 050) The anticipated effect size of050 was used rather than the anticipated effect of 020used in Study 1 because our focus in this study was ontestndashretest reliability which we expected to be large

412 Materials

A questionnaire was used to measure environmentalattitudes and the demographic variables of age andgender The questionnaire consisted of four environ-mental measures a revised version of Dunlap et alrsquosNew Environmental Paradigm (NEP 2000) SchultzrsquosEnvironmental Motives scale (2001) a self-reportedproenvironmental behavior scale (Schultz amp Zelezny1998) and a revised version of Aronrsquos Inclusion ofOther in Self scale (1992 see Schultz 2001) The IAT-nature computer program developed for Study 1 wasused to measure implicit connections with natural andbuilt environments

One of the items included in the questionnaire was ameasure of the perceived relationship between self andnature The item was an adaptation of Aron et al (1992)Inclusion of Other in Self (IOS) scale (see also Aronet al 1991) Participants were asked to select one ofseven different sets of overlapping circles labeled lsquolsquoselfrsquorsquoand lsquolsquonaturersquorsquo The item read lsquolsquoPlease circle the picturethat best describes your relationship with the naturalenvironment How interconnected are you with naturersquorsquoScores ranged from 1 (where the circles touched but didnot overlap) to 7 (where the two circles entirelyoverlapped) We refer to this modified scale as theInclusion of Nature in Self (INS) scale (see Schultz2001)

The questionnaire also contained Witkinrsquos Embedded

Figures Test (EFT Witken 1950 Witken Oltman

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4236

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

worldview serves as a cognitive paradigm or beliefsystem Dunlap and Van Liere (1978) and Dunlap et al(2000) have argued that a personrsquos environmentalparadigm constitutes a fundamental part of a personrsquosbelief system it is a lsquolsquoprimitive beliefrsquorsquo and influences awide range of concerns and attitudes At the broadestlevel are values which are conceptualized as importantlife goals or principles (Rokeach 1973 Olson amp Zanna1993) Values function as an organizing system forattitudes and beliefs and they are viewed as determi-nants of attitudes Examples of values would includelsquolsquoequalityrsquorsquo lsquolsquowisdomrsquorsquo lsquolsquoambitionrsquorsquo and lsquolsquofreedomrsquorsquo (cfSchwartz 1994) The term environmental values refers tothose values that are specifically related to nature or thathave been found to correlate with specific environmentalattitudes or concerns

2 Conceptualizing environmental concernmdashwhy do

people care

For the last 30 years researchers have examined theunderlying factors that influence peoplersquos attitudes andconcerns about the environment and environmentalissues A host of demographic variables experiencespersonality dimensions beliefs about control efficacyand responsibility have all received considerable re-search attention (Schultz Oskamp amp Mainieri 1995Hwang Kim amp Jeng 2000 Kals amp Maes 2002) Mostof this research has been based on traditional attitudetheory in which attitudes about environmental issuesare measured on cognitive affective and behavioraldimensions Although fruitful this approach to thestudy of environmental attitudes lacks a clear theoreticalfoundation for why a person develops the types ofattitudes she does

One notably different approach was that taken byDunlap and his colleagues in the New EnvironmentalParadigm (NEP Dunlap amp Van Liere 1978 Dunlapet al 2000) Rather than measuring specific attitudesDunlap and his colleagues developed a series of items toassess lsquolsquoecological worldviewrsquorsquo In examining the coremessage of the environmental movement during the1970s Dunlap and Van Liere (1978) noted that therewas more than just attitudes and concerns aboutenvironmental issuesmdashthe movement was challengingfundamental views about the relationship betweenpeople and nature The NEP scale was designed tomeasure this new emerging worldview which includedbeliefs about the limits to growth humanityrsquos ability toupset the balance of nature and humanityrsquos right to ruleover nature

More recently research has begun to differentiatebetween different types of environmental attitudes andto develop a theoretical model for the relationshipbetween worldview and specific attitudes Stern and his

colleagues have developed a Value-Belief-Norm (VBN)model to explain environmental attitudes and behaviors(Stern Dietz amp Guagnano 1995 Stern Dietz AbelGuagnano amp Kalof 1999 Stern 2000) The VBNtheory postulates a causal chain of variables that leadsto behavior values worldview awareness of adverseconsequences for valued objects perceived ability toreduce the threat and personal norms for proenviron-mental behavior In essence the theory predicts that anindividualrsquos values interact with specific perceptions of agiven situation (perceived adverse consequences to avalued object and perceived ability to do somethingabout it) to yield behavior

Within the VBN theory values provide the source ofconcern for environmental issues and for proenviron-mental behavior Stern and his colleagues (cf Stern ampDietz 1994 Stern et al 1995) have identified three setsof values associated with environmental attitudes whichthey labeled egoistic altruistic and biospheric Egoistic

values are focused on self and self-oriented goals (egsocial power wealth personal success) altruistic valuesfocus on other people (eg family community human-ity friends) biospheric values focus on the well-being ofliving things (eg plants animals trees) Conceptuallyeach of these sets of values can lead to attitudes ofconcern for environmental issues and ultimately tobehavior when activated

There is considerable evidence for the existence ofvalue-based environmental concerns (Thompson ampBarton 1994 Stern et al 1995a b Schultz amp Zelezny1999 Schultz 2000 2001) For example Schultz (2001)asked participants in 14 countries to rate their concernfor the harm caused by environmental problemsto a number of objects Analyses showed a clearstructure corresponding to egoistic (me my lifestylemy health and my future) altruistic (people in mycommunity all people children future generations)and biospheric (plants marine life birds animals)concerns The structure of these concerns was largelyconsistent across the 14 countries sampled See alsoSchultz (2000 2002b)

These findings are consistent with the VBN theorybut we have provided a slightly different interpretationWe have argued that an individualrsquos belief about theextent to which she is part of the natural environmentprovides the foundation for the types of concerns aperson develops and the types of situations that willmotivate them to act At one extreme is the individualwho believes that she is separate from naturemdashthatpeople (and specifically him or her) are exempt from thelaws of nature and superior to plants and animals (seealso Opotow 1994 Opotow amp Weiss 2000) At theother end of the continuum is the individual whobelieves that she is just as much a part of nature as areother animals and (taken to the extreme) that the samerights that apply to humans should apply to plants and

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4232

animals We have referred to this core belief asconnectedness with nature (Schultz 2002a)

The notion of connectedness has a rich history in theenvironmental literature Philosophers have long arguedfor the importance of the belief that lsquolsquoIrsquorsquo am part ofnature (cf Leopold 1949 Bateson amp Bateson 1987Callicott 1999 2002) These philosophical underpin-nings have found their way into sociology (Weigert1997 Dunlap et al 2000) and psychology (Kahn 1999Kidner 2001) Researchers and scholars writing aboutthis topic use terms like lsquolsquoecological identityrsquorsquo lsquolsquorelation-shiprsquorsquo lsquolsquoidentificationrsquorsquo or lsquolsquoonenessrsquorsquo to refer to this corebelief At the individual level connectedness is apsychological variable and yet it has not been carefullystudied or operationalized by psychologists In my ownwork (Schultz 2001 2002a) I have used a single-itemmeasure modified from Aronrsquos (Aron Aron Tudor ampNelson 1991 Aron Aron amp Smollanrsquos 1992) work oninterpersonal relationships consisting of a series ofoverlapping circles labeled lsquolsquoselfrsquorsquo and lsquolsquonaturersquorsquo (see alsoDutcher 2000) We refer to this measure as theInclusion of Nature in Self scale (INS) The resultsfrom studies using this measure (Schultz 2001) havefound connectedness to correlate with biospheric con-cerns (r frac14 031) and with self-reported environmentalbehavior (r frac14 041)

Despite the encouraging results with the INS scalethere is reason to believe that it might not be the bestway to assess connectedness First it is only a singleitem and therefore not subject to tests of internalreliability Second and more importantly it relies onself-report This requires that the participant have anexplicit belief about his or her relationship with natureIn debriefing sessions with participants it became clearthat many of our respondents had not thought aboutthis issue and did not have an existing belief to express(although they were able to create one in response to ourquestion) Dunlap et al (2000) has suggested that apersonrsquos worldview is lsquolsquoprimitiversquorsquo Indeed it seemspossible that a personrsquos sense of connectedness is not aconscious one or at least not a belief that is thoughtabout on a regular basis or readily available for retrieval

Recent developments in social cognition have pro-vided an alternative measurement technique for asses-sing the degree to which people associate themselveswith nature Importantly this relatively new measuredoes not require conscious awareness of the associationGreenwald McGhee and Schwartz (1998) have devel-oped the Implicit Association Test (IAT) to measureautomatic concept-attribute associations The IATmeasures associations through reaction time to pairsof concepts presented on a computer screen For eachtrial participants are instructed to match an item (egDaisy or Moth) with the appropriate concept (egFlower or Insect) as quickly as possible Two conceptsare then combined (Flower and Good Insect and Bad)

The degree of association between the two concepts ismeasured by the difference in response time tocompatible (Flower and Good Insect and Bad) andincompatible (Flower and Bad Insect and good) trialsResearch using the IAT has found it to correlatemoderately with explicit measures (ie self report) ofattitudes (Greenwald amp Banaji 1995 Greenwald et al1998 Farnham Greenwald amp Banaji 1999) Greenwaldand Banaji (1995) suggest that implicit attitudes areautomatic and influence decisions and actions withoutawareness

The IAT has been utilized in various types ofresearch Greenwald and Farnham (2000) describedhow the IAT can be used as an indirect measure of self-esteem and how it can further the understanding of anindividualrsquos social identity The study suggests thatexplicit measures can be biased but that implicitmeasures like the IAT are not affected by such biasesAnother study using the IAT found that racial prejudicewas more accurately measured through implicit mea-sures such as the IAT (Greenwald et al 1998) ratherthan with self report measures Indeed the IAT does notappear to be subject to the same desireability effectsroutinely found in self-report measures and it is difficultto lsquolsquofakersquorsquo a good score on the IAT (see Greenwald et al(2002) for a review of IAT research)

The current set of studies were designed to develop anIAT procedure to measure connectedness with natureOur primary goal was to test hypotheses about therelationship between implicit associations with thenatural environment and explicit environmental con-cerns as well as to examine the stability of implicit andexplicit measures across time Specifically for thepurpose of this paper we predicted that low connected-ness would be associated with egoistic concerns aboutenvironmental issues while high connectedness wouldbe associated with biospheric concerns In the first studyreported below we developed an Implicit AssociationTest to measure the extent to which an individualassociated lsquolsquoselfrsquorsquo with lsquolsquonaturersquorsquo We expected to findthat differences on this IAT task would correlate withexplicit measures of environmental concerns Given theprevious research using the IAT to measure self concept(cf Greenwald amp Farnham 2000) we expected thecorrelations between explicit measures of environmentalattitudes and the IAT to be small (rB020)

3 Study 1

31 Method

311 Participants

Participants in the study were 160 undergraduatesfrom California State University San Marcos Partici-pants were recruited from the Psychology Departmentrsquos

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 33

Human Participant Pool A sample size of 160 wasselected in order to provide 80 power for a correlationcoefficient of 020 (Cohen 1988 p 87)

312 Materials

A questionnaire and computerized test were devel-oped to measure environmental attitudes implicitattitudes toward natural and built environments valuesand demographics Measures included the revisedversion of Dunlaprsquos New Environmental ParadigmScale (NEP) Schultzrsquos Environmental Motives Scaleand a revised version of Greenwaldsrsquo Implicit Associa-tion Test (IAT)

New Environmental Paradigm Scale The revisedversion of the NEP scale was used containing 15 itemsItems in the scale were rated on a 5-point Likert scaleranging from 1 (strongly disagree) to 5 (strongly agree)The instrument is designed to measure the degree towhich people view humans as a part of nature ratherthan as consumers or protectors of nature (Dunlap et al2000)

Environmental Motives Scale This instrument mea-sures concern about environmental problems causedby human behavior (Schultz 2000 2001) Concern forenvironmental issues is divided into three categoriesegoistic altruistic and biospheric Participants rated12 items about which they were concerned from 1 (notimportant) to 7 (supreme importance) Egoisticitems were me my future my prosperity and myhealth altruistic items were future generationshumanity people in the community children bio-spheric items were plants animals marine life andbirds

Implicit Association Test A modified version ofGreenwaldrsquos IAT was used to assess automatic con-ceptndashattribute associations (cf Farnham et al 1999) Acomputerized test was created to measure response time(in ms) needed to classify words associated with naturaland built environments The procedure consisted ofseven blocks of trials Prior to beginning the testparticipants were informed that the task involvedmatching words with categories They were then givenexample items and allowed to ask questions Partici-pants were instructed to go as quickly as possible and tokeep their fingers on the keyboard at all times The IATwas administered using Superlab 174 running on aPowermac 7300 computer with a 1500 color monitor

Following the introduction participants were pre-sented with seven blocks of 10 trials Each trial consistedof a word (selected at random from the relevantcategories shown in Table 1) and a set of categoriesThere were four categories of words used Me Not meNature and Built The five words and their correctcategory classification are shown in Table 1 The Meand Not me words were taken from previous studiesusing the IAT to measure self-concept The Nature and

Built words were selected on the basis of face validityThe blocks were presented as follows

Block 1 NaturendashBuilt Block 2 MendashNot me Block 3 NatureMendashBuiltNot me Block 4 NatureMendashBuiltNot me Block 5 BuiltndashNature Block 6 BuiltMendashNatureNot me Block 7 BuiltMendashNatureNot me

An instruction screen along with examples of Blocks1 3 and 6 are presented in Fig 1 The words werepresented in random order within each of the blocksBlocks 1 and 2 were considered practice as was Block 5Blocks 3 and 4 were considered lsquolsquocompatiblersquorsquo pairingsand Blocks 6 and 7 were considered lsquolsquoincompatiblersquorsquoThe strength of the association between lsquolsquoselfrsquorsquo andlsquolsquonaturersquorsquo is reflected in the ease with which a participantcould complete the compatible trials relative to theincompatible trials That is if a person has a cognitiveassociation between self and nature then it should beeasier for them to complete those trials than when thecategories are incompatible The reverse would be truefor a person with a strong association with builtenvironments

Two versions of the IAT procedure were developed inorder to counter-balance for order of NatureMe(Blocks 3 and 4 above) and BuiltMe (Blocks 6 and 7above) trials In the first rotation (shown above) NatureMe was presented first in the second rotation BuiltMe(shown as Blocks 6 and 7 above) was presented first (asBlocks 3 and 4) Participants were randomly assigned toone of two rotations The average error rate across the70 trials was 413 or 59 One subject had a high errorrate and was dropped yielding a working sample of 159Response latencies were averaged within each of theblocks Prior to averaging the data were screened foroutliers and errors Reaction time for trials that wereanswered incorrectly were removed The data were thenscreened for outliers on an individual basis such thatreaction times that were more than 3 standard devia-tions above or below the participantrsquos mean score wereremoved

ARTICLE IN PRESS

Table 1

Categories and words used to measure connectedness with nature

Categories

Nature Built Me Not me

Animals Building I It

Birds Car Me Other

Plants City Mine Their

Whales Factory Myself Them

Trees Street Self They

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4234

The IAT effect was then produced by subtracting theaverage response latency for the compatible trails(Blocks 3 and 4 above) from the average responselatency for the incompatible trials (Blocks 6 and 7above) Higher scores indicate faster responses forNatureMe trials which we interpret as a greaterassociation between self and nature (ie connectedness)

313 Procedure

Upon arrival participants provided informed con-sent Each participant was then randomly assigned torotation (1 or 2) and order (1 or 2) Rotation refers tothe sequence of IAT blocks described above Orderrefers to whether the participants completed the IAT orthe questionnaire first Once the researcher determinedthe order and rotation they proceeded to administer thetest Participants were then debriefed

32 Results

The first set of analyses examined the psychometricproperties of each of the measures The 12 items fromthe Environmental Motives scale were factor analyzedusing a Principal Axis analysis with an obliminrotation The results revealed a three-factor structure

that explained 66 of the total variance with a clearpattern of factor loadings The results are shown inTable 2 The factors were consistent with those found inprevious research and represent biospheric egostic andaltruistic concerns Correlations between the rotated

ARTICLE IN PRESS

NATUREOR

ME

Instructions

Press d FAST for words described by the left concept name Press k FAST for words described by the right concept name Nothing will happen if you press the wrong key press the other key immediately When you are ready press the spacebar to proceed GO FAST Some mistakes are OK

NATURE BUILT

ANIMALS

NATUREOR

ME

CAR

BUILTOR

ME

MINE

Sample Instructions

Block 3 Sample Compatible Trials The correct response here is ldquoBUILT OR NOT MErdquo

Block 6 Sample Incompatible Trials The correct response here is ldquoBUILT OR MErdquo

Block 1 Sample Practice Trials The correct response here is ldquoNATURErdquo

BUILTOR

NOT ME

BUILTOR

NOT ME

NATUREOR

NOT ME

Fig 1 Sample screenshots from the Implicit Association Test developed to measure connectedness with nature

Table 2

Factor loadings from the Environmental Motives Scale with oblimin

rotation

Item Factor 1

(egoistic)

Factor 2

(biospheric)

Factor 3

(altruistic)

Plants 076

Marine life 088

Birds 088

Animals 082

Me 081

My future 078

My prosperity 065

My health 093

Future generations 072

Humanity 070

People in the

community

060

Children 077

Note Factor loadings smaller than 020 are not shown The results

reported in this table are from Study 1 Factor loadings from Study 2

were nearly identical

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 35

factors were r frac14 002 (egoistic and biospheric) r frac14 028(biospheric and altruistic) and r frac14 051 (egoistic andaltruistic) Scores for each were created by averaging theitems Biospheric (M frac14 554 sd=139) Egoistic(M frac14 549 sd=140) Altruistic (M frac14 620 sd=096) To create scores that controlled for differencesin response tendency a mean correction procedure wasused The average response to all 12 of the environ-mental motives items was computed for each respon-dent This score was subtracted from each of the threescale scores to produce mean corrected biosperic(M frac14 020) egoistic (M frac14 025) and altruistic(M frac14 046) scores

Alpha reliability for the NEP scale was 078 Themean score was 357 (sd=046)

Scores for the IAT-Nature scale were produced byaveraging the 10 trials within each block The IAT effectproduced by subtracting the mean score for the twoblocks of NatureMe items (M frac14 1049 sd=350) fromthe mean score for two blocks of BuiltMe trials(M frac14 1335 sd=418) was 286 (sd=286 N frac14 159)An IAT effect of 0 would indicate no association and apositive IAT effect in this context indicates a generaltendency among our participants to associate self morewith nature than with built environments We alsoanalyzed the IAT data separately by rotation The scoreswere similar for the two rotations with the IAT effect forthe rotation with NatureMe BuiltNot me (ie compa-tible trials) presented first of 264 (N frac14 83) The secondrotation with BuiltMe NatureNot me (ie incompa-tible trials) first had an IAT effect of 309 (N frac14 76)

The second set of analyses examined the relationshipbetween implicit associations with nature (connected-ness) and explicit measures of environmental concernWe hypothesized that participants with a greaterimplicit connection with nature (measured with theIAT) would tend to have higher biospheric concernsand lower egoistic concerns Correlations were calcu-lated using the mean corrected egoistic altruistic andbiospheric scale scores NEP and IAT-Nature scoresAs seen in past research all three environmentalconcerns correlated significantly with the NEP egoistic(reth158THORN frac14 038 po0001) and altruistic (reth158THORN frac14018 po001) negatively and biospheric (reth158THORN frac14049 po001) positively

Biospheric and egoistic concerns had similar relation-ships to the IAT-Nature scores Biospheric concernswere positively correlated with the IAT reth158THORN frac14 021p frac14 0009 while egoistic concerns were negativelycorrelated reth158THORN frac14 016 p frac14 004

4 Study 2

The results from Study 1 showed a small butinterpretable pattern of relationships between implicit

associations with nature and specific types of environ-mental attitudes In order to more fully understandthese relationships a second study was conducted usinga similar procedure Our goals in this second study wereto replicate the findings from Study 1 (particularly therelationship between connectedness biospheric andegoistic concerns) to examine the relationship betweenglobalndashlocal processing styles and environmental atti-tudes (both explicit and implicit) and to examine thestability of IAT-Nature scores across time using a testndashretest procedure

41 Method

411 Participants

Data were obtained from 100 undergraduate students(40 males 60 females mean age=2362 sd=568)Participants were recruited from the Psychology De-partmentrsquos Human Participant Pool at California StateUniversity San Marcos A sample size of 99 was selectedin order to allow for 33 participants in each of threeretest conditions providing 80 power to detect a largeeffect (estimated r frac14 050) The anticipated effect size of050 was used rather than the anticipated effect of 020used in Study 1 because our focus in this study was ontestndashretest reliability which we expected to be large

412 Materials

A questionnaire was used to measure environmentalattitudes and the demographic variables of age andgender The questionnaire consisted of four environ-mental measures a revised version of Dunlap et alrsquosNew Environmental Paradigm (NEP 2000) SchultzrsquosEnvironmental Motives scale (2001) a self-reportedproenvironmental behavior scale (Schultz amp Zelezny1998) and a revised version of Aronrsquos Inclusion ofOther in Self scale (1992 see Schultz 2001) The IAT-nature computer program developed for Study 1 wasused to measure implicit connections with natural andbuilt environments

One of the items included in the questionnaire was ameasure of the perceived relationship between self andnature The item was an adaptation of Aron et al (1992)Inclusion of Other in Self (IOS) scale (see also Aronet al 1991) Participants were asked to select one ofseven different sets of overlapping circles labeled lsquolsquoselfrsquorsquoand lsquolsquonaturersquorsquo The item read lsquolsquoPlease circle the picturethat best describes your relationship with the naturalenvironment How interconnected are you with naturersquorsquoScores ranged from 1 (where the circles touched but didnot overlap) to 7 (where the two circles entirelyoverlapped) We refer to this modified scale as theInclusion of Nature in Self (INS) scale (see Schultz2001)

The questionnaire also contained Witkinrsquos Embedded

Figures Test (EFT Witken 1950 Witken Oltman

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4236

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

animals We have referred to this core belief asconnectedness with nature (Schultz 2002a)

The notion of connectedness has a rich history in theenvironmental literature Philosophers have long arguedfor the importance of the belief that lsquolsquoIrsquorsquo am part ofnature (cf Leopold 1949 Bateson amp Bateson 1987Callicott 1999 2002) These philosophical underpin-nings have found their way into sociology (Weigert1997 Dunlap et al 2000) and psychology (Kahn 1999Kidner 2001) Researchers and scholars writing aboutthis topic use terms like lsquolsquoecological identityrsquorsquo lsquolsquorelation-shiprsquorsquo lsquolsquoidentificationrsquorsquo or lsquolsquoonenessrsquorsquo to refer to this corebelief At the individual level connectedness is apsychological variable and yet it has not been carefullystudied or operationalized by psychologists In my ownwork (Schultz 2001 2002a) I have used a single-itemmeasure modified from Aronrsquos (Aron Aron Tudor ampNelson 1991 Aron Aron amp Smollanrsquos 1992) work oninterpersonal relationships consisting of a series ofoverlapping circles labeled lsquolsquoselfrsquorsquo and lsquolsquonaturersquorsquo (see alsoDutcher 2000) We refer to this measure as theInclusion of Nature in Self scale (INS) The resultsfrom studies using this measure (Schultz 2001) havefound connectedness to correlate with biospheric con-cerns (r frac14 031) and with self-reported environmentalbehavior (r frac14 041)

Despite the encouraging results with the INS scalethere is reason to believe that it might not be the bestway to assess connectedness First it is only a singleitem and therefore not subject to tests of internalreliability Second and more importantly it relies onself-report This requires that the participant have anexplicit belief about his or her relationship with natureIn debriefing sessions with participants it became clearthat many of our respondents had not thought aboutthis issue and did not have an existing belief to express(although they were able to create one in response to ourquestion) Dunlap et al (2000) has suggested that apersonrsquos worldview is lsquolsquoprimitiversquorsquo Indeed it seemspossible that a personrsquos sense of connectedness is not aconscious one or at least not a belief that is thoughtabout on a regular basis or readily available for retrieval

Recent developments in social cognition have pro-vided an alternative measurement technique for asses-sing the degree to which people associate themselveswith nature Importantly this relatively new measuredoes not require conscious awareness of the associationGreenwald McGhee and Schwartz (1998) have devel-oped the Implicit Association Test (IAT) to measureautomatic concept-attribute associations The IATmeasures associations through reaction time to pairsof concepts presented on a computer screen For eachtrial participants are instructed to match an item (egDaisy or Moth) with the appropriate concept (egFlower or Insect) as quickly as possible Two conceptsare then combined (Flower and Good Insect and Bad)

The degree of association between the two concepts ismeasured by the difference in response time tocompatible (Flower and Good Insect and Bad) andincompatible (Flower and Bad Insect and good) trialsResearch using the IAT has found it to correlatemoderately with explicit measures (ie self report) ofattitudes (Greenwald amp Banaji 1995 Greenwald et al1998 Farnham Greenwald amp Banaji 1999) Greenwaldand Banaji (1995) suggest that implicit attitudes areautomatic and influence decisions and actions withoutawareness

The IAT has been utilized in various types ofresearch Greenwald and Farnham (2000) describedhow the IAT can be used as an indirect measure of self-esteem and how it can further the understanding of anindividualrsquos social identity The study suggests thatexplicit measures can be biased but that implicitmeasures like the IAT are not affected by such biasesAnother study using the IAT found that racial prejudicewas more accurately measured through implicit mea-sures such as the IAT (Greenwald et al 1998) ratherthan with self report measures Indeed the IAT does notappear to be subject to the same desireability effectsroutinely found in self-report measures and it is difficultto lsquolsquofakersquorsquo a good score on the IAT (see Greenwald et al(2002) for a review of IAT research)

The current set of studies were designed to develop anIAT procedure to measure connectedness with natureOur primary goal was to test hypotheses about therelationship between implicit associations with thenatural environment and explicit environmental con-cerns as well as to examine the stability of implicit andexplicit measures across time Specifically for thepurpose of this paper we predicted that low connected-ness would be associated with egoistic concerns aboutenvironmental issues while high connectedness wouldbe associated with biospheric concerns In the first studyreported below we developed an Implicit AssociationTest to measure the extent to which an individualassociated lsquolsquoselfrsquorsquo with lsquolsquonaturersquorsquo We expected to findthat differences on this IAT task would correlate withexplicit measures of environmental concerns Given theprevious research using the IAT to measure self concept(cf Greenwald amp Farnham 2000) we expected thecorrelations between explicit measures of environmentalattitudes and the IAT to be small (rB020)

3 Study 1

31 Method

311 Participants

Participants in the study were 160 undergraduatesfrom California State University San Marcos Partici-pants were recruited from the Psychology Departmentrsquos

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 33

Human Participant Pool A sample size of 160 wasselected in order to provide 80 power for a correlationcoefficient of 020 (Cohen 1988 p 87)

312 Materials

A questionnaire and computerized test were devel-oped to measure environmental attitudes implicitattitudes toward natural and built environments valuesand demographics Measures included the revisedversion of Dunlaprsquos New Environmental ParadigmScale (NEP) Schultzrsquos Environmental Motives Scaleand a revised version of Greenwaldsrsquo Implicit Associa-tion Test (IAT)

New Environmental Paradigm Scale The revisedversion of the NEP scale was used containing 15 itemsItems in the scale were rated on a 5-point Likert scaleranging from 1 (strongly disagree) to 5 (strongly agree)The instrument is designed to measure the degree towhich people view humans as a part of nature ratherthan as consumers or protectors of nature (Dunlap et al2000)

Environmental Motives Scale This instrument mea-sures concern about environmental problems causedby human behavior (Schultz 2000 2001) Concern forenvironmental issues is divided into three categoriesegoistic altruistic and biospheric Participants rated12 items about which they were concerned from 1 (notimportant) to 7 (supreme importance) Egoisticitems were me my future my prosperity and myhealth altruistic items were future generationshumanity people in the community children bio-spheric items were plants animals marine life andbirds

Implicit Association Test A modified version ofGreenwaldrsquos IAT was used to assess automatic con-ceptndashattribute associations (cf Farnham et al 1999) Acomputerized test was created to measure response time(in ms) needed to classify words associated with naturaland built environments The procedure consisted ofseven blocks of trials Prior to beginning the testparticipants were informed that the task involvedmatching words with categories They were then givenexample items and allowed to ask questions Partici-pants were instructed to go as quickly as possible and tokeep their fingers on the keyboard at all times The IATwas administered using Superlab 174 running on aPowermac 7300 computer with a 1500 color monitor

Following the introduction participants were pre-sented with seven blocks of 10 trials Each trial consistedof a word (selected at random from the relevantcategories shown in Table 1) and a set of categoriesThere were four categories of words used Me Not meNature and Built The five words and their correctcategory classification are shown in Table 1 The Meand Not me words were taken from previous studiesusing the IAT to measure self-concept The Nature and

Built words were selected on the basis of face validityThe blocks were presented as follows

Block 1 NaturendashBuilt Block 2 MendashNot me Block 3 NatureMendashBuiltNot me Block 4 NatureMendashBuiltNot me Block 5 BuiltndashNature Block 6 BuiltMendashNatureNot me Block 7 BuiltMendashNatureNot me

An instruction screen along with examples of Blocks1 3 and 6 are presented in Fig 1 The words werepresented in random order within each of the blocksBlocks 1 and 2 were considered practice as was Block 5Blocks 3 and 4 were considered lsquolsquocompatiblersquorsquo pairingsand Blocks 6 and 7 were considered lsquolsquoincompatiblersquorsquoThe strength of the association between lsquolsquoselfrsquorsquo andlsquolsquonaturersquorsquo is reflected in the ease with which a participantcould complete the compatible trials relative to theincompatible trials That is if a person has a cognitiveassociation between self and nature then it should beeasier for them to complete those trials than when thecategories are incompatible The reverse would be truefor a person with a strong association with builtenvironments

Two versions of the IAT procedure were developed inorder to counter-balance for order of NatureMe(Blocks 3 and 4 above) and BuiltMe (Blocks 6 and 7above) trials In the first rotation (shown above) NatureMe was presented first in the second rotation BuiltMe(shown as Blocks 6 and 7 above) was presented first (asBlocks 3 and 4) Participants were randomly assigned toone of two rotations The average error rate across the70 trials was 413 or 59 One subject had a high errorrate and was dropped yielding a working sample of 159Response latencies were averaged within each of theblocks Prior to averaging the data were screened foroutliers and errors Reaction time for trials that wereanswered incorrectly were removed The data were thenscreened for outliers on an individual basis such thatreaction times that were more than 3 standard devia-tions above or below the participantrsquos mean score wereremoved

ARTICLE IN PRESS

Table 1

Categories and words used to measure connectedness with nature

Categories

Nature Built Me Not me

Animals Building I It

Birds Car Me Other

Plants City Mine Their

Whales Factory Myself Them

Trees Street Self They

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4234

The IAT effect was then produced by subtracting theaverage response latency for the compatible trails(Blocks 3 and 4 above) from the average responselatency for the incompatible trials (Blocks 6 and 7above) Higher scores indicate faster responses forNatureMe trials which we interpret as a greaterassociation between self and nature (ie connectedness)

313 Procedure

Upon arrival participants provided informed con-sent Each participant was then randomly assigned torotation (1 or 2) and order (1 or 2) Rotation refers tothe sequence of IAT blocks described above Orderrefers to whether the participants completed the IAT orthe questionnaire first Once the researcher determinedthe order and rotation they proceeded to administer thetest Participants were then debriefed

32 Results

The first set of analyses examined the psychometricproperties of each of the measures The 12 items fromthe Environmental Motives scale were factor analyzedusing a Principal Axis analysis with an obliminrotation The results revealed a three-factor structure

that explained 66 of the total variance with a clearpattern of factor loadings The results are shown inTable 2 The factors were consistent with those found inprevious research and represent biospheric egostic andaltruistic concerns Correlations between the rotated

ARTICLE IN PRESS

NATUREOR

ME

Instructions

Press d FAST for words described by the left concept name Press k FAST for words described by the right concept name Nothing will happen if you press the wrong key press the other key immediately When you are ready press the spacebar to proceed GO FAST Some mistakes are OK

NATURE BUILT

ANIMALS

NATUREOR

ME

CAR

BUILTOR

ME

MINE

Sample Instructions

Block 3 Sample Compatible Trials The correct response here is ldquoBUILT OR NOT MErdquo

Block 6 Sample Incompatible Trials The correct response here is ldquoBUILT OR MErdquo

Block 1 Sample Practice Trials The correct response here is ldquoNATURErdquo

BUILTOR

NOT ME

BUILTOR

NOT ME

NATUREOR

NOT ME

Fig 1 Sample screenshots from the Implicit Association Test developed to measure connectedness with nature

Table 2

Factor loadings from the Environmental Motives Scale with oblimin

rotation

Item Factor 1

(egoistic)

Factor 2

(biospheric)

Factor 3

(altruistic)

Plants 076

Marine life 088

Birds 088

Animals 082

Me 081

My future 078

My prosperity 065

My health 093

Future generations 072

Humanity 070

People in the

community

060

Children 077

Note Factor loadings smaller than 020 are not shown The results

reported in this table are from Study 1 Factor loadings from Study 2

were nearly identical

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 35

factors were r frac14 002 (egoistic and biospheric) r frac14 028(biospheric and altruistic) and r frac14 051 (egoistic andaltruistic) Scores for each were created by averaging theitems Biospheric (M frac14 554 sd=139) Egoistic(M frac14 549 sd=140) Altruistic (M frac14 620 sd=096) To create scores that controlled for differencesin response tendency a mean correction procedure wasused The average response to all 12 of the environ-mental motives items was computed for each respon-dent This score was subtracted from each of the threescale scores to produce mean corrected biosperic(M frac14 020) egoistic (M frac14 025) and altruistic(M frac14 046) scores

Alpha reliability for the NEP scale was 078 Themean score was 357 (sd=046)

Scores for the IAT-Nature scale were produced byaveraging the 10 trials within each block The IAT effectproduced by subtracting the mean score for the twoblocks of NatureMe items (M frac14 1049 sd=350) fromthe mean score for two blocks of BuiltMe trials(M frac14 1335 sd=418) was 286 (sd=286 N frac14 159)An IAT effect of 0 would indicate no association and apositive IAT effect in this context indicates a generaltendency among our participants to associate self morewith nature than with built environments We alsoanalyzed the IAT data separately by rotation The scoreswere similar for the two rotations with the IAT effect forthe rotation with NatureMe BuiltNot me (ie compa-tible trials) presented first of 264 (N frac14 83) The secondrotation with BuiltMe NatureNot me (ie incompa-tible trials) first had an IAT effect of 309 (N frac14 76)

The second set of analyses examined the relationshipbetween implicit associations with nature (connected-ness) and explicit measures of environmental concernWe hypothesized that participants with a greaterimplicit connection with nature (measured with theIAT) would tend to have higher biospheric concernsand lower egoistic concerns Correlations were calcu-lated using the mean corrected egoistic altruistic andbiospheric scale scores NEP and IAT-Nature scoresAs seen in past research all three environmentalconcerns correlated significantly with the NEP egoistic(reth158THORN frac14 038 po0001) and altruistic (reth158THORN frac14018 po001) negatively and biospheric (reth158THORN frac14049 po001) positively

Biospheric and egoistic concerns had similar relation-ships to the IAT-Nature scores Biospheric concernswere positively correlated with the IAT reth158THORN frac14 021p frac14 0009 while egoistic concerns were negativelycorrelated reth158THORN frac14 016 p frac14 004

4 Study 2

The results from Study 1 showed a small butinterpretable pattern of relationships between implicit

associations with nature and specific types of environ-mental attitudes In order to more fully understandthese relationships a second study was conducted usinga similar procedure Our goals in this second study wereto replicate the findings from Study 1 (particularly therelationship between connectedness biospheric andegoistic concerns) to examine the relationship betweenglobalndashlocal processing styles and environmental atti-tudes (both explicit and implicit) and to examine thestability of IAT-Nature scores across time using a testndashretest procedure

41 Method

411 Participants

Data were obtained from 100 undergraduate students(40 males 60 females mean age=2362 sd=568)Participants were recruited from the Psychology De-partmentrsquos Human Participant Pool at California StateUniversity San Marcos A sample size of 99 was selectedin order to allow for 33 participants in each of threeretest conditions providing 80 power to detect a largeeffect (estimated r frac14 050) The anticipated effect size of050 was used rather than the anticipated effect of 020used in Study 1 because our focus in this study was ontestndashretest reliability which we expected to be large

412 Materials

A questionnaire was used to measure environmentalattitudes and the demographic variables of age andgender The questionnaire consisted of four environ-mental measures a revised version of Dunlap et alrsquosNew Environmental Paradigm (NEP 2000) SchultzrsquosEnvironmental Motives scale (2001) a self-reportedproenvironmental behavior scale (Schultz amp Zelezny1998) and a revised version of Aronrsquos Inclusion ofOther in Self scale (1992 see Schultz 2001) The IAT-nature computer program developed for Study 1 wasused to measure implicit connections with natural andbuilt environments

One of the items included in the questionnaire was ameasure of the perceived relationship between self andnature The item was an adaptation of Aron et al (1992)Inclusion of Other in Self (IOS) scale (see also Aronet al 1991) Participants were asked to select one ofseven different sets of overlapping circles labeled lsquolsquoselfrsquorsquoand lsquolsquonaturersquorsquo The item read lsquolsquoPlease circle the picturethat best describes your relationship with the naturalenvironment How interconnected are you with naturersquorsquoScores ranged from 1 (where the circles touched but didnot overlap) to 7 (where the two circles entirelyoverlapped) We refer to this modified scale as theInclusion of Nature in Self (INS) scale (see Schultz2001)

The questionnaire also contained Witkinrsquos Embedded

Figures Test (EFT Witken 1950 Witken Oltman

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4236

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

Human Participant Pool A sample size of 160 wasselected in order to provide 80 power for a correlationcoefficient of 020 (Cohen 1988 p 87)

312 Materials

A questionnaire and computerized test were devel-oped to measure environmental attitudes implicitattitudes toward natural and built environments valuesand demographics Measures included the revisedversion of Dunlaprsquos New Environmental ParadigmScale (NEP) Schultzrsquos Environmental Motives Scaleand a revised version of Greenwaldsrsquo Implicit Associa-tion Test (IAT)

New Environmental Paradigm Scale The revisedversion of the NEP scale was used containing 15 itemsItems in the scale were rated on a 5-point Likert scaleranging from 1 (strongly disagree) to 5 (strongly agree)The instrument is designed to measure the degree towhich people view humans as a part of nature ratherthan as consumers or protectors of nature (Dunlap et al2000)

Environmental Motives Scale This instrument mea-sures concern about environmental problems causedby human behavior (Schultz 2000 2001) Concern forenvironmental issues is divided into three categoriesegoistic altruistic and biospheric Participants rated12 items about which they were concerned from 1 (notimportant) to 7 (supreme importance) Egoisticitems were me my future my prosperity and myhealth altruistic items were future generationshumanity people in the community children bio-spheric items were plants animals marine life andbirds

Implicit Association Test A modified version ofGreenwaldrsquos IAT was used to assess automatic con-ceptndashattribute associations (cf Farnham et al 1999) Acomputerized test was created to measure response time(in ms) needed to classify words associated with naturaland built environments The procedure consisted ofseven blocks of trials Prior to beginning the testparticipants were informed that the task involvedmatching words with categories They were then givenexample items and allowed to ask questions Partici-pants were instructed to go as quickly as possible and tokeep their fingers on the keyboard at all times The IATwas administered using Superlab 174 running on aPowermac 7300 computer with a 1500 color monitor

Following the introduction participants were pre-sented with seven blocks of 10 trials Each trial consistedof a word (selected at random from the relevantcategories shown in Table 1) and a set of categoriesThere were four categories of words used Me Not meNature and Built The five words and their correctcategory classification are shown in Table 1 The Meand Not me words were taken from previous studiesusing the IAT to measure self-concept The Nature and

Built words were selected on the basis of face validityThe blocks were presented as follows

Block 1 NaturendashBuilt Block 2 MendashNot me Block 3 NatureMendashBuiltNot me Block 4 NatureMendashBuiltNot me Block 5 BuiltndashNature Block 6 BuiltMendashNatureNot me Block 7 BuiltMendashNatureNot me

An instruction screen along with examples of Blocks1 3 and 6 are presented in Fig 1 The words werepresented in random order within each of the blocksBlocks 1 and 2 were considered practice as was Block 5Blocks 3 and 4 were considered lsquolsquocompatiblersquorsquo pairingsand Blocks 6 and 7 were considered lsquolsquoincompatiblersquorsquoThe strength of the association between lsquolsquoselfrsquorsquo andlsquolsquonaturersquorsquo is reflected in the ease with which a participantcould complete the compatible trials relative to theincompatible trials That is if a person has a cognitiveassociation between self and nature then it should beeasier for them to complete those trials than when thecategories are incompatible The reverse would be truefor a person with a strong association with builtenvironments

Two versions of the IAT procedure were developed inorder to counter-balance for order of NatureMe(Blocks 3 and 4 above) and BuiltMe (Blocks 6 and 7above) trials In the first rotation (shown above) NatureMe was presented first in the second rotation BuiltMe(shown as Blocks 6 and 7 above) was presented first (asBlocks 3 and 4) Participants were randomly assigned toone of two rotations The average error rate across the70 trials was 413 or 59 One subject had a high errorrate and was dropped yielding a working sample of 159Response latencies were averaged within each of theblocks Prior to averaging the data were screened foroutliers and errors Reaction time for trials that wereanswered incorrectly were removed The data were thenscreened for outliers on an individual basis such thatreaction times that were more than 3 standard devia-tions above or below the participantrsquos mean score wereremoved

ARTICLE IN PRESS

Table 1

Categories and words used to measure connectedness with nature

Categories

Nature Built Me Not me

Animals Building I It

Birds Car Me Other

Plants City Mine Their

Whales Factory Myself Them

Trees Street Self They

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4234

The IAT effect was then produced by subtracting theaverage response latency for the compatible trails(Blocks 3 and 4 above) from the average responselatency for the incompatible trials (Blocks 6 and 7above) Higher scores indicate faster responses forNatureMe trials which we interpret as a greaterassociation between self and nature (ie connectedness)

313 Procedure

Upon arrival participants provided informed con-sent Each participant was then randomly assigned torotation (1 or 2) and order (1 or 2) Rotation refers tothe sequence of IAT blocks described above Orderrefers to whether the participants completed the IAT orthe questionnaire first Once the researcher determinedthe order and rotation they proceeded to administer thetest Participants were then debriefed

32 Results

The first set of analyses examined the psychometricproperties of each of the measures The 12 items fromthe Environmental Motives scale were factor analyzedusing a Principal Axis analysis with an obliminrotation The results revealed a three-factor structure

that explained 66 of the total variance with a clearpattern of factor loadings The results are shown inTable 2 The factors were consistent with those found inprevious research and represent biospheric egostic andaltruistic concerns Correlations between the rotated

ARTICLE IN PRESS

NATUREOR

ME

Instructions

Press d FAST for words described by the left concept name Press k FAST for words described by the right concept name Nothing will happen if you press the wrong key press the other key immediately When you are ready press the spacebar to proceed GO FAST Some mistakes are OK

NATURE BUILT

ANIMALS

NATUREOR

ME

CAR

BUILTOR

ME

MINE

Sample Instructions

Block 3 Sample Compatible Trials The correct response here is ldquoBUILT OR NOT MErdquo

Block 6 Sample Incompatible Trials The correct response here is ldquoBUILT OR MErdquo

Block 1 Sample Practice Trials The correct response here is ldquoNATURErdquo

BUILTOR

NOT ME

BUILTOR

NOT ME

NATUREOR

NOT ME

Fig 1 Sample screenshots from the Implicit Association Test developed to measure connectedness with nature

Table 2

Factor loadings from the Environmental Motives Scale with oblimin

rotation

Item Factor 1

(egoistic)

Factor 2

(biospheric)

Factor 3

(altruistic)

Plants 076

Marine life 088

Birds 088

Animals 082

Me 081

My future 078

My prosperity 065

My health 093

Future generations 072

Humanity 070

People in the

community

060

Children 077

Note Factor loadings smaller than 020 are not shown The results

reported in this table are from Study 1 Factor loadings from Study 2

were nearly identical

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 35

factors were r frac14 002 (egoistic and biospheric) r frac14 028(biospheric and altruistic) and r frac14 051 (egoistic andaltruistic) Scores for each were created by averaging theitems Biospheric (M frac14 554 sd=139) Egoistic(M frac14 549 sd=140) Altruistic (M frac14 620 sd=096) To create scores that controlled for differencesin response tendency a mean correction procedure wasused The average response to all 12 of the environ-mental motives items was computed for each respon-dent This score was subtracted from each of the threescale scores to produce mean corrected biosperic(M frac14 020) egoistic (M frac14 025) and altruistic(M frac14 046) scores

Alpha reliability for the NEP scale was 078 Themean score was 357 (sd=046)

Scores for the IAT-Nature scale were produced byaveraging the 10 trials within each block The IAT effectproduced by subtracting the mean score for the twoblocks of NatureMe items (M frac14 1049 sd=350) fromthe mean score for two blocks of BuiltMe trials(M frac14 1335 sd=418) was 286 (sd=286 N frac14 159)An IAT effect of 0 would indicate no association and apositive IAT effect in this context indicates a generaltendency among our participants to associate self morewith nature than with built environments We alsoanalyzed the IAT data separately by rotation The scoreswere similar for the two rotations with the IAT effect forthe rotation with NatureMe BuiltNot me (ie compa-tible trials) presented first of 264 (N frac14 83) The secondrotation with BuiltMe NatureNot me (ie incompa-tible trials) first had an IAT effect of 309 (N frac14 76)

The second set of analyses examined the relationshipbetween implicit associations with nature (connected-ness) and explicit measures of environmental concernWe hypothesized that participants with a greaterimplicit connection with nature (measured with theIAT) would tend to have higher biospheric concernsand lower egoistic concerns Correlations were calcu-lated using the mean corrected egoistic altruistic andbiospheric scale scores NEP and IAT-Nature scoresAs seen in past research all three environmentalconcerns correlated significantly with the NEP egoistic(reth158THORN frac14 038 po0001) and altruistic (reth158THORN frac14018 po001) negatively and biospheric (reth158THORN frac14049 po001) positively

Biospheric and egoistic concerns had similar relation-ships to the IAT-Nature scores Biospheric concernswere positively correlated with the IAT reth158THORN frac14 021p frac14 0009 while egoistic concerns were negativelycorrelated reth158THORN frac14 016 p frac14 004

4 Study 2

The results from Study 1 showed a small butinterpretable pattern of relationships between implicit

associations with nature and specific types of environ-mental attitudes In order to more fully understandthese relationships a second study was conducted usinga similar procedure Our goals in this second study wereto replicate the findings from Study 1 (particularly therelationship between connectedness biospheric andegoistic concerns) to examine the relationship betweenglobalndashlocal processing styles and environmental atti-tudes (both explicit and implicit) and to examine thestability of IAT-Nature scores across time using a testndashretest procedure

41 Method

411 Participants

Data were obtained from 100 undergraduate students(40 males 60 females mean age=2362 sd=568)Participants were recruited from the Psychology De-partmentrsquos Human Participant Pool at California StateUniversity San Marcos A sample size of 99 was selectedin order to allow for 33 participants in each of threeretest conditions providing 80 power to detect a largeeffect (estimated r frac14 050) The anticipated effect size of050 was used rather than the anticipated effect of 020used in Study 1 because our focus in this study was ontestndashretest reliability which we expected to be large

412 Materials

A questionnaire was used to measure environmentalattitudes and the demographic variables of age andgender The questionnaire consisted of four environ-mental measures a revised version of Dunlap et alrsquosNew Environmental Paradigm (NEP 2000) SchultzrsquosEnvironmental Motives scale (2001) a self-reportedproenvironmental behavior scale (Schultz amp Zelezny1998) and a revised version of Aronrsquos Inclusion ofOther in Self scale (1992 see Schultz 2001) The IAT-nature computer program developed for Study 1 wasused to measure implicit connections with natural andbuilt environments

One of the items included in the questionnaire was ameasure of the perceived relationship between self andnature The item was an adaptation of Aron et al (1992)Inclusion of Other in Self (IOS) scale (see also Aronet al 1991) Participants were asked to select one ofseven different sets of overlapping circles labeled lsquolsquoselfrsquorsquoand lsquolsquonaturersquorsquo The item read lsquolsquoPlease circle the picturethat best describes your relationship with the naturalenvironment How interconnected are you with naturersquorsquoScores ranged from 1 (where the circles touched but didnot overlap) to 7 (where the two circles entirelyoverlapped) We refer to this modified scale as theInclusion of Nature in Self (INS) scale (see Schultz2001)

The questionnaire also contained Witkinrsquos Embedded

Figures Test (EFT Witken 1950 Witken Oltman

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4236

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

The IAT effect was then produced by subtracting theaverage response latency for the compatible trails(Blocks 3 and 4 above) from the average responselatency for the incompatible trials (Blocks 6 and 7above) Higher scores indicate faster responses forNatureMe trials which we interpret as a greaterassociation between self and nature (ie connectedness)

313 Procedure

Upon arrival participants provided informed con-sent Each participant was then randomly assigned torotation (1 or 2) and order (1 or 2) Rotation refers tothe sequence of IAT blocks described above Orderrefers to whether the participants completed the IAT orthe questionnaire first Once the researcher determinedthe order and rotation they proceeded to administer thetest Participants were then debriefed

32 Results

The first set of analyses examined the psychometricproperties of each of the measures The 12 items fromthe Environmental Motives scale were factor analyzedusing a Principal Axis analysis with an obliminrotation The results revealed a three-factor structure

that explained 66 of the total variance with a clearpattern of factor loadings The results are shown inTable 2 The factors were consistent with those found inprevious research and represent biospheric egostic andaltruistic concerns Correlations between the rotated

ARTICLE IN PRESS

NATUREOR

ME

Instructions

Press d FAST for words described by the left concept name Press k FAST for words described by the right concept name Nothing will happen if you press the wrong key press the other key immediately When you are ready press the spacebar to proceed GO FAST Some mistakes are OK

NATURE BUILT

ANIMALS

NATUREOR

ME

CAR

BUILTOR

ME

MINE

Sample Instructions

Block 3 Sample Compatible Trials The correct response here is ldquoBUILT OR NOT MErdquo

Block 6 Sample Incompatible Trials The correct response here is ldquoBUILT OR MErdquo

Block 1 Sample Practice Trials The correct response here is ldquoNATURErdquo

BUILTOR

NOT ME

BUILTOR

NOT ME

NATUREOR

NOT ME

Fig 1 Sample screenshots from the Implicit Association Test developed to measure connectedness with nature

Table 2

Factor loadings from the Environmental Motives Scale with oblimin

rotation

Item Factor 1

(egoistic)

Factor 2

(biospheric)

Factor 3

(altruistic)

Plants 076

Marine life 088

Birds 088

Animals 082

Me 081

My future 078

My prosperity 065

My health 093

Future generations 072

Humanity 070

People in the

community

060

Children 077

Note Factor loadings smaller than 020 are not shown The results

reported in this table are from Study 1 Factor loadings from Study 2

were nearly identical

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 35

factors were r frac14 002 (egoistic and biospheric) r frac14 028(biospheric and altruistic) and r frac14 051 (egoistic andaltruistic) Scores for each were created by averaging theitems Biospheric (M frac14 554 sd=139) Egoistic(M frac14 549 sd=140) Altruistic (M frac14 620 sd=096) To create scores that controlled for differencesin response tendency a mean correction procedure wasused The average response to all 12 of the environ-mental motives items was computed for each respon-dent This score was subtracted from each of the threescale scores to produce mean corrected biosperic(M frac14 020) egoistic (M frac14 025) and altruistic(M frac14 046) scores

Alpha reliability for the NEP scale was 078 Themean score was 357 (sd=046)

Scores for the IAT-Nature scale were produced byaveraging the 10 trials within each block The IAT effectproduced by subtracting the mean score for the twoblocks of NatureMe items (M frac14 1049 sd=350) fromthe mean score for two blocks of BuiltMe trials(M frac14 1335 sd=418) was 286 (sd=286 N frac14 159)An IAT effect of 0 would indicate no association and apositive IAT effect in this context indicates a generaltendency among our participants to associate self morewith nature than with built environments We alsoanalyzed the IAT data separately by rotation The scoreswere similar for the two rotations with the IAT effect forthe rotation with NatureMe BuiltNot me (ie compa-tible trials) presented first of 264 (N frac14 83) The secondrotation with BuiltMe NatureNot me (ie incompa-tible trials) first had an IAT effect of 309 (N frac14 76)

The second set of analyses examined the relationshipbetween implicit associations with nature (connected-ness) and explicit measures of environmental concernWe hypothesized that participants with a greaterimplicit connection with nature (measured with theIAT) would tend to have higher biospheric concernsand lower egoistic concerns Correlations were calcu-lated using the mean corrected egoistic altruistic andbiospheric scale scores NEP and IAT-Nature scoresAs seen in past research all three environmentalconcerns correlated significantly with the NEP egoistic(reth158THORN frac14 038 po0001) and altruistic (reth158THORN frac14018 po001) negatively and biospheric (reth158THORN frac14049 po001) positively

Biospheric and egoistic concerns had similar relation-ships to the IAT-Nature scores Biospheric concernswere positively correlated with the IAT reth158THORN frac14 021p frac14 0009 while egoistic concerns were negativelycorrelated reth158THORN frac14 016 p frac14 004

4 Study 2

The results from Study 1 showed a small butinterpretable pattern of relationships between implicit

associations with nature and specific types of environ-mental attitudes In order to more fully understandthese relationships a second study was conducted usinga similar procedure Our goals in this second study wereto replicate the findings from Study 1 (particularly therelationship between connectedness biospheric andegoistic concerns) to examine the relationship betweenglobalndashlocal processing styles and environmental atti-tudes (both explicit and implicit) and to examine thestability of IAT-Nature scores across time using a testndashretest procedure

41 Method

411 Participants

Data were obtained from 100 undergraduate students(40 males 60 females mean age=2362 sd=568)Participants were recruited from the Psychology De-partmentrsquos Human Participant Pool at California StateUniversity San Marcos A sample size of 99 was selectedin order to allow for 33 participants in each of threeretest conditions providing 80 power to detect a largeeffect (estimated r frac14 050) The anticipated effect size of050 was used rather than the anticipated effect of 020used in Study 1 because our focus in this study was ontestndashretest reliability which we expected to be large

412 Materials

A questionnaire was used to measure environmentalattitudes and the demographic variables of age andgender The questionnaire consisted of four environ-mental measures a revised version of Dunlap et alrsquosNew Environmental Paradigm (NEP 2000) SchultzrsquosEnvironmental Motives scale (2001) a self-reportedproenvironmental behavior scale (Schultz amp Zelezny1998) and a revised version of Aronrsquos Inclusion ofOther in Self scale (1992 see Schultz 2001) The IAT-nature computer program developed for Study 1 wasused to measure implicit connections with natural andbuilt environments

One of the items included in the questionnaire was ameasure of the perceived relationship between self andnature The item was an adaptation of Aron et al (1992)Inclusion of Other in Self (IOS) scale (see also Aronet al 1991) Participants were asked to select one ofseven different sets of overlapping circles labeled lsquolsquoselfrsquorsquoand lsquolsquonaturersquorsquo The item read lsquolsquoPlease circle the picturethat best describes your relationship with the naturalenvironment How interconnected are you with naturersquorsquoScores ranged from 1 (where the circles touched but didnot overlap) to 7 (where the two circles entirelyoverlapped) We refer to this modified scale as theInclusion of Nature in Self (INS) scale (see Schultz2001)

The questionnaire also contained Witkinrsquos Embedded

Figures Test (EFT Witken 1950 Witken Oltman

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4236

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

factors were r frac14 002 (egoistic and biospheric) r frac14 028(biospheric and altruistic) and r frac14 051 (egoistic andaltruistic) Scores for each were created by averaging theitems Biospheric (M frac14 554 sd=139) Egoistic(M frac14 549 sd=140) Altruistic (M frac14 620 sd=096) To create scores that controlled for differencesin response tendency a mean correction procedure wasused The average response to all 12 of the environ-mental motives items was computed for each respon-dent This score was subtracted from each of the threescale scores to produce mean corrected biosperic(M frac14 020) egoistic (M frac14 025) and altruistic(M frac14 046) scores

Alpha reliability for the NEP scale was 078 Themean score was 357 (sd=046)

Scores for the IAT-Nature scale were produced byaveraging the 10 trials within each block The IAT effectproduced by subtracting the mean score for the twoblocks of NatureMe items (M frac14 1049 sd=350) fromthe mean score for two blocks of BuiltMe trials(M frac14 1335 sd=418) was 286 (sd=286 N frac14 159)An IAT effect of 0 would indicate no association and apositive IAT effect in this context indicates a generaltendency among our participants to associate self morewith nature than with built environments We alsoanalyzed the IAT data separately by rotation The scoreswere similar for the two rotations with the IAT effect forthe rotation with NatureMe BuiltNot me (ie compa-tible trials) presented first of 264 (N frac14 83) The secondrotation with BuiltMe NatureNot me (ie incompa-tible trials) first had an IAT effect of 309 (N frac14 76)

The second set of analyses examined the relationshipbetween implicit associations with nature (connected-ness) and explicit measures of environmental concernWe hypothesized that participants with a greaterimplicit connection with nature (measured with theIAT) would tend to have higher biospheric concernsand lower egoistic concerns Correlations were calcu-lated using the mean corrected egoistic altruistic andbiospheric scale scores NEP and IAT-Nature scoresAs seen in past research all three environmentalconcerns correlated significantly with the NEP egoistic(reth158THORN frac14 038 po0001) and altruistic (reth158THORN frac14018 po001) negatively and biospheric (reth158THORN frac14049 po001) positively

Biospheric and egoistic concerns had similar relation-ships to the IAT-Nature scores Biospheric concernswere positively correlated with the IAT reth158THORN frac14 021p frac14 0009 while egoistic concerns were negativelycorrelated reth158THORN frac14 016 p frac14 004

4 Study 2

The results from Study 1 showed a small butinterpretable pattern of relationships between implicit

associations with nature and specific types of environ-mental attitudes In order to more fully understandthese relationships a second study was conducted usinga similar procedure Our goals in this second study wereto replicate the findings from Study 1 (particularly therelationship between connectedness biospheric andegoistic concerns) to examine the relationship betweenglobalndashlocal processing styles and environmental atti-tudes (both explicit and implicit) and to examine thestability of IAT-Nature scores across time using a testndashretest procedure

41 Method

411 Participants

Data were obtained from 100 undergraduate students(40 males 60 females mean age=2362 sd=568)Participants were recruited from the Psychology De-partmentrsquos Human Participant Pool at California StateUniversity San Marcos A sample size of 99 was selectedin order to allow for 33 participants in each of threeretest conditions providing 80 power to detect a largeeffect (estimated r frac14 050) The anticipated effect size of050 was used rather than the anticipated effect of 020used in Study 1 because our focus in this study was ontestndashretest reliability which we expected to be large

412 Materials

A questionnaire was used to measure environmentalattitudes and the demographic variables of age andgender The questionnaire consisted of four environ-mental measures a revised version of Dunlap et alrsquosNew Environmental Paradigm (NEP 2000) SchultzrsquosEnvironmental Motives scale (2001) a self-reportedproenvironmental behavior scale (Schultz amp Zelezny1998) and a revised version of Aronrsquos Inclusion ofOther in Self scale (1992 see Schultz 2001) The IAT-nature computer program developed for Study 1 wasused to measure implicit connections with natural andbuilt environments

One of the items included in the questionnaire was ameasure of the perceived relationship between self andnature The item was an adaptation of Aron et al (1992)Inclusion of Other in Self (IOS) scale (see also Aronet al 1991) Participants were asked to select one ofseven different sets of overlapping circles labeled lsquolsquoselfrsquorsquoand lsquolsquonaturersquorsquo The item read lsquolsquoPlease circle the picturethat best describes your relationship with the naturalenvironment How interconnected are you with naturersquorsquoScores ranged from 1 (where the circles touched but didnot overlap) to 7 (where the two circles entirelyoverlapped) We refer to this modified scale as theInclusion of Nature in Self (INS) scale (see Schultz2001)

The questionnaire also contained Witkinrsquos Embedded

Figures Test (EFT Witken 1950 Witken Oltman

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4236

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

Raskin amp Karp 1971) The EFT measures the extent towhich participants are either global or local processorsof information (see also Kuhnen Hannover amp Schu-bert 2001) The test consists of 25 complex figures and 8simple figures Participants are instructed to locate andtrace a simple geometric figure located inside a morecomplex figure The 25 items are divided into threetimed test sections Participants are given 2 min tocomplete the first section 5 min to complete thesecond section and 5 min to complete the third sectionDuring the testing period participants look at adesignated simple figure located on the back of the testbooklet and then attempted to find that figure in themore complex figure Scale scores for this measure wereproduced by summing the number of correct responsesin the second and third sections Higher scores areindicative of more local processing Conversely lowerscores are indicative of more global processing such thatit was more difficult to separate the small part from thewhole

We included the EFT in order to test for differences ininformation processing styles associated with egoisticand biospheric concerns We reasoned that a more localstyle of information processing might be positivelyrelated to egoistic concerns while a more global styleof processing might be associated with biosphericconcerns The globalndashlocal distinction is used widely incognitive psychology but to our knowledge it has yet tobe extended to studies of environmental attitudes

413 Procedure

Testing occurred in two sessions Participants wererandomly assigned to receive one of two rotations of theIAT and one of three retest conditions (immediate 1week or 4 weeks) Upon arriving at the lab participantsread and signed a consent form Next participants wereseated in front of the computer to complete the IATThe researcher read the instructions to the participantand informed the participant that the IAT would takeapproximately 5 min to complete and that they shouldgo as quickly as possible Once the IAT was completedthe questionnaire was administered

After completing the questionnaire the researcheradministered the Embedded Figures Test The research-er instructed the participant to read the directions and tocomplete two practice problems Once the participantfinished the practice problems the researcher remindedthe participant of the important aspects of the directionssuch as erasing completely and being careful to trace allof the lines of the simple figure Once the EFT wascompleted participants were informed which of thethree retest conditions they were assigned to During theretest session participants were first given the sameversion of the IAT that they had taken in the firstsession and then they were given the questionnaire tocomplete a second time Because testndashretest data is

already available for the EFT it was not administeredduring the second testing session

Participants in the immediate retest condition werethanked debriefed and asked not to discuss the studywith their peers until the end of the semesterParticipants in the 1- and 4-week retest conditions weretold that they would be fully debriefed when theyreturned and completed the second portion of the study

42 Results

Of the initial 100 participants two were droppedbecause of incomplete data leaving a useable sample of98 participants Of the 98 participants 32 were in theimmediate retest condition 33 were in the 1-week retestand 33 were in the 4-week

Chronbachrsquos alpha was used to evaluate the reliabilityfor all of the questionnaire measures All scales showedan acceptable level of reliability NEP (080 test 087retest) biospheric environmental concerns (090 test090 retest) altruistic concerns (080 test 083 retest)egoistic concerns (091 test 093 retest) self-ratedbehavior (080 test 083 retest)

A principal axis factor analysis of the 12 environ-mental motives items (with an oblimin rotation)revealed a 3-factor structure that accounted for 68of the variance The pattern of item loadings was similarto that found in Study 1 Factor 1 (egoistic) me (090)my lifestyle (072) my health (082) my future (095)Factor 2 (biospheric) plants (085) marine life (086)birds (082) animals (080) Factor 3 (altruistic) peoplein the community (062) future generations (063)humanity (092) and children (054) Correlationsbetween the rotated factors were r frac14 007 (egoistic andbiospheric) r frac14 016 (biospheric and altruistic) and r frac14043 (egoistic and altruistic) Scores for the threeenvironmental motives scales were produced by aver-aging the four items in each biospheric (M frac14 547sd=110) egoistic (M frac14 513 sd=158) and altruis-tic (M frac14 606 sd=095) Following the analyticprocedure used in Study 1 a mean corrected score wascreated by subtracting the total of all 12 items (grandmean=551) from the scale score for each of the threemeasures A similar procedure was used for the retestdata

On the IAT participants made an average of 364(sd=352) errors and had an average of 102(sd=204) outliers (response latencies below 300 msor above 3000 ms)1 Participants were faster for thecompatible (NatureMe and BuiltNot me) trials(M frac14 963 sd=25694) than they were for the

ARTICLE IN PRESS

1 Following Greenwald et al (1998) we screened for outliers on an

individual basis Any score below 300 ms was replaced with a score of

300 any score above 3000 was replaced with a score of 3000 Response

times for trials answered incorrectly were excluded

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 37

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

incompatible (BuiltMe and NatureNot me) trials(M frac14 1207 sd=32965) Thus the overall IAT effectshowing a preference for nature was 244 ms The IATeffect was also examined separately by rotation Resultsshowed a slightly smaller IAT effect when the compa-tible (ie NatureMe BuiltNot me) trials were firstparticipants were faster (mean compatible=928 meanincompatible=1211 IAT effect=283 N frac14 49) thanwhen the incompatible (BuiltMe NatureNot me) trialswere presented first (mean compatible=882 meanincompatible =1066 IAT effect=428 N frac14 49) Rota-tion did not have a significant effect on any of thecorrelations reported below

The testre-test reliability was evaluated for all explicitmeasures as well as the computerized IAT usingcorrelation coefficients The results are shown inTable 3 As seen in the table the level of re-testreliability across the measures was extremely high Theexplicit measures showed a higher degree of stabilityacross time but the IAT-nature test also showed areasonable level of testndashretest reliability r frac14 045 p frac14

0001 (immediate condition) r frac14 046 p frac14 001 (1-weekcondition) r frac14 040 p frac14 002 (4-week condition)

To test the hypotheses predicting a relationshipbetween the explicit measures and the computerizedimplicit test Pearson correlations were calculated usingthe data from the first testing session The resultingmatrix is shown in Table 4 As predicted there was asignificant and positive relationship between the bio-spheric environmental concerns and the IAT-naturemeasure r frac14 027 p frac14 001 Also as predicted there wasa significant and negative relationship between theegoistic concerns and IAT scores r frac14 022 p frac14 004Finally there was a significant and positive correlationbetween the Inclusion of Nature in Self scale (M frac14 374sd=132) and IAT-Nature scores r frac14 026 p frac14 001Similar results were obtained using the data from theretest testing session

Correlation coefficients were also used to test therelationship between environmental attitudes (bothimplicit and explicit) and information processing style(lower scores indicate global processing and higherscores indicate local processing) As predicted there wasa significant and negative relationship between IATscores and the embedded figures test (M frac14 1112sd=474) r frac14 023 p frac14 002 There were no sig-nificant relationships between the EFT and the meancorrected egoistic altruistic or biospheric scales

5 Discussion

The two studies reported in this paper were designedto examine implicit connections with nature Our goalswere (1) to develop an implicit measure of the extent towhich individuals associate themselves with the naturalenvironment (2) to examine the relationship betweenthese implicit connections and explicit measures ofenvironmental attitudes and (3) to examine the stabilityof implicit connections with nature across time

The results provide clear evidence for the usefulness ofimplicit measures in environmental research and of the

ARTICLE IN PRESS

Table 3

Testretest correlations for immediate 1 and 4 weeks

Scale Immediate

retest

(N frac14 32)

1-Week

retest

(N frac14 33)

4-Week

retest

(N frac14 33)

New Environmental

Paradigm (NEP)

094 092 083

Inclusion of Nature in Self

(INS)

098 090 084

Self-reported

environmental behavior

097 088 086

Egoistic concerns (mean

corrected)

093 080 076

Altruistic concerns (mean

corrected)

087 074 088

Biospheric concerns

(mean corrected)

095 086 086

Implicit Association Test

(IAT)

045 046 040

po005 po001

Table 4

Correlation matrix

NEP INS EFT Behavior Egoistic Biospheric Altruistic

INS 030

EFT 007 009

Behavior 027 026 008

Egoistic 019 028 011 025

Biospheric 036 040 008 024 085

Altruistic 032 025 004 000 021 034

IAT 018 026 023 006 022 027 011

Note Results based on data from the first testing session Similar results were obtained for the retest data INS Inclusion of Nature in Self EFT

Embedded Figures Test (higher scores indicate more local processing) IAT Implicit Association Test po005 po001 N frac14 98

PW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4238

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

importance of connectedness with nature in under-standing environmental attitudes The data showed abasic IAT effect wherein it was easier for participants toassociate themselves with nature than with builtenvironments In Study 1 the basic IAT effect was286 ms and in Study 2 the effect was 244 ms Thispositive IAT must be interpreted with caution Oneinterpretation would suggest that people have a basictendency to associate themselves with naturemdashaninterpretation that is consistent with the biophiliahypothesis (cf Kellert 1996) Relatedly it could bethat people prefer natural over built environments(Kaplan amp Kaplan 1989) and this preference makesthe association with self easier Such an interpretation isconsistent with Greenwald et alrsquos (2002) recent theoriz-ing about implicit attitudes However tempting suchinterpretations are it is also possible that the tendencytoward connectedness is a result of our college studentsample and that a more representative sample wouldnot show the same tendency Thus the interpretation ofthe positive IAT scores in our two studies remains to beaddressed in future research

The results across the two studies also showed a smallbut consistent pattern of correlations between IATscores and explicit measures of environmental attitudesImplicit connections with nature were positively corre-lated with biospheric environmental concerns (r frac14 021in Study 1 and r =027 in Study 2) and negativelycorrelated with egoistic environmental concerns(r frac14 016 in Study 1 and r frac14 022 in Study 2) Inaddition the results from Study 2 showed a positiverelationship between the implicit IAT measure and anexplicit measure of inclusion with nature (r frac14 026) andwith scores on the Embedded Figures Test (r frac14 023)

On the surface these effects may appear smallHowever previous research with the IAT has alsofound correlations between implicit and explicit mea-sures to be around 020 For example CunninghamPreacher and Banaji (2001) report correlations rangingfrom 008 to 026 between an IAT measure of prejudiceand an explicit measure of prejudice However Cun-ningham et al (2001) also reported a confirmatoryfactor analysis showing a clear higher order factor forimplicit measures of prejudice and a latent correlationcoefficient of 045 between implicit and explicit pre-judice See also Greenwald et al (1998) Although thesize of the relationships between implicit and explicitmeasures is small we believe that the statisticalsignificance of the correlations combined with theoverall pattern of correlations is theoretically mean-ingful and suggest that implicit measures can be usefulin future research on environmental attitudes

Scores on the IAT-Nature procedure were relativelystable across time The results from Study 2 showed atestndashretest correlation of 045 046 and 040 forimmediate 1-week and 4-week periods respectively

These are consistent with findings by other researchersusing the IAT For example Cunningham et al (2001)reported testndashretest correlations ranging from 016 to050 across four retest periods separated by 2 weekswith an average correlation of 032 It is interesting tonote that the retest correlations in our study were similaracross time Often in testndashretest analyses the relation-ship between the measures decreases over time That iswith explicit measures of attitudes recall for previousresponses changes in attitudes or fluctuations due todaily experiences lead to a decrease in the stability of thescores We believe that an individualrsquos implicit connec-tion with nature is more stable across time than explicitmeasures because it is not affected by memory it is notconcealable or influenced by response bias and it is lessaffected by day-to-day experiences This is not tosuggest that implicit connections with nature do notchange across time or are not influenced by surround-ings but merely that they are less influenced by suchvariables than are explicit measures See Karpinski andHilton (2001) for an examination of the contextualinfluence on implicit attitudes

51 Implications

The finding that implicit connections with nature arecorrelated with environmental attitudes has severaltheoretical implications The issue of whether humansare part of nature or separate from nature has beendiscussed and debated at length among philosophersand environmentalists (Lovejoy 1936 Pepper 1984Rothenberg 1987) Beliefs about the relationshipbetween self and nature is a core element of the writingand theorizing of deep ecology (Naess 1988) land ethics(Callicott 1999 2002) ecological identity (Bragg 1996Light 2000) and ecopsychology (Cohen 1997 Spilner1997) As Ponting (1991) stated in his Green History of

the World

One of the fundamental issues addressed by alltraditions is the relationship between humans andthe rest of nature Are humans an integral part ofnature or are they separate from it and in some waysuperior to it The answer to this question is crucial indetermining how different thinkers and religionsdecide which human actions can be regarded aslegitimate or morally justified From this flow otherrelated questions about whether all plants andanimals in the world are there solely for the benefitof humans about whether humans have a responsi-bility to guard and take care of the rest of nature(p 141)

The data reported in this paper affirm the importanceof connectedness in understanding attitudes aboutenvironmental issues At a psychological level thedegree to which an individual associates him or herself

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 39

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

with nature is directly related to the types of attitudesthat she develops In essence individuals who associatethemselves with the natural environment tend to holdbroader sets of concerns for environmental issues (iebiospheric attitudes) An individual with less of anassociation between self and nature can still beconcerned about environmental issues but these con-cerns are more narrow and focused on issues thatdirectly affect the individual Yet despite the centralityof the concept of connectedness in the environmentalliterature only a few studies have operationalized itusing explicit measures and the studies reported in thecurrent article represent the first attempt to do soimplicitly

But why use an implicit measure Indeed implicitmeasures are more difficult to develop and administerthe effect sizes tend to be smaller and the amount ofvariability and lsquolsquonoisersquorsquo in the data are larger An explicitmeasure like the Inclusion of Nature in Self scale usedin Study 2 provides an useful tool for measuringconnectedness However in addition to the standardissues of response set and bias common in self-reportmeasures explicit measures also make a very importantassumption they assume that the individual knows andcan articulate his or her belief If as Dunlap et al (2000)have suggested such beliefs are lsquolsquoprimitiversquorsquo they maynot be readily available for recall or easily articulatedon a self-report questionnaire In this case an implicitmeasure provides a useful alternative and may offer abetter technique for measuring the strength of anassociation

Two other aspects of our results warrant additionalcomment First are the results found with the EmbeddedFigures Test To our knowledge this is the first reportedstudy in which the EFT has been used to predictenvironmental attitudes and the purpose of using it andthe meaning of the findings may not be readily apparentIn developing our inclusion model of environmentalattitudes we have argued that egoistic concerns aremore narrowly focused on issues that directly affect theindividual while biospheric concerns reflect a broader(ie more inclusive) set of concerns (Schultz 20002002b) Indeed this was the basis for our interpretationof earlier research showing that biospheric attitudeswere positively related to Schwartzrsquo self-transcendentvalues while egoistic attitudes were positively associatedwith values of self-enhancement (cf Schultz amp Zelezny1999 Schultz 2001)

Following this line of reasoning and supported withrecent research by Kuhnen et al (2001) on the semanticndashprocedural interface model of the self we postulatedthat there may be differences in information processingstyle associated with connectedness egoistic and bio-spheric concerns We predicted that less connection withnature would be associated with more local informationprocessing while a greater connection would be

associated with global processing Although our findingswith the EFT were somewhat disappointing the small-but-significant correlation with IAT nature does suggestprocessing differences We believe this finding and thebasic line of reasoning deserves further research

The second aspect of our results that warrantsdiscussion is our approach to measuring egoisticaltruistic and biospheric attitudes Following Sternand Dietzrsquos (1994) influential article on the value basesof environmental attitudes we set out to develop a set ofitems to measure each of these concerns (rather than thevalues associated with them) However unlike Stern andDietz (1994) we view these three sets of attitudes asrelated rather than distinct Our basic approach is toview egoistic altruistic and biospheric attitudes asaspects of a second-order factor which is reflected inour use of correlated (ie oblique) factor structures Inprevious research we have used the mean score of theitems to form 3 subscales This approach generates ahigh Chronbachrsquos alpha (even with only four items)However the scale scores appear to reflect a fair degreeof response set across the items as well as socialdesirability

In an attempt to correct for these problems theanalyses reported in this paper are based on mean-corrected scores Although more sophisticated transfor-mations or partialling procedures could be developedwe find that the mean-correction procedure yieldsscale scores that are meaningful stable acrosstime and that generate consistent patterns of relation-ships with other measures It is interesting to notehowever that results based on the full score (not meancorrected) yields results that are very similar to thosegenerated by the mean corrected scores In futureresearch using the Environmental Motives scale werecommend using the mean-correction procedureadopted in this paper

52 Directions for future research

The finding that associations between self and natureare relatively stable across time and that they correlatewith explicit attitudes of concern raises several othertestable hypotheses First are implicit connections withnature associated with specific patterns of behaviorlifestyle or life decisions Several recent publicationshave suggested that many people living in industrializedcountries have begun to reject the consumptive lifestylesso prevalent today (Ray amp Anderson 2000 Degenhardt2002) Are people who choose to live alternativelifestyles higher in connectedness with nature Similarlyare associations with nature related to specific day-to-day behaviors like energy conservation green buyingand recycling The VBN theory of Stern et al (1999)and Stern (2000) would suggest that connectedness

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4240

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

would only lead to behavior in instances where valuesassociated with the beliefs are activated

Second does connectedness vary across culturesSeveral recent publications have documented country-level differences in attitudes about environmental issues(Dunlap et al 1993 Schultz 2001) In essencerespondents from the United States and WesternEurope tend to be less biospheric and more egoistic intheir approach to environmental issues while respon-dents from Central America and South America tend tobe more biospheric Perhaps these differences are theresult of underlying differences in inclusion Consider-able social psychological research has suggested that oneof the fundamental dimensions along which culturesvary is self-construal In some cultures self is definedbroadly and interconnected with other people while inother cultures self is defined more narrowly andindependent of other people (cf Markus amp Kitayama1991 Singelis 1994 Kuhnen et al 2001) Perhaps aconnection with nature is simply an extension of aninterdependent self not only is who I am dependent onmy relationship with other people but who I am is alsodependent on my relationship with the environmentaround me As Bateson and Bateson (1987 p 177)stated lsquolsquoWhat we believe ourselves to be should becompatible with what we believe of the world aroundusrsquorsquo

Finally how malleable is connectedness Our resultsshow a moderate degree of stability in connectednessacross time However our measures were obtained in acontrolled laboratory environment Perhaps connected-ness would vary depending on the context in which itwas assessed (eg a lab room a classroom an officebuilding a park a forest) A considerable amount ofenvironmental research has demonstrated the trans-forming ability of encounters with nature (Nabhan ampTrimble 1994 Chawla 1998 Palmberg amp Kuru 2000)Perhaps one of the ways in which encounters withnature can transform an individual is through a sense ofconnectedness Similarly environmental education haslong sought to promote proenvironmental attitudes andbehaviors but has achieved only limited success(Zelezny 1999) Perhaps educational activities thatpromote a connection with nature will have longerlasting effects

In closing the research reported in this paper hasprovided evidence for individual differences in the extentto which people associate themselves with nature Wedeveloped an Implicit Association Test to measureconnectedness with nature The results suggest thatconnectedness is fairly stable across time and that it isassociated with biospheric concerns (positively) andegoistic concerns (negatively) Armed with this new toolfor measuring implicit connections with nature futureresearch can begin to explore a number of theoreticaland applied questions

References

Aron A Aron E N amp Smollan D (1992) Inclusion of other in the

self scale and the structure of interpersonal closeness Journal of

Personality and Social Psychology 63 596ndash612

Aron A Aron E N Tudor M amp Nelson G (1991) Close

relationships as including other in the self Journal of Personality

and Social Psychology 60 241ndash253

Bateson G amp Bateson M C (1987) Angels fear Towards an

epistemology of the sacred New York Macmillan

Bragg E A (1996) Towards ecological self Deep ecology meets

constructionist self theory Journal of Environmental Psychology

16 93ndash108

Callicott J B (1999) Beyond the land ethic More essays in

environmental philosophy Albany State University of New York

Press

Callicott J B (2002) In W Ouderkirk amp H Jim (Eds) Land value

community Callicott and environmental philosophy New York

State University of New York Press

Chawla L (1998) Significant life experiences revisited A review of

research on sources of environmental sensitivity Environmental

Education Research 4 369ndash383

Cohen J (1988) Statistical power analysis for the behavioral sciences

(2nd ed) Hillsdale NJ Lawrence Erlbaum

Cohen M J (1997) Reconnecting with nature Finding wellness

through restoring your bond with the earth Corvallis OR Ecopress

Cunningham W A Preacher K J amp Banaji M (2001) Implicit

attitude measures Consistency stability and convergent validity

Psychological Science 12 163ndash170

Degenhardt L (2002) Why do people act in sustainable ways Results

of an empirical survey of lifestyle pioneers In P Schmuck amp P

Schultz (Eds) Psychology of sustainability New York Kluwer

Dunlap R E (1991) Trends in public opinion toward environmental

issues 1965ndash1990 Society and Natural Resources 4 285ndash312

Dunlap R E amp Van Liere K (1978) The new environmental

paradigm Journal of Environmental Education 9 10ndash19

Dunlap R E Gallup G amp Gallup A (1993) Global environmental

concern Results from an international public opinion survey

Environment 35 7ndash15 33ndash39

Dunlap R E Van Liere K Mertig A amp Jones R E (2000)

Measuring endorsement of the New Ecological Paradigm A

revised NEP scale Journal of Social Issues 56 425ndash442

Dutcher D (2000) Landowner perceptions of protecting and establish-

ing riparian forests in central Pennsylvania Doctoral dissertation

Pennsylvania State University School of Forest Resources

Eagly A amp Chaiken S (1993) The psychology of attitudes Fort

Worth TX Harcourt Brace Jocanovich

Farnham S D Greenwald A G amp Banaji M R (1999) Implicit

self-esteem In D Abrams amp M A Hogg (Eds) Social identity

and social cognition (pp 230ndash248) Malden MA Blackwell

Greenwald A G amp Banaji M (1995) Implicit social cognition

Attitudes self-esteem and stereotypes Psychological Review 102

4ndash28

Greenwald A G Banaji M R Rudman L A Farnham S D

Nosek B A amp Mellot D S (2002) A unified theory of implicit

attitudes stereotypes self-esteem and self-concept Psychological

Review 109 3ndash25

Greenwald A G amp Farnham S D (2000) Using the implicit

association test to measure self-esteem and self-concept Journal of

Personality and Social Psychology 79 1022ndash1039

Greenwald A G McGhee D E amp Schwartz J L K (1998)

Measuring individual differences in implicit cognition The implicit

association test Journal of Personality and Social Psychology 74

1464ndash1480

Hwang Y H Kim S I amp Jeng J M (2000) Examining the

causal relationship among selected antecedents of responsible

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash42 41

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242

environmental behavior Journal of Environmental Education 2000

19ndash26

Kahn P (1999) The human relationship with nature Development and

culture Cambridge MA MIT Press

Kals E amp Maes J (2002) Sustainable development and emotions In

P Schmuck amp P W Schultz (Eds) Psychology and sustainability

(pp 97ndash122) New York Kluwer

Kaplan R amp Kaplan S (1989) The experience of nature A

psychological perspective Cambridge Cambridge University Press

Karpinski A amp Hilton J L (2001) Attitudes and the Implicit

Association Test Journal of Personality and Social Psychology 81

774ndash788

Kellert S (1996) The value of life Washington DC Island Press

Kidner D (2001) Nature and psyche Radical environmentalism and

the politics of subjectivity New York State University of New

York Press

Kuhnen U Hannover B amp Schubert B (2001) The semantic-

procedural interface model of the self The role of self-knowledge

for context dependent versus context-independent modes

of thinking Journal of Personality and Social Psychology 80

397ndash409

Leopold A (1949) A sand county almanac and sketches here and there

New York Oxford University Press

Light A (2000) What is an ecological identity Environmental

Politics 9 59ndash82

Lovejoy A (1936) The great chain of being Cambridge MA Harvard

University Press

Markus H R amp Kitayama S (1991) Culture and the self

Implications for cognition emotion and motivation Psychological

Review 98 224ndash253

Nabhan G P amp Trimble S (1994) The geography of childhood why

children need wild places Boston Beacon

Naess A (1988) Identification as a source for deep ecological

attitudes In M Tobias (Ed) Deep ecology (pp 256ndash270) San

Diego Avant books

Olson J amp Zanna M (1993) Attitudes and attitude change Annual

Review of Psychology 44 117ndash154

Opotow S (1994) Predicting protection Scope of justice and the

natural world Journal of Social Issues 50(3) 49ndash63

Opotow S amp Weiss L (2000) Denial and the process of moral

exclusion in environmental conflict Journal of Social Issues 56

475ndash490

Palmberg I E amp Kuru J (2000) Outdoor activities as a basis for

environmental responsibility Journal of Environmental Education

31 32ndash36

Pepper D (1984) The roots of modern environmentalism London

Croom Helm

Ponting C (1991) A green history of the world the environment and

the collapse of great civilizations New York Penguin

Ray P H amp Anderson S R (2000) The cultural creatives How 50

million people are changing the world New York Harmony Books

Rokeach M (1973) The nature of human values New York Free

Press

Rothenberg D (1987) A platform of deep ecology The Environmen-

talist 7 185ndash190

Schultz P W (2000) Empathizing with nature The effects of

perspective taking on concern for environmental issues Journal of

Social Issues 56 391ndash406

Schultz P W (2001) Assessing the structure of environmental

concern Concern for the self other people and the biosphere

Journal of Environmental Psychology 21 327ndash339

Schultz P W (2002a) Inclusion with nature Understanding human-

nature interactions In P Schmuck amp P W Schultz (Eds) The

psychology of sustainable development (pp 61ndash78) New York

Kluwer

Schultz P W (2002b) Environmental attitudes and behaviors across

cultures In W J Lonner D L Dinnel S A Hayes amp D N

Sattler (Eds) Online readings in psychology and culture Western

Washington University Department of Psychology Center for

Cross-Cultural Research Bellingham WA Web site http

wwwwwueduBculture

Schultz P W amp Zelezny L (1998) Values and proenvironmental

behavior A five-country survey Journal of Cross-Cultural Psychol-

ogy 29 540ndash558

Schultz P W amp Zelezny L (1999) Values as predictors of

environmental attitudes Evidence for consistency across 14

countries Journal of Environmental Psychology 19 255ndash265

Schultz P W Oskamp S amp Mainieri T (1995) Who recycles and

when A review of personal and situational factors Journal of

Environmental Psychology 15 105ndash121

Schwartz S (1994) Are there universal aspects in the structure

and contents of human values Journal of Social Issues 50

19ndash45

Singelis T M (1994) The measurement of independent and

interdependent self-construals Personality and Social Psychology

Bulletin 20 580ndash591

Spilner M (1997) Connecting with nature Prevention 49

128ndash132

Stern P C amp Dietz T (1994) The value basis of environmental

concern Journal of Social Issues 56 407ndash424

Stern P C (2000) Toward a coherent theory of environmentally

significant behavior Journal of Social Issues 56 407ndash424

Stern P C Dietz T amp Guagnano G A (1995a) The new

environmental paradigm in social psychological perspective

Environment and Behavior 27 723ndash745

Stern P C Dietz T Kalof L amp Guagnano G A (1995b) Values

beliefs and proenvironmental action Attitudes formation toward

emergent attitude change Journal of Applied Social Psychology 25

1611ndash1636

Stern P C Dietz T Abel T Guagnano G A amp Kalof L

(1999) A value-belief-norm theory of support for social move-

ments The case of environmental concern Human Ecology Review

6 81ndash97

Thompson S G amp Barton M A (1994) Ecocentric and anthropo-

centric attitudes toward the environment Journal of Environmental

Psychology 14 149ndash157

Weigert A J (1997) Self interaction and natural environment

Refocusing our eyesight New York SUNY Press

Witken H A (1950) Individual differences in ease of perception of

embedded figures Journal of Personality 19 1ndash15

Witken H Oltman P Raskin E amp Karp S (1971) Manual Group

embedded figures test Palo Alto CA Consulting Psychologists

Press

Zelezny L (1999) Educational interventions that improve environ-

mental behaviors A meta-analysis Journal of Environmental

Education 31 5ndash15

ARTICLE IN PRESSPW Schultz et al Journal of Environmental Psychology 24 (2004) 31ndash4242