, 44 W
2003e 5 N
future behavior. 2004 Elsevier Inc. All rights reserved.
behavior. These investigations have focused both onley, 2002; Macrae & Johnston, 1998).1
q The authors thank Kevin Carlsmith, Robyn Leboeuf, BenotMonin, and Sam Sommers for their helpful comments on an earlierversion of the manuscript.
* Corresponding author.E-mail address: firstname.lastname@example.org (L.D. Nelson).
1 We note that Darley and Batson (1973) used another classichelpful exemplarthe Good Samaritanand failed to nd signicanteects. Because participants in this experiment were also exposed totwo unhelpful exemplars (the priest and the Levite), it should perhapsnot be a surprise that this manipulation was not entirely successful.
Journal of Experimental Social Psychology 41 (2005) 4234300022-1031/$ - see front matter 2004 Elsevier Inc. All rights reservedthe stable individual dierences that reect altruistic mo-tives and the subtle situational factors that can impacthelping, reective of a dichotomy in the broader eortof psychologists to document the determinants of hu-man behavior. Although research has suggested thatbehavior reects the conscious workings of a complexpsychological system (e.g., Ajzen, 1991; Carver & Sche-ier, 1998), a growing subset of ndings has indicated
Early research on helping focused on the ways inwhich modest manipulations could lead to dramaticchanges in behavior. Darley and Batson (1973), forexample, showed that merely telling people that theywere running late reduced the likelihood that they wouldstop to help someone slumped in a doorway. Macraeand Johnston (1998) showed that an even more subtlemanipulation could impact helping behavior, as par-ticipants primed with helping-related words wereLike. . .Superman you will come to save me. . .-Aimee Mann Save Me
Psychologists have exhaustively researched factorsthat promote and inhibit altruism, with two primarygoals: Understanding the processes which underlie help-ing, and developing strategies for increasing helping
that subtle priming techniques can cause behavior with-out conscious regulation. Such primes have been shownto impact an increasingly diverse set of behaviors, fromintellectual performance (Dijksterhuis & van Knippen-berg, 1998), conformity (Epley & Gilovich, 1999), andwalking speed (Bargh, Chen, & Burrows, 1996) to onemost relevant to the present investigation, spontaneoushelping behavior (Garcia, Weaver, Moskowitz, & Dar-From student to superhero: Situa
Leif D. Nelsona,*,a Stern School of Business, New York University
b Sloan School of Management, Ma
Received 16 SeptemberAvailable onlin
The present research uses priming techniques to modify commcontrol condition, people primed with the exemplar Supermanprimed with the category superhero saw themselves as more likthese eects to real-world planned helping behavior, by demonsism. Finally, Study 3 showed that these changes in initial cominitial exposure. These results demonstrate that eeting situatdoi:10.1016/j.jesp.2004.08.003nal primes shape future helpingq
ichael I. Nortonb
est Fourth Street, New York, NY 10012, USA
setts Institute of Technology, USA
; revised 16 June 2004ovember 2004
ent to and engagement in future helping behavior. Relative to athemselves as less likely (Studies 1a and 1b), and participantsStudy 1a), to help in hypothetical situations. Study 2 extendedg that these primes impacted commitment to future volunteer-ent impacted volunteering behavior up to three months afterprimes can impact not only spontaneous behavior, but also
erimesubsequently more likely to help someone pick upspilled pens. While this line of research has focused onthe ease with which such spontaneous helping behaviorcan be manipulated, a parallel line of research has exam-ined a dierent type of helping behaviorplanned, orlong-term, helping behaviorwith often very dierentdeterminants. We chose volunteerism as our instantia-tion of this kind of helping behavior, a form of helpingthat has received increased attention in recent years (seePutnam, 2000; Wilson, 2000). Volunteering, unlike thespontaneous helping behaviors examined in many inves-tigations, may require a great deal of time and eort(Omoto & Snyder, 1995). The act of volunteering, more-over, has consequences for long-term behavior: a split-second decision to volunteer may lead to weeks, months,or even years of commitment. Research on volunteeringhas shown a relative insensitivity to situational inu-ences: Because the decision to volunteer involves com-mitment beyond the immediate future, volunteeringhas been shown to be best predicted by more stable fac-tors, such as individual dierences in prosocial orienta-tion (e.g., Penner & Finkelstein, 1998), and, as is thecase with many behaviors, prospective volunteers ownpast behavior (e.g., Piliavin & Callero, 1991). Thesetwo stable factors, individual dierences and past behav-ior, are by their very denition situationally inalterable.Given these constraints, one possible inference is thatthe subliminal priming procedures shown to inuencemany types of spontaneous behavior would be unlikelyto impact behaviors that are predicted by more stablefactors, like volunteering.
Though research on volunteerism suggests that situa-tional factors should have little eect, some ndings indi-cate that even stable dispositions can be impacted bypriming manipulations. In a prisoners dilemma para-digm, for example, individuals primed with competitive-ness were more likely to compete, but only if they hadalready shown a predisposition towards competing(Neuberg, 1988). More recent work has further suggestedthat pro-social and pro-self dispositions are further mod-erated by individual self-consistency (Smeesters, Warlop,Van Avermaet, Corneille, & Yzerbyt, 2003). Unlike thenon-competitive response in a prisoners dilemma situa-tion, volunteeringdue to its positive connotationsmay be a domain towards which individuals generallymight consider themselves disposed: The vast majorityof people, for example, see themselves as more likelythan the average person to donate blood (Allison, Mes-sick, & Goethals, 1989). Primes that activate helpingconstructs, therefore, may have the potential to inuencebehaviors that are generally seen as resistant to the im-pact of eeting situational forces.
Echoing the behavioral priming research cited above,we used a category and exemplar prime paradigm (e.g.,Dijksterhuis et al., 1998) to prime helpfulness, selecting
424 L.D. Nelson, M.I. Norton / Journal of Expthe category superheroes and the exemplar Super-Overview
In a series of studies, we used situational primes de-signed to elicit increased or decreased helping behavior.Participants were primed to think about a helpful cate-gory (e.g., superheroes), or an exemplar member of thatcategory (e.g., Superman). We had three primary goalsin the studies reported below. First, we wanted to dem-onstrate that situational primes can both make peoplethink of themselves as more helpful and cause them topredict more helpful behavior in the future (Studies 1aand 1b). Second, we wanted to show that these primescould move beyond impacting spontaneous behaviorsand make people more likely to volunteer for a realcommunity service group (Study 2). Our third and mostimportant goal was to show that such commitment tovolunteering, even when induced through priming,would lead to increased volunteering behavior in the fu-ture, many months after initial exposure (Study 3).Moreover, because we apply the exemplar/category par-adigm in our attention to helping behavior, some of thestudies specically compare neutral controls with Super-man and/or superhero primes (Studies 1a and 1b, andStudy 2), while others compare Superman to superheroprimes (Study 1a and Study 3). Though the quote withwhich we opened this paper illustrates Manns faith thatmanboth highly altruistic constructsas our targetstimuli. Previous research has shown that individualscompare themselves to the standards set by such socialstimuli (e.g., Festinger, 1954; Mussweiler, 2003). Typi-cally, these comparison processes result in assimilationin both judgments and behavior (e.g., Bargh et al.,1996; Dijksterhuis & Bargh, 2001; Dijksterhuis & vanKnippenberg, 1998; Kawakami, Dovidio, & Dijkster-huis, 2003), in part because when making comparisons,people rst focus on shared features (e.g., Srull & Gae-lick, 1983), a focus which frequently leads to assimila-tion due to activation of this information (Mussweiler,2003). Although people default to similarity testingand the assimilation that resultspeople do engage indissimilarity testing as well (Mussweiler, 2003). This lesscommon comparison is more likely to occur when com-parisons are made with extreme, unambiguous stan-dards (e.g., Dijksterhuis et al., 1998; Herr, 1986; Herr,Sherman, & Fazio, 1983; LeBoeuf & Estes, in press;Lockwood & Kunda, 1997; Stapel, Koomen, & vander Plight, 1997), precisely the kind of standard that asuperhuman target such as Superman represents. Thus,we predicted that people would contrast from helpfulexemplar primes (Superman), but assimilate to helpfulcategory primes (superheroes) in judgments of them-selves, predictions of their behavior, and their actual
ntal Social Psychology 41 (2005) 423430people like Supermanthe most helpful exemplar the
similar in format but asked questions that were notexplicitly related to the primed constructs. These ques-
pleted this page, they continued to work through thepacket until nished, at which point they were probed
erimental Social Psychology 41 (2005) 423430 425tions were intended to be either self-enhancing, butnot directly related to the prime (e.g., likelihood of win-ning an essay contest), or unrelated (e.g., likelihood ofchoosing pizza or Chinese food for dinner), and usedauthors could generatewill behave altruistically, wepredict that participants primed with this construct willhelp less than the average person, while those primedwith superheroes will help more.
Fifty-six Princeton undergraduates participated inthe experiment as partial fulllment of a course require-ment, were recruited via electronic mail and telephone,and were scheduled in groups of three to participate inthe experiment. Participants were seated in three sepa-rate quiet rooms, and completed large questionnairepackets containing the manipulation and dependentmeasures. Participants were told to answer every ques-tion and to complete the questionnaire in order, withoutreturning to previous pages.
Participants rst answered a brief set of demograph-ics questions, then completed the priming manipulation.Analogous to the procedures used in other research(Dijksterhuis & van Knippenberg, 1998; Dijksterhuiset al., 1998), participants in the superhero (Superman)condition were asked the following: For this task wewould like you to describe the characteristics of a super-hero (Superman). Think of a superhero (Superman) andlist the behaviors, values, lifestyle, and appearance asso-ciated with these characters (this character). Partici-pants in the control condition were given nearlyidentical instructions, but were asked to describe a dormroom. Following the priming procedure, participantscompleted the (purportedly unrelated) dependent mea-sure, which asked participants to evaluate their behaviorin a series of scenarios. The critical measure of plannedhelping behavior was, An elderly woman gets on acrowded subway on which you are riding. Althoughall the seats are taken and many people are standing,you have a seat. Relative to the average Princeton stu-dent how likely is it that you would oer your seat tothis woman? (1: much less likely, 8: same, 15: muchmore likely). We used the relative to average measurefor two reasons. First, it provides a reference point thatis modestly meaningful for participants. Moreover, thecomparison to a typical Princeton student provides thebest test of our hypothesisthis was the one group thatall participants belonged to and felt fairly knowledge-able about. The remaining questions on the page were
L.D. Nelson, M.I. Norton / Journal of Expthe same scale as above. When participants had com-for suspicion and debriefed.
Results and discussion
In debrieng, one participant expressed suspicionabout the unrelatedness of the various tasks; though un-able to identify the hypothesis, we excluded this partici-pants data from further analysis.
As predicted, participants primed with superhero re-ported being most likely to help (M = 11.32), followedby participants in the control (M = 9.93) and Superman(M = 8.95) conditions. The omnibus ANOVA was sig-nicant, F (2,54) = 6.93, p = .002, as was the predictedlinear contrast, F (1,54) = 13.76, p < .001. These prim-ing eects were limited only to prime-specic behaviors;as expected, none of the alternative measures (of bothself-enhancing and irrelevant behaviors) were aectedby the priming manipulation (Fs < 1).
To ensure that our manipulations were successful wehad an independent coder, blind to condition, identifythe Superman-specic features (e.g., kryptonite, phonebooth, etc.) listed by each participant. Consistent withour predictions, participants reported more of these fea-tures when primed with Superman (M = 1.1) than whenprimed with superhero (M = .04) or dorm room(M = .00), F (2,63) = 14.54, p < .001. Most importantlyfor rejecting alternative hypotheses about the manipula-tions, there was a reliable dierence between the Super-man and superhero conditions specically, t (43) = 3.78,p < .001.
Given that the two experimental conditions criticallydiered, we also wanted to show that dierences in thedependent variable did not occur as the result of partic-ipants general approach to the original feature-listingtask. We further analyzed the content of the feature list-ings in order to show that our primes generated the pre-dicted types of thoughts. Two independent coders, blindto condition, rated the overall valence of each feature(positive, negative, or neutral), and whether a listed fea-ture was related to helping behavior or not.2 There wasan acceptable 85.2% agreement rate; a second pair ofcoders resolved any dierences between the ratings.We computed an overall score for the valence of thelisted features by computing the dierence between thenumber of positive features listed and the number ofnegative features listed. Not surprisingly, features listedby participants primed with Superman (M = 6.6) andsuperhero (M = 7.4) were signicantly more positivethan the features listed by participants in the controlcondition (M = .87), F (2,54) = 116.4, p < .001. Most
2 Typical positive features: strong, handsome, and honest. Typicalnegative features: arrogant, secretive, and violent. Neutral terms were
typically non-valenced descriptors: male, big, and white.
from the category prime, and judged themselves as more
bound). Those phrases, followed by look up in the sky-its Superman were originally associated with...