26
RESIDENTS’ SUPPORT FOR TOURISM An Identity Perspective Robin Nunkoo University of Waterloo, Canada Dogan Gursoy Washington State University, USA Abstract: Studies on residents’ support for tourism have primarily been conducted in the developed world. This study analyzes community support in the island economy of Mauritius by testing a model based on the social exchange theory and the identity theory. The model proposes that the resource-based occupational identity, environmental identity, and gender identity of the residents influence attitudes to tourism impacts and support (behavior). Results indicate that one’s identity has a direct bearing on support, but may not always influ- ence attitudes. Findings confirm the relevance of the social exchange theory and the identity theory in explaining community support for tourism in island economies. The study’s prac- tical implications and limitations are discussed. Keywords: social exchange theory, identity theory, community support, island economy. Ó 2011 Elsevier Ltd. All rights reserved. INTRODUCTION Over the past years, host communities’ attitudes to tourism impacts have attracted the attention of several researchers resulting in numer- ous studies on the topic (e.g. Andereck & Nyaupane, 2011; Gursoy, Chi, & Dyer, 2010; Lee, Kang, Long, & Reisinger, 2010). The prolifer- ation of such studies confirms the importance of understanding resi- dents’ responses and their involvement in tourism to achieve sustainability in any destination. Scholars disagree with the pre-suppo- sition that sustainable tourism development issues are similar from place to place (Bramwell & Sharman, 1999). Researchers argue that such assumption is incompatible with progressive thinking in sustain- ability science and sustainable tourism development. Studies suggest that principles of ‘mutuality’ and ‘locality’ should be applied in the planning process rather than stereo-type approaches which tend to de- grade the social networks and the distinctiveness of a destination (Potts Robin Nunkoo (BA, MA, MA, MPhil) is a doctoral candidate in the Department of Recreation and Leisure Studies, University of Waterloo (200 University Ave. W. Waterloo ON N2L 3G1. Canada. Email <[email protected]>). He is also a Senior Lecturer in the Department of Management, University of Mauritius. He conducts research on community impacts of tourism. Dogan Gursoy, PhD is the Taco Bell Distinguished Professor at the School of Hospitality Business Management, Washington State University, USA. He has research interests in tourism impacts, mega events, consumer behavior and tourism marketing. Annals of Tourism Research, Vol. 39, No. 1, pp. 243–268, 2012 0160-7383/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. Printed in Great Britain doi:10.1016/j.annals.2011.05.006 www.elsevier.com/locate/atoures 243

Residents’ support for tourism: An Identity Perspective

Embed Size (px)

Citation preview

Annals of Tourism Research, Vol. 39, No. 1, pp. 243–268, 20120160-7383/$ - see front matter � 2011 Elsevier Ltd. All rights reserved.

Printed in Great Britain

doi:10.1016/j.annals.2011.05.006www.elsevier.com/locate/atoures

RESIDENTS’ SUPPORT FOR TOURISMAn Identity Perspective

Robin NunkooUniversity of Waterloo, Canada

Dogan GursoyWashington State University, USA

Abstract: Studies on residents’ support for tourism have primarily been conducted in thedeveloped world. This study analyzes community support in the island economy of Mauritiusby testing a model based on the social exchange theory and the identity theory. The modelproposes that the resource-based occupational identity, environmental identity, and genderidentity of the residents influence attitudes to tourism impacts and support (behavior).Results indicate that one’s identity has a direct bearing on support, but may not always influ-ence attitudes. Findings confirm the relevance of the social exchange theory and the identitytheory in explaining community support for tourism in island economies. The study’s prac-tical implications and limitations are discussed. Keywords: social exchange theory, identitytheory, community support, island economy. � 2011 Elsevier Ltd. All rights reserved.

INTRODUCTION

Over the past years, host communities’ attitudes to tourism impactshave attracted the attention of several researchers resulting in numer-ous studies on the topic (e.g. Andereck & Nyaupane, 2011; Gursoy,Chi, & Dyer, 2010; Lee, Kang, Long, & Reisinger, 2010). The prolifer-ation of such studies confirms the importance of understanding resi-dents’ responses and their involvement in tourism to achievesustainability in any destination. Scholars disagree with the pre-suppo-sition that sustainable tourism development issues are similar fromplace to place (Bramwell & Sharman, 1999). Researchers argue thatsuch assumption is incompatible with progressive thinking in sustain-ability science and sustainable tourism development. Studies suggestthat principles of ‘mutuality’ and ‘locality’ should be applied in theplanning process rather than stereo-type approaches which tend to de-grade the social networks and the distinctiveness of a destination (Potts

Robin Nunkoo (BA, MA, MA, MPhil) is a doctoral candidate in the Department ofRecreation and Leisure Studies, University of Waterloo (200 University Ave. W. Waterloo ONN2L 3G1. Canada. Email <[email protected]>). He is also a Senior Lecturer in theDepartment of Management, University of Mauritius. He conducts research on communityimpacts of tourism. Dogan Gursoy, PhD is the Taco Bell Distinguished Professor at the Schoolof Hospitality Business Management, Washington State University, USA. He has researchinterests in tourism impacts, mega events, consumer behavior and tourism marketing.

243

244 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

& Harril, 1998). These arguments indicate the importance of placespecific policies and raise the need for destinations to identify the un-ique factors which are responsible for change in the local context(Twining-Ward & Butler, 2002).

Despite the importance of destination specific policies (Laws, Faulk-ner, & Moscardo, 1998), the majority of studies on residents’ attitudesand support for tourism have been conducted on communities fromdeveloped and industrialized countries (Nepal, 2008; Sirakaya, Teye,& Sonmez, 2002) although a few qualitative studies have been carriedout in the developing world. However, very few scholars studied com-munity support for tourism in small islands developing economies(Andriotis, 2005) despite the fact that some researchers argue thatthe determinants of residents’ support for tourism are likely to be dif-ferent between the developed and developing regions of the world(Sirakaya et al., 2002). The February 1993 United Nations GeneralAssembly Resolution 47/186 noted that small islands are not only con-fronted by factors common to the developing economies, but they arealso characterized by more specific factors. Their distinctive character-istics of insularity, vulnerability, and fragility in environment (Brigu-glio, 1995; Douglas, 2006) mean that they face significant challengesin terms of sustainable tourism (Bardolet & Sheldon, 2008; Twining-Ward & Butler, 2002). As a result of the unique characteristics of is-lands, sustainable development of tourism in such economies is closelylinked to local community involvement (Lim & Cooper, 2009).

Though numerous studies have been conducted on community sup-port for tourism, conceptual ideas, findings, and policy implicationsemerging from existing research may not be applicable to island desti-nations because of their uniqueness in several respects. In an attemptto fill this gap, this study tests a model of community support for tour-ism in Mauritius, an island economy (Figure 1). Using the Social Ex-change Theory (SET), the model postulates that attitudes toward thepositive and negative impacts of tourism influence support for theindustry, the ultimate dependent construct in the model. Drawingfrom the identity theory which provides a conceptual framework estab-lishing the link between identity, attitude, and behavior, the model alsoproposes that the resource-based occupational identity, environmentalidentity, and gender identity of the island residents are the indepen-dent variables influencing attitudes toward the positive and negativeimpacts of tourism as well as support for the industry. These three iden-tity constructs were chosen as being the likely determinants of residentattitudes for tourism because literature (e.g. Douglas, 2006; Petrzelka,Krannich, & Brehm, 2006; Petrzelka, Krannich, Brehm, & Trentelman,2005) suggests that these can be viewed as important concepts for is-land economies and can be relevant to the study of community supportfor island tourism.

Many previous studies (e.g. Andereck & Vogt, 2000; Huh & Vogt,2008) have conceptualized residents’ attitudes to tourism by measuringtheir attitudes to the positive and negative impacts of the industry.Although residents’ support for tourism is also often understood asan attitude to tourism (Gursoy, Jurowski, & Uysal, 2002), some

Resource-based occupational

identity

Environmental identity

Gender identity

Attitudes to positive impacts

Attitudes to negative impacts

Support for island tourism

Figure 1. The Proposed Model of the Study

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 245

researchers (e.g. Jackson & Inbarakan, 2006; Mackay & Campbell,2004; Nunkoo & Ramkissoon, 2010a) have considered residents’ sup-port/opposition for tourism as behavioral intent or behaviors residentsexhibit toward tourism. This is based on the premise that residents’ acton their attitudes toward the perceived impacts of tourism by support-ing/opposing the industry. The present study continues along a similarline of research and operationalizes residents’ attitudes to tourism interms of their attitudes to the positive and negative impacts of theindustry and considers the residents’ level of support for tourism as aform of behavior. Such a conceptualization of attitude and behaviormakes the identity theory a suitable framework for understandingthe influence of residents’ identities on their attitudes to tourismand support for the industry. The main theoretical contribution ofthe study results from the use of the identity theory, which has to-date,remained an underutilized theoretical framework for investigatingcommunity support for tourism. While several studies focus on the im-pacts of tourism development on community identities, very limited re-search analyzes the influence of identities on tourism (Ballesteros &Ramirez, 2007). By developing and testing a model based onjoint use of the SET and the identity theory in an island state, the studyalso contributes to the comparatively limited number of studies con-ducted in the developing world, and more specifically in islandeconomies.

246 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

RESIDENTS’ SUPPORT: AN IDENTITY PERSPECTIVE

Many previous studies that examine locals’ attitudes utilize the SETas their theoretical base (e.g. Lee et al., 2010; Nunkoo & Ramkissoon,2010b). SET postulates that an individual is likely to enter an exchangeif the resulting rewards are valued, the exchange is likely to produceappreciated rewards, and the perceived costs resulting from the ex-change do not exceed the benefits derived from the exchange (Skid-more, 1975). From a tourism perspective, residents who consider theindustry as valuable and believe that the costs do not exceed the ben-efits will be inclined to enter the exchange and consequently supportthe industry (Ap, 1992). Studies using SET test the relationships be-tween the perceived positive impacts and support and perceived nega-tive impacts and support (Nunkoo & Ramkissoon, 2010b; Vargas-Sanchez, Plaza-Mejia, & Porras-Bueno, 2009). These studies suggestthat positive attitudes to tourism are usually accompanied by higher le-vel of support for the industry while negative attitudes leads to lowersupport. Though SET remains one of the most acceptable frameworksto study community support (Gursoy & Rutherford, 2004), its predic-tive power is questioned. While some studies support the theory’s pos-tulates (e.g. Gursoy & Rutherford, 2004; Gursoy et al., 2002), othershave not been conclusive (e.g. Andereck, Valentine, Knopf, & Vogt,2005). The mixed support for SET is an indication that attitudes to-ward tourism impacts are not sufficient in explaining residents’ sup-port (behavior).

Studies based on attitude theories to explain behaviors omit an expli-cit reference to self identity (Biddle, Bank, & Slavings, 1987; Sparks &Shepherd, 1992). An identity is ‘‘a set of meanings attached to the selfthat serves as a standard or reference that guides behavior in situations’’(Stets & Biga, 2003, p. 401). In an attempt to understand one’s behavior,it is important to give due consideration to the identities that a personclaims and the corresponding meanings of these identities (Stets & Biga,2003). Previous studies suggest that a person’s identity influences his/her behaviors (Hagger, Anderson, Kyriakaki, & Darkings, 2007; Man-netti, Peirro, & Livi, 2004). The identity-behavior relationship can beunderstood within the context of the identity theory (Stryker, 1968,1980). Identity affects behavior by serving as a source of informationwhen an individual makes plans to act (Hagger et al., 2007). When behav-iors are consistent with identity, self-verification exists (Burke & Stets,1999). When there is a lack of self-verification, that is, when an individualdoes not engage in behaviors consistent with their identities, this createsa state of internal conflict between identity and actions (Callero, 1985).In this case, the identity system alters the behavior to counteract the sit-uational disturbances in an attempt to restore perceptions to match thebehavior with the individual’s identity (Stets & Biga, 2003). While atti-tude-based theories focus on how individuals make decision and choicesregarding a specific object or in a specific situation, identity theory notonly focuses on specific choices, but it also explains how individualswho are multifaceted and are embedded in the social structures makechoices and behave (Stets & Biga, 2003).

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 247

An individual possesses many identities because there are several so-cial relationships and roles which are connected to them (by virtue oftheir commitment to them) in the social structure (Stryker & Burke,2000). More salient and prominent identities influence behaviors to alarger extent than less salient and prominent ones. Identity theory pro-vides an important link between individuals to the larger social structure,a conceptualization which is missing in attitude theory (Stets & Biga,2003). Research indicates that while an individual’s identity influencesbehavior, such effects are independent of the influence of attitudes onbehavior (Biddle et al., 1987; Hagger et al., 2007). This reflects occasionswhen individuals engage in behavior because it is consistent with theirsense of identity (Hagger et al., 2007). Taking this into account, ourmodel proposes that the resource-based occupational identity, environ-mental identity, and gender identity of the residents influence their sup-port (behavior) directly, reflecting a spontaneous route, as well as a moredeliberate route via the mediation of perceived positive and negative im-pacts (attitude). The model implies that one’s attitude may not always beimplicated in support and may not always mediate the effects of identityon support. Attitudes may not always be influential in the process bywhich the different facets of a person’s identity affect his/her support.Identity theory may be considered as a more general theory than atti-tude-based ones. It has the potential of explaining a wide array of behav-iors across situations, including a host’s support for tourism.

Hypothetical Constructs

Residents’ Attitudes to Positive Impacts. Economic impacts of tourism arethe most valued elements for the host community (Liu, Sheldon, &Var, 1987). Tourism improves the local economy (Gursoy & Ruther-ford, 2004; Perdue, Long, & Allen, 1990), contributes to income andstandard of living (Liu & Var, 1986; Milman & Pizam, 1988), bringsin new businesses, and creates investment opportunities (Dyer, Gursoy,Sharma, & Carter, 2007; Kwan & McCartney, 2005). Tourism may alsoenrich community fabrics, preserve cultural values, and improve self es-teem (Andereck et al., 2005; Stronza & Gordillo, 2008). It also im-proves quality of life of the residents (Milman & Pizam, 1988; Perdueet al., 1990), creates new opportunities and instigates social change(Harrison, 1992), creates flexible working patterns and new opportuni-ties for females (Crompton & Sanderson, 1990), improves quality offire protection (Milman & Pizam, 1988; Pizam, 1978) and improvesthe quality of security such as police protection (Pizam, 1978). Studiesgenerally report a positive relationship between positive impacts andsupport (Gursoy & Rutherford, 2004; Lee et al., 2010; Nunkoo & Ram-kissoon, 2010b). Based on the postulates of the SET and the empiricaldiscussion from the literature, the following hypothesis is developed:

Hypothesis 1a. There is a direct positive relationship between residents’attitudes to the positive impacts of tourism and their support for theindustry.

248 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

Residents’ Attitudes to Negative Impacts. Tourism contributes to increasein cost of living (Liu & Var, 1986; Perdue et al., 1990), increases theprice of land and housing (Belisle & Hoy, 1980; Tovar & Lockwood,2008) and leads to a lack of economic diversification (Jackson & Inbar-akan, 2006). Other studies report that residents view social and culturalimpacts of tourism negatively (Ap & Crompton, 1998; Pizam, 1978).Tourism leads to prostitution (Liu & Var, 1986; Liu et al., 1987), trafficcongestion, crowding (Andereck et al., 2005; Dyer et al., 2007), smug-gling (Belisle & Hoy, 1980; Milman & Pizam, 1988), heightened ten-sion and crime (Andereck et al., 2005), and psychological tension(Andereck et al., 2005). Tourism also creates pollution and litter (Dyeret al., 2007), and destroys cultural and historical resources (Nepal,2008). Some of the previous research indicates that higher perceptionsof the negative impacts lead to lower support for tourism (Gursoy &Rutherford, 2004; Gursoy et al., 2010; Nunkoo & Ramkissoon,2010b). Based on the prior theoretical discussion on the SET andthe empirical evidences, the following hypothesis is proposed:

Hypothesis 1b. There is a direct negative relationship between residents’attitudes to the negative impacts of tourism and their support for theindustry.

Resource-Based Occupational Identity. Carroll and Lee (1990) defineoccupational identity as one in which members’ sense of identity is clo-sely linked to their occupation. Resource-based occupational identity isconsidered to be a factor influencing residents’ support for tourism(Petrzelka et al., 2006). Tourism impacts negatively on the occupa-tional distribution by sector and a community’s traditional employ-ment patterns (Haralambopoulous & Pizam, 1996). As a resultresidents often oppose tourism as an economic development strategy(Petrzelka et al., 2006; Smith & Krannich, 1998). Locals consider tour-ism as disruptive and a threat to local cultures if the industry leads todecrease in traditional sectors in a region (Lindberg & Johnson,1997). Mbaiwa’s (2011) study reveals how modernization caused bytourism contributes to livelihood insecurity. Previous studies indicatethat residents who work in resource-based industries and who hold ahigh level of occupational identity oppose tourism as a developmentstrategy (Petrzelka et al., 2006). Several other studies indicate that res-idents’ reliance on resource-based industries is associated with resis-tance toward tourism (e.g. Haukeland, 1984; Johnson, Snepenger, &Akis, 1994; Mason & Cheyne, 2000).

In general, residents oppose tourism if they feel that their identitiesare being threatened by such development (Mason & Cheyne, 2000).Many small islands traditionally relied on primary sectors such as agri-culture, farming, and fishing. As a result of economic restructuring, in-creased focus is now being placed on tourism, leading to the decline oftraditional industries in many of such economies. Such economicchanges shake the foundations of residents whose values and attitudes

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 249

are often deeply embedded with the occupational identity associatedwith resource-based industries (Reed, 2003). The implication of thisis that the resource-based occupational identity of residents can be adeterminant of their support for tourism development. However, therelationship between resource-based occupational identity and supporthas not been systematically documented though the literature indi-cates that residents with high resource-based occupational identity willbe more likely to oppose tourism than those with a low resource-basedoccupational identity. Based on the theoretical and the empirical dis-cussion from the literature, the following hypotheses are proposed:

Hypothesis 2a. There is a direct negative relationship between resource-based occupational identity and residents’ attitudes to the positive impactsof tourism.

Hypothesis 2b. There is a direct positive relationship between resource-based occupational identity and residents’ attitudes to the negativeimpacts of tourism.

Hypothesis 2c. There is a direct negative relationship between resource-based occupational identity and residents’ support for tourism.

Environmental Identity. Environmental identity is defined as an ‘‘expe-rienced social standing of who we are in relation to, and how we inter-act with the natural environment’’ (Weigert, 1997, p. 159). Itconstitutes of a set of meanings attached to an individual as he/sheinteracts with the natural environment (Stets & Biga, 2003). Strongerenvironmental identity is associated with more positive attitudes towardthe environment (Hinds & Sparks, 2008). Previous research indicatesthat one explanation for variance in residents’ support is environmen-tal attitudes (Gursoy & Rutherford, 2004; Gursoy et al., 2002). Nunkooand Ramkissoon (2010b) reported a direct negative relationship be-tween ecocentric attitudes and perceived benefits and a positive rela-tionship between ecocentric attitudes and perceived costs. Ingeneral, residents with ecocentric attitudes support policies that canlead to preservation of resources and the environment, while thoseholding anthropocentric attitude favor transforming the environmentto fulfill human needs (Jones, Jurowski, & Uysal, 2000).

Environmental identity has a particular meaning and significance forsmall island communities. The majority of people in the developingworld are close to the environment on a day-to-day basis (Franks,1996). This is very applicable to residents of small islands where theirdaily interaction with land, water, plants, and animals are an importantpart of their survival strategies (Douglas, 2006). Consequently, themanagement of resources impacts directly on the quality of life andlivelihoods of islands’ communities (Douglas, 2006). Bramwell’s

250 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

(2003) study indicates residents’ rising concerns for the environmentin Malta and how the level of environmentalism influences their atti-tudes to tourism. The research indicates that Maltese are reluctant toaccept loss of resources as a result of tourism. This discussion rein-forces the need to understand the relationship between environmentalidentity and tourism impacts in small islands. Such economies areincreasingly becoming dependent on tourism which is known to con-sume a high level of resources and put considerable pressure on them.Use and consumption of resources influence the social developmentand the governance responses of island residents (Douglas, 2006).The conceptual and empirical discussion led to the development ofthe following hypotheses:

Hypothesis 3a. There is a direct negative relationship between environ-mental identity and residents’ attitudes to the positive impacts of tourism.

Hypothesis 3b. There is a direct positive relationship between environ-mental identity and residents’ attitudes to the negative impacts of tourism.

Hypothesis 3c. There is a direct negative relationship between environ-mental identity and residents’ support for tourism.

Gender Identity. Some studies indicate that gender is a determinant ofresidents’ attitudes to tourism impacts (e.g. Huh & Vogt, 2008; Mason& Cheyne, 2000). However, Mason and Cheyne (2000) argue that gen-der differences in attitudes to tourism should be investigated further.Chhabra (2007, p. 145) further notes that ‘‘although a plethora of stud-ies have appeared in tourism literature that explore tourism impact per-ceptions of local residents based on sex, recent years have indicated aneed to delve deeper into gender based explorations and identify factorsother than biological that drive different or similar behaviors amongmen and women.’’ So far, researchers have approached the study of gen-der and tourism impacts uniquely from a biological sex perspective, dis-tinguishing between males and females. However, attitudes may have lessto do with being male or female, but more to do with the meanings that aperson attributes to himself/herself as masculine or feminine—his/hergender identity. Gender identity is the essential psychological attributesdistinguishing males from females and includes the characteristics asso-ciated with maleness or masculinity and femaleness or femininity that ex-ist in each individual (Bem, 1981). Gender identity is independent of anindividual’s biological sex and accordingly, cross-sex-typed individualsmay exist with varying levels of masculine or feminine traits (Martin &Gnoth, 2009).

Masculinity is agency-focused, emphasizing on competition, andindependence, while characteristics associated with femininity includecommunion-oriented, sensitivity and concern for others (e.g. about

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 251

tourism’s negative socio-cultural, environmental, and economic im-pacts). In support of the identity theory which postulates that individualsengage in behaviors which are consistent with their identity, studies con-firm that gender identity is a good predictor of behavior (Fisher & Ar-nold, 1990). Research suggests that individuals with feminine identitycare more for society, are more likely to engage in environmentalresponsive behaviors, and are more aware of the consequences of envi-ronmental conditions (Spence & Helmreich, 1978; Stets & Biga,2003). Caring is considered to be an orientation of feminine identity be-cause women’s role in society (e.g. domestic workers, primary caretak-ers) promotes the welfare of people (Cancian & Oliker, 2000). In linewith the identity theory and the empirical evidences, it is reasonable toextrapolate that residents with higher feminine traits are more likelyto view tourism impacts negatively and be less supportive of developmentthan those with lower feminine traits. Therefore, effects of residents’gender identity on their attitudes to tourism impacts should be investi-gated further, particularly in small islands where special attention needsto be given to population issues such as gender (United Nations, 1994).Accordingly, the following hypotheses are proposed:

Hypothesis 4a. Residents’ level of femininity is negatively related toattitudes to the positive impacts of tourism.

Hypothesis 4b. Residents’ level of femininity is positively related to atti-tudes to the negative impacts of tourism.

Hypothesis 4c. Residents’ level of femininity is negatively related to sup-port for tourism.

Research Methods

Study Site and Context. Data were collected from the residents of Mau-ritius. The economy has made considerable progress, transforming it-self from a low-income country to a middle income economy relyingmainly on the production and export of agricultural products, textile,and tourism. The country’s progress was made possible by a number ofstructural and macroeconomic policies and preferential access to theEuropean Union market under the sugar protocol, and to the worldmarkets under the Multifibre Agreement. However, these agreementshave come to an end, requiring the government to rethink its eco-nomic priorities. The state has renewed its interest in the developmentand expansion of tourism. The island depends on up-market and all-inclusive forms of tourism and received a total of 930, 456 tourists in2008. One of the main objectives of the government is to increase tour-ist arrivals to two million by the year 2015. Construction of integratedresorts, development of the cultural and heritage sector, and thecreation of tourist villages are the product diversification strategies

252 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

being implemented. Movement toward mass tourism (in the context oftwo million tourist arrivals) is also clearly visible. The current tourismpolicies promoting all inclusive development largely exclude the localpopulation from the industry. The transition of Mauritius from aneconomy traditionally based on agriculture and the primary sector toone based on tourism makes the island an ideal case study for the pres-ent research.

Survey Method and Sample. The sample included residents who were atleast 18 years of age or older, who had the most recent birthday, andwas a permanent resident of the household. This procedure is consid-ered to be a nonintrusive method for randomly selecting individualrespondents (Petrzelka et al., 2006). A stratified random samplingwas utilized to determine the number of respondents required fromeach village and town of Mauritius. Afterwards, a random samplingwas used to select the pre-determined number of respondents fromeach region. With the help of street maps, every third house in everythird street in each region was selected for the interview. This methodensured the random selection of respondents from the study popula-tion. A self administered survey questionnaire was used for data collec-tion. The questionnaire was delivered to 360 residents using the drop-off/pick up method (Steele et al., 2001). Thirty-two questionnaireswere unreturned, leading to 328 responses. However, since missing val-ues can lead to biased statistical results (Hair, Anderson, Tathman, &Black, 1998), questionnaires with a total of five or more unanswereditems across the measurement scales were eliminated from the analysis(28 questionnaires). After elimination, a total of 300 questionnaireswere retained for further analysis.

Measurement of Variables. Five items were used to measure residents’ le-vel of resource-based occupational identity. These items were adoptedfrom Petrzelka et al. (2005) and Petrzelka et al. (2006) and intended tocapture residents’ patterns of involvement with primary/resourcebased organizations in Mauritius. Residents were asked to state their le-vel of involvement with these organizations on a 1–5 Likert scale, where1 represented ‘not at all involved’ and 5 represented ‘very highly in-volved’. Petrzelka et al. (2005, p. 1129) consider that such a measure-ment scale provides a ‘‘reasonable proxy for getting at one facet ofoccupational identity.’’ Items used to measure environmental identitywere borrowed from Stets and Biga (2003) and consisted of elevenbipolar statements. Respondents were asked to think about how theyview themselves with respect to the environment, identifying wherethey would place themselves between each bipolar statement referenc-ing the natural environment. Responses ranged from 1 to 5 where 1represented agreement with one bipolar statement and 5 reflectedagreement with the other extreme, and 3 indicated that the respon-dents is in between the two statements. A high score indicated anenvironmental friendly or ecocentric attitude. While answering

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 253

these questions, the focus was the person rather than the positionwhich the individual occupies in society (Stets & Biga, 2003).

Gender identity was measured using 24 items borrowed from the Per-sonal Attribute Questionnaire (Spence & Helmreich, 1978). PersonalAttribute Questionnaire is one of the most widely used measures tocapture the meanings of masculinity and femininity in society (Stets& Burke, 2000). The instrument employs a set of bio-polar scales tocapture gender identity and lists a series of attributes that are positivelyvalued for both sexes but are more normative for either males or fe-males to endorse (Stets & Biga, 2003). Respondents rated themselveson these bio-polar scales. Responses range from 1 to 5 where 1 pre-sented agreement with one bipolar statement and 5 reflected agree-ment with the other extreme and 3 indicated that the respondents isin between the two statements. A high score represented femininitytraits while a low score represented masculinity traits. Attitudes tothe positive (12 items) and negative impacts (12 items) were measuredusing scales adopted from previous tourism impact studies (Dyer et al.,2007; Gursoy & Rutherford, 2004; Gursoy et al., 2002; Liu & Var, 1986).The items were measured on a 1–5 Likert scale where 1 represented‘strongly disagree’ and 5 represented ‘strongly agree’. These itemswere chosen because they fit the specific characteristics of tourismdevelopment in Mauritius and they reflect the standard environmental,socio-cultural, and economic impacts of tourism as espoused in the lit-erature. Support for tourism was measured by 5 items asking residentstheir level of support with specific forms of tourism development on a1–5 Likert scale where 1 represented ‘strongly oppose’ and 5 repre-sented ‘strongly support’. The first two items used were adopted fromGursoy and Rutherford (2004) and Gursoy et al. (2002) and Yoon, Gur-soy, and Chen (2001). Three additional items were added as they wereother types of tourism development specific to Mauritius at the timethe research was conducted. Operationalizing ‘support for tourism’by asking residents their level of support/opposition for different typesof tourism development is considered to be an ideal way to measurethe construct (Andereck & Vogt, 2000).

Scale Refinement. The content and validity of the items were first as-sessed by a number of tourism professors who were asked to providefeedback on the content and understandability of each measurementitem. The survey instrument was also sent to a number of tourism prac-titioners working in the Mauritian tourism industry. They were asked tocomment on the relevance of the items to Mauritius. In both cases, sug-gestions regarding the phrasing of statements were made. These weretaken into account and the questionnaire was refined and pretestedusing a convenience sample of 120 residents. The purpose of the pre-test was to establish a uni-dimensional scale for the measurement of theconstructs. To detect scale dimensionality, an exploratory factor analy-sis with principal component method was conducted for eachconstruct. In order to ensure that each factor identified by the explor-atory factor analysis had only one dimension and that each attribute

254 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

loaded only on one factor, attributes that had factor loadings of lowerthan 0.40 and attributes loading on more than one factor with a load-ing score of equal to or greater than 0.40 on each factor were elimi-nated from the analysis (Chen & Hsu, 2001). The items whichremained after this process and the reliability coefficient for each con-struct are given in Table 1. All Cronbach’s alpha values exceeded theacceptable level of 0.70.

Table 1. Items and Cronbach’s Alpha Values

Construct and items Cronbach’s Alpha

Resource Based Occupational Identity: (OCI) 0.72

Agricultural production organizations

Sugar Investment Trust

Small planters associations

Farmers’ service corporations

Environmental Identity: (ENI) 0.73

Not at all protective of the natural environment . . . very protective of the natural environment

Not respectful toward the natural environment . . . very respectful toward the environment

Indifferent about the natural environment . . . very concerned about the natural environment

Not at all passionate toward the environment . . . very passionate toward the environment

Disinterested in the natural environment . . . an advocate of the natural environment

Detached from the natural environment . . . connected to the natural environment

Independent from the natural environment . . . dependent on the natural environment

Gender Identity: (GNI) 0.73

Can make decisions easily . . . Has difficulty making decisions

Not at all emotional . . . Very emotional

Not at all able to devote self completely to others . . . Able to devote self completely to others

Feelings not easily hurt . . . Feelings easily hurt

Very rough . . . Very gentle

Not at all helpful to others . . . Very helpful to others

Very little need for security . . . Very strong need for security

Attitudes to Positive Impacts: (API) 0.85

Cultural exchange

Development of a variety of cultural activities

Positive impact on the cultural identity of Mauritius

Conservation of natural resources

Restoration of historic buildings

More jobs

More business for local people

More investment

Attitudes to Negative Impacts: (ANI) 0.93

Increase in vandalism

Traffic congestion

Noise and air pollution

Crime rate

Destruction of the natural environment

Increase in prostitution

Crowded beaches and other outdoor places

Residents are likely to suffer from living in a tourist destination

Change the Mauritian culture

Support: (SFT) 0.82

Small businesses

Cultural and historic based attractions

All inclusive form of tourism development

Expansion in tourist arrivals to 2 million tourists

Integrated resort projects

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 255

Data Analysis and Assumption Checks. Stepwise multiple regression anal-ysis was used to test the proposed model and hypothesized relation-ships. This data analysis technique was chosen because the identityvariables presented in the models are exploratory in nature as theyhave not been systematically tested in previous research. Multipleregression analysis is considered useful for investigating exploratoryvariables. Stepwise multiple regression analysis also eliminates any var-iable already in the prediction equation at each step if doing so doesnot result in a significant loss of predictive power (Diekhoff, 1992).The outcome is a readily interpretable model containing the mostimportant predictors of the dependent variables. Consequently, multi-ple regression analysis was considered to be a suitable technique to testthe hypothesized relationships. The following assumptions were testedto ascertain that the data is suitable for multiple regression analysis:normality and multicollinearity. In order to assess the normality ofthe distribution of the data, the skewness of each construct was exam-ined. The acceptable range of values for skewness falls between �1 and+1 (Hair et al., 1998). Results indicated that all values fell within thisrange. Multicollinearity was verified by analyzing the tolerance and Var-iance Inflation Factor (VIF) values for each predictor variable in themodel. The Variance Inflation Factor value is computed as 1/toler-ance. Only variables which have Variance Inflation Factor values belowthe threshold of 10 are to be retained in the analysis (Ho, 2006; Pedha-zur, 1997). Results indicated that all values fell within below the recom-mended threshold of 10 (Table 2).

Results from a multiple regression analysis indicate the explanatorypower of all predictor variables with measures of R2 and adjusted R2 to-gether with the relative importance of each individual predictor vari-able after calculating the b coefficients (Musil, Jones, & Warner,1998). The multiple regression coefficient R is an indication of the cor-relation between the weighted sum of the predictor variables and thecriterion variable (Kachigan, 1986). R2 is the square of this measureof correlation and indicates the proportion of the variance in the

Table 2. Regression Analysis (n = 300)

Step Predictor R R2 Adjusted R2 Increment

to R2

F change Final b t Values VIF Tolerance

Stepwise Regression of Support for Tourism

1 API .622 .387 .385 .387 188.101* .586* 12.075 .711 1.407

2 ENI .648 .420 .416 .033 17.137* �.286* �5.852 .699 1.430

3 OCI .688 .473 .467 .052 29.432* �.298* �6.069 .696 1.438

4 ANI .707 .501 .494 .028 16.365* �.207* �4.398 .752 1.329

5 GNI .713 .508 .500 .008 4.662** �.107** �2.159 .677 1.476

Stepwise Regression of Attitudes to Positive Impacts

1 OCI .462 .213 .211 .213 80.796* �.462* �8.989 1.000 1.000

Stepwise Regression of Attitudes to Negative Impacts

1 OCI .379 .144 .141 .144 49.981* .337* 6.146 .932 1.074

2 GNI .409 .167 .162 .024 8.489** .160** 2.914 .932 1.074

* p < 0.001.; ** p < 0.05.;

256 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

criterion variable which is accounted for by the model. The model pre-diction accuracy is measured by the value of the adjusted R2, which isexpressed as a percentage. Other components of the multiple regres-sion analysis are the F-test and the t-test. The F-test indicates thestrength of the regression model while the t-test assesses whether theindependent variables predict the dependent variables.

RESULTS

Predicting Support

Results indicated that all the predictor variables (attitudes to the positiveimpacts, attitudes to the negative impacts, occupational identity, environ-mental identity, and gender identity) entered the equation, and resultedin an overall significant model, R2 = 0.51, p < 0.05 (Table 2). Attitudes tothe positive impacts accounted for 39% of the variance in support for tour-ism (F change = 188.10, p < 0.001). Entry of the environmental identity var-iable resulted in an R2 change of .033, increasing the explained variance insupport for tourism to 42% (F change = 17.137, p < 0.001). Finding sug-gested that the entry of the occupational identity variable contributed toan additional 5.2% to total variance (F change = 29.43, p < 0.001). Inclusionof the attitudes to the negative impacts construct contributed to an addi-tional 2.8% increase (F change = 16.365, p < 0.001), increasing total vari-ance in support to around 50%. Entry of the gender identity constructcontributed to an additional 0.8%, increasing total variance to 51% (Fchange = 4.662, p < 0.05).

Results indicated support for Hypothesis 1a which proposed a direct po-sitive relationship between residents’ attitudes to the positive impacts andsupport for tourism (b = .586, t = 12.075, p < .001) and Hypothesis 1b whichproposed a direct negative relationship between residents’ attitudes to thenegative impacts and support (b = �.207, t =�4.398, p < .001). Results alsoindicated a direct negative relationship between occupational identity andsupport, supporting Hypothesis 2c (b = �.298, t = �6.069, p < .001).Hypothesis 3c which proposed a direct negative relationship between resi-dents’ environmental identity and support was also supported (b = �.286,t =�5.852, p < .001). Hypothesis 4c which postulated a negative relation-ship between residents’ level of femininity and support for tourism was alsosupported (b = �.107, t = �2.159, p < .05).

Predicting Residents’ Attitudes to Positive Impacts

Occupational identity, environmental identity, and gender identitywere entered into the regression equation. Results indicated that onlythe occupational identity variable entered the equation while the envi-ronmental identity and gender identity variables were removed fromthe regression analysis based on the following criteria: Probability-of-F-to-enter 6.050, Probability-of-F-to-remove P.100. As presented inTable 2, results indicated that the occupational identity accountedfor around 21% variance in the dependent variable (R2 = 0.21,

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 257

p < 0.001). Results indicated support for Hypothesis 2a which postu-lated a direct negative relationship between resource-based occupa-tional identity and attitudes to the positive impacts (b = �.462t = �8.989, p < 0.001). Hypotheses 6 which postulated that there is a di-rect negative relationship between residents’ environmental identityand attitudes to the positive impacts and Hypothesis 4a which postu-lated that there is a direct negative relationship between residents’ le-vel of femininity and attitudes to the positive impacts were bothrejected (b = .042 t = .795, p > 0.05; b = .040, t = .745, p > 0.05), indicat-ing that environmental identity and gender identity are not significantpredictors of attitudes to the positive impacts.

Predicting Residents’ Attitudes to Negative Impacts

Stepwise multiple regression analysis was conducted to analyze theinfluence of the predictor variables (occupational identity, environ-mental identity, and gender identity) on the dependent variable (atti-tudes to the negative impacts). Results indicated that with theexception of the environmental identity variable, both the occupa-tional identity and the gender identity variables entered the predictionequation based on the following criteria: Probability-of-F-to-enter6.050, Probability-of-F-to-remove P.100, yielding a significant model,R2 = 0.17, p < 0.05 (Table 2). This indicated that environmental iden-tity was not a significant determinant of attitudes to the negative im-pacts. Findings showed that inclusion of the occupational identityvariable in the model accounted for 14.4% of the variance (Fchange = 49.98, p < 0.001). Entry of the gender identity variable re-sulted in an R2 change of .024, meaning that entry of the variable in-creased the explained variance in attitudes to the negative impactsby 2.4% to a total of around 17% (F change = 8.489, p < 0.05). Resultsindicated support for Hypothesis 2b which postulated that a direct po-sitive relationship exists between resource-based occupational identityand attitudes to the negative impacts (b = .337, t = 6.146, p < 0.001).Findings further suggested that Hypothesis 3b, which postulated thata direct negative relationship exists between environmental identityand attitudes to the negative impacts was not supported (b = .053t = .836, p > 0.05). Hypothesis 4b which proposed that a direct positiverelationship exists between level of femininity and attitudes to the neg-ative impacts of tourism was supported (b = .160, t = 2.914, p < 0.05).

Mediating Effects

Baron and Kenny’s (1986) method was used for mediation testing.Mediating effects can be tested if there is a significant relationship be-tween (1) the predictor variable and the dependent variable, (2) thepredictor variable and the mediating variable, and (3) the mediatorand the dependent variable (Baron & Kenny, 1986). Sobel z test(Sobel, 1986) was used to test the mediating effects of attitudes to

Table 3. Mediating Effects: Sobel Test Results

Test statistics p Values

Model 1: Attitudes to the positive impacts mediatesthe relationship between occupational identityand support for tourism

Sobel test 0.40 0.69

Model 2: Attitudes to the negative impacts mediatesthe relationship between occupational identityand support for tourism

Sobel test 4.23 0.00

Model 3: Attitudes to negative impacts mediatesthe relationship between gender identityand support for tourism

Sobel test �0.07 0.94

Note: Mediating effects were calculated only for those independent variables which had asignificant influence on the mediating variables.

258 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

the positive and negative impacts on the relationship between occupa-tional identity and support for tourism and the mediating effects ofattitudes to the negative impacts on the relationship between genderidentity and support for tourism. Sobel z test is considered to havegreater statistical power than other formal methods of assessing medi-ation. One assumption necessary for the Sobel z test is the normality ofthe data (Preacher & Hayes, 2004). As noted before, the skewness val-ues of the constructs indicated that our data meets this assumption.

Sobel z test is conducted by comparing the strength of the indirecteffect of the independent variable on the dependent variable (Preach-er & Hayes, 2004). Mediation effect is established if the indirect effectbetween the predictor variable and the dependent variable is signifi-cant by the Sobel z test (Baron & Kenny, 1986). Results from Table 3suggest that attitudes to the negative impacts mediates the relationshipbetween occupational identity and support for tourism (z = 4.23,p < 0.001), indicating that the indirect effect of occupational identityon support for tourism (behavior) is statistically significant. However,results indicated that attitudes to the positive impacts does not mediatethe relationship between occupational identity and support for tourism(z = 0.40, p > 0.05). Attitudes to the negative impacts was also not foundto mediate the relationship between gender identity and support fortourism (z = �0.07, p > 0.05). These results suggest that the influenceof one’s identity on support for tourism (behavior) is not always med-iated by one’s attitude toward tourism impacts.

DISCUSSION

As indicated in Table 4, the findings supported eight of the elevenoriginally proposed hypotheses. Hypothesis 1a which proposed a direct

Table 4. Result of Hypotheses Testing

Hypothesized relationship Standardized coefficients Results

H1a: API fi SFT (+ve) .586 SupportedH1b: ANI fi SFT (�ve) �.207 SupportedH2a: OCI fiAPI (�ve) �.462 SupportedH2b: OCI fi ANI (+ve) .337 SupportedH2c: OCI fi SFT (�ve) �.298 SupportedH3a: ENI fi API (�ve) .042 Not supportedH3b: ENI fi ANI (+ve) .053 Not supportedH3c: ENI fiSFT (�ve) �.286 SupportedH4a: GNI fi API (�ve) .040 Not supportedH4b: GNI fi ANI (+ve) .160 SupportedH4c: GNI fi SFT (�ve) �.107 Supported

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 259

positive relationship between attitudes to the positive impacts and sup-port was supported. This indicates that the more positively residentsviewed tourism, the higher was their support. These findings are in linewith the results of other studies (Dyer et al., 2007; Gursoy & Ruther-ford, 2004; Lee et al., 2010), but contradicts the results of Gursoyand Kendall’s (2006) and Deccio and Baloglu’s (2002) studies. Thiscan be explained by the fact that the latter two studies considered res-idents’ support for mega events, an obtrusive form of tourism, suggest-ing that a community’s support is dependent on the type ofdevelopment taking place. Results also suggest support for Hypothesis1b which proposed a direct negative relationship between the attitudesto the negative impacts and support for tourism. This finding is in linewith the results of previous studies, confirming that residents’ attitudesto the negative impacts of tourism negatively influence support for theindustry (Milman & Pizam, 1988; Nunkoo & Ramkissoon, 2010b; Shel-don & Var, 1984).

Hypotheses 2a that predicted a direct negative relationship betweenresource-based occupational identity and attitudes to the positive im-pacts, 2b that proposed a direct positive relationship between occupa-tional identity and attitudes to the negative impacts, and 2c whichpostulated a direct negative relationship between occupational identityand support have been supported. Findings indicate that residents whohold a high level of resource-based occupational identity are less likely toview tourism as having positive effects and are less supportive of develop-ment. Results also suggest that residents with a low level of resource-based occupational identity are likely to view tourism as having positiveimpacts and are more supportive of the industry. These findings are con-sistent with previous studies in the literature (Haukeland, 1984; Lind-berg & Johnson, 1997; Mason & Cheyne, 2000). Results also indicatethe relationship between occupational identity and support for tourismis mediated by the residents’ attitudes to the negative impacts of theindustry. The strength of the path relationships between resource-basedoccupational identity and attitudes to the positive impacts, attitudes tothe negative impacts and support for tourism indicates the importance

260 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

of this construct for understanding attitudes within the context of tour-ism development in island economies where residents’ identities areembedded in resource-based occupations.

Hypothesis 3a which proposed a direct negative relationship be-tween environmental identity and attitudes to the positive impactsand 3b which proposed a direct positive relationship between environ-mental identity and attitudes to the negative impacts were both re-jected, indicating that environmental identity is not a good predictorof attitudes to tourism. These findings are different from what the lit-erature has reported so far as they contradict those of Nunkoo andRamkissoon (2010b), Gursoy and Rutherford (2004), and Gursoyand Kendall (2006) whose results reveal that environmental attitudeis a good predictor of attitudes to the positive and negative impactsof tourism. Interestingly, while Hypotheses 3a and 3b have been re-jected, results indicated support for Hypothesis 3c which proposed thata direct negative relationship exists between environmental identityand support. Thus, environmental identity was found to be an inde-pendent predictor of one’s support (behavior), but not a determinantof one’s attitude. This contradictory finding pertaining to the hypoth-eses related to the environmental identity construct can be explainedby the postulates of the identity theory. While identity theory postulatesthat a person’s identity can influence his/her attitudes (toward tourismimpacts), such identities can also have a direct bearing on behavior(support) even if it does not influence one’s attitudes (Leary & Jones,1993; Mannetti et al., 2004).

Hypothesis 4a predicting a direct negative relationship between res-idents’ level of femininity and attitudes to the positive impacts was re-jected, indicating that the gender identity was not a significantpredictor of attitudes to the positive impacts of tourism. Hypothesis3b postulating a direct positive relationship between residents’ levelof femininity and attitudes to the negative impacts and 3c predictinga direct negative relationship between residents’ level of femininityand support (behavior) were both supported. These findings indicatethat residents displaying more female oriented traits were more likelyto view tourism impacts negatively and be supportive of developmentthan residents with masculine oriented characteristics. These findingscan be explained by the fact that masculinity is agency-focused, empha-sizing on competition and independence while femininity is commu-nion-oriented, emphasizing on sensitivity and concern for others(Spence & Helmreich, 1978). In line with this, previous studies indi-cate that individuals with femininity traits are more likely to expressgreater concern over environmental conditions and are more likelyto engage in pro-environmental behaviors (Stets & Biga, 2003). The re-sults support the identity theory’s postulates that individuals engage inbehaviors that are similar in meanings to the meanings of their identi-ties (Burke & Reitzes, 1981). The non-significant path between genderidentity and attitudes to the positive impacts, and the supportedhypothesis between the former and support for tourism confirms thatone’s gender identity can influence support for tourism (behavior),but not necessarily one’s attitudes toward the industry.

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 261

Policy Implications

The results can help tourism planners and businesses to better inte-grate the local communities in tourism development and understandareas of frustration and dissatisfaction. Tourism developers shouldunderstand the importance of residents’ identity-based concernsregarding tourism development. Residents should not be treated ascustomers who need to be convinced to endorse tourism, but rather,they should be considered as a group of important stakeholder whoseattitudes and interests are important for the sustainable developmentof tourism. Findings suggest that resource-based occupational identityinfluences residents’ attitudes to the positive and negative impacts aswell as support for tourism in the destination. Local planners shoulddevelop tourism in a way which is not disruptive to the island commu-nities’ occupational identities. Residents with high levels of occupa-tional identity were more likely to view tourism impacts negativelyand were less supportive of the industry. Tourism authorities and re-lated businesses should attempt to segment residents based on their le-vel of occupational identity. Such information can be collectedthrough survey research and focus group meetings with residents work-ing in resource-based and primary sectors. An internal marketing pro-gram can be conducted with the aim of making residents working inresource-based industries more supportive of development. Authoritiescan provide monetary and non-monetary compensations to residentsworking in resource-based sectors to gain their endorsement for futuredevelopment. Tourism development should also be based around suchresource-based occupational identities. Planners can focus on develop-ing those types of tourism which are the least disruptive to the islandcommunities’ occupational identity. Tourism planning should alsobe sensitive to the traditional economic system prevailing in islandeconomies.

Results also indicate that environmental identity is a good determi-nant of support. Residents with high environmental identity (ecocen-tric attitudes) were likely to be less supportive of developmentsuggesting that developers and investors should be very sensitive aboutthe environmental impacts of development. Policymakers should alsounderstand that developing tourism is highly resource-intensive andcould be a major challenge in island destinations where communitiesare usually highly dependent on the environment and natural re-sources. The high environmental dependence of island residentsmeans that authorities should also provide them with alternative op-tions for livelihoods if their survival options are threatened by tourismdevelopment. Any tourism related development in the island shouldalso incorporate projects for environmental improvement, conserva-tion, and restoration. Investors and developers should provide clearenvironmental management plans and environmental restoration pol-icies in case of environmental damage resulting from development.These may help to ease the concerns of residents displaying a highlevel of environmental identity. To be more effective, these plans

262 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

should be well publicized to the island communities who need to bemade aware of the same to ensure their full support.

Results suggest that gender identity is a good predictor of attitudes tothe negative impacts and support for tourism. Residents displaying fem-ininity traits were more likely to view tourism impacts negatively and wereless supportive of the industry. These findings imply that understandinghow biological gender differences influences attitudes to tourism im-pacts may not be enough to gain public support for development. Plan-ners should understand that gender is a complex variable andunderstanding how males and females residents react to tourism devel-opment may provide limited information for planning purposes. Plan-ners should gear their market research not only toward understandinghow biological gender differences (male and female) influence attitudesto tourism impacts, but they should also investigate psychological genderdifferences (gender identity) among residents of a community and howsuch differences influence support for tourism. It is also important torecognize that female residents may possess masculinity traits while maleresidents may have femininity traits and these might influence attitudesto tourism and support in a way different from biological sex. Plannerscan collect such information by engaging in survey research asking resi-dents about their gender identity.

Results also indicate that support was influenced by the residents’attitudes to the positive and negative impacts of tourism. Communityleaders and planners may consider conducting an educational pro-gram informing residents about the benefits of tourism to gain theirendorsement. Planners can identity residents having negative attitudestoward the industry and conduct an internal marketing program aim-ing at changing their attitudes favorably. Businesses and plannersshould also ensure that the industry’s negative impacts on the econ-omy, environment, and society are mitigated while the positive onesare enhanced. Tourism businesses can engage in corporate socialresponsibility programs aiming at community improvement, environ-mental enhancement, and improving the general welfare of the resi-dents. Attempts should be made to ensure community involvementin decision and policy making. Dissemination of the benefits or ex-pected benefits to the local residents may also contribute to gainingtheir support.

CONCLUSION

The majority of studies on community support have been conductedin the developed and industrialized countries leading to a paucity ofresearch in small island economies. As a result of their uniqueness, is-lands face several challenges for sustainable tourism. This study reaf-firms the importance for researchers, scholars, and practitioners tounderstand community support as a component of sustainable tourismdevelopment in small economies. It tested a community support modelbased on the SET and identity theory from data collected from resi-dents of the island of Mauritius. The model addressed the influence

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 263

of resource-based occupational identity, environmental identity, andgender identity on residents’ attitudes and support. Findings sup-ported eight of the eleven hypotheses proposed. However, as withany research, this study has certain limitations which should be takeninto account while interpreting the findings.

First, the small sample size of the study should be mentioned as alimitation of the research. All other things being equal, smaller sam-ples tend to have greater sampling error than larger samples. Thismakes it less likely that any statistically significant relationships willbe detected in the sample data. Issues relating to potential non-re-sponse bias should also be considered given the small sample size ofthe study. Non-response bias means that results of this study may notbe fully representative of everyone in the population of the study.These sampling issues may limit the generalizability of the findings.Therefore, findings of this study may not be generalized beyond thesample examined. Future studies with larger samples in different loca-tions are needed to validate the findings of this study. Secondly, thestudy considered the influence of identities on support but did notanalyze the effects of their prominence (McCall & Simmons, 1978)and salience (Stryker, 1980) on support. The prominence of an iden-tity reflects one’s ideal self and the salience refers to the probabilityof enacting (Stets & Biga, 2003). Inclusion of these identity relatedconstructs could have increased the predictive power of the model.

Thirdly, the research considered the consequences of tourism asbeing two-dimensional, comprising of positive and negative impactswhen in fact research suggests that tourism impacts are categorizedinto environmental, economic, social, and cultural (Pizam, 1978).Delineating the impacts into these categories could have enhancedour understanding of the relationships between one’s identities andsupport. Finally, the findings may also have limited external validityas the model has been tested using data collected from one island des-tination only. Scholars should attempt to test the model in other islanddestinations positioned a different stages in their life cycle and experi-encing different forms of development. In doing so, the relationshipsamong the constructs may become clearer. Future researchers can alsoconsider incorporating other dimensions of identity relevant to theirstudy area in the model.

Despite the above limitations, the study makes some useful theoret-ical contributions to the literature. Findings indicate that residents’decision to support development is based on the expected benefitsand costs of development (Vargas-Sanchez et al., 2009). The studynot only joins other research which supports the underlying principlesof SET (Gursoy et al., 2010; Lee et al., 2010), but its findings also indi-cate that the predictive power of SET in explaining residents’ supportis equally relevant for small island economies and not only to the devel-oped and industrialized world where the theory has been mostly tested.Another important contribution the study makes to the theoretical do-main relates to the use of the identity theory principles in explainingresidents’ attitudes and support. While one’s identities were found toinfluence attitude toward impacts, they exerted a higher influence

264 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

on support, confirming that self identity is a good determinant ofbehavior (Mannetti et al., 2004; Sparks, Shepherd, Wieringa, &Zimmermanns, 1995).

The theoretical implications emerging from the results of the medi-ating analysis further imply that attitudes toward impacts do not alwaysmediate the influence of identity on support (behavior) and are there-fore not always influential in the process by which identity influenceresidents’ behavior (Hagger et al., 2007). For further theoretical devel-opments in the field, it is therefore important to note that it is possiblethat residents’ identities significantly influence their support for tour-ism even if they do not influence their attitudes to tourism impacts.Researchers should also understand that support for tourism is notonly influenced by residents’ attitudes to tourism impacts, but alsoby more complex factors such as their identities. Findings suggest thatidentities serve as a guide for behaviors which serve to verify and main-tain the self-meanings associated with an identity. Thus, support fortourism may not only be influenced by residents’ attitudes to impacts,but also by their identities relevant to the context.

Another theoretical implication relate to the influence of genderidentity on support. Findings imply that understanding the influenceof biological sex (male and female) on support may provide research-ers with a restricted view and suggest that existing studies provide uswith a limited representation of different gendered approaches to res-idents’ attitudes to tourism. The study confirms other researchers’ (e.g.Fischer & Arnold, 1994) arguments that gender effects on behavior arenot only influence by biological sex differences, but also by psycholog-ical differences between genders. Thus, scholars should integrate thegender identity construct in their research to further our understand-ing on its relationship with support. Inclusion of identity factors mayimprove the prediction of host behavior because identity theory isbased on the assumption that individuals behave not on the basis ofdiscrete and personal decisions but on the basis that individuals areembedded in a social structure which influences their behavior. Thus,while SET remains important in understanding residents’ support fortourism, future researchers should also incorporate the postulates ofthe identity theory in their investigations on host’s attitudes to tourismand support. Joint use of the identity theory and the SET may provevaluable in understanding community support for tourism.

REFERENCES

Andereck, K. L., & Nyaupane, G. P. (2011). Exploring the nature of tourism andquality of life perceptions among residents. Journal of Travel Research, 50,248–260.

Andereck, K. L., & Vogt, C. A. (2000). The relationship between residents’attitudes toward tourism and tourism development options. Journal of TravelResearch, 39, 27–36.

Andereck, K. L., Valentine, K. M., Knopf, R. C., & Vogt, C. A. (2005). Residents’perceptions of community tourism impacts. Annals of Tourism Research, 32(4),1056–1076.

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 265

Andriotis, K. (2005). Community groups’ perceptions of and preferences fortourism development: Evidence from Crete. Journal of Hospitality and TourismResearch, 29, 67–90.

Ap, J. (1992). Residents’ perceptions on tourism impacts. Annals of TourismResearch, 19(4), 665–690.

Ap, J., & Crompton, J. (1998). Developing and testing a tourism impact scale.Journal of Travel Research, 37, 120–130.

Ballesteros, E. R., & Ramirez, M. H. (2007). Identity and community–Reflectionson the development of mining heritage tourism in Southern Spain. TourismManagement, 28, 677–687.

Bardolet, E., & Sheldon, P. J. (2008). Tourism in archipelagos: Hawaii and theBalearics. Annals of Tourism Research, 35(4), 900–923.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinctionin social psychological research: Conceptual, strategic, and statistical consid-erations. Journal of Personality and Social Psychology, 51, 1173–1182.

Belisle, F. J., & Hoy, D. R. (1980). The perceived impact of tourism by residents: Acase study in Santa Maria, Columbia. Annals of Tourism Research, 7(1), 83–101.

Bem, S. L. (1981). Gender schema theory: A cognitive account of sex typing.Psychological Review, 88, 354–364.

Biddle, B. J., Bank, B. J., & Slavings, R. L. (1987). Norms, preferences, identitiesand retention decisions. Social Psychology Quarterly, 50, 322–337.

Bramwell, B. (2003). Maltese responses to tourism. Annals of Tourism Research,30(3), 581–605.

Bramwell, B., & Sharman, A. (1999). Collaboration in local tourism policy making.Annals of Tourism Research, 26(2), 392–415.

Briguglio, L. (1995). Small Island Developing States and their economicvulnerabilities. World Development, 23, 1615–1632.

Burke, P. J., & Reitzes, D. C. (1981). The link between identity and roleperformance. Social Psychology, 44, 83–92.

Burke, P. J., & Stets, J. E. (1999). Trust and commitment through self-verification.Social Psychology Quarterly, 62, 347–360.

Callero, P. L. (1985). Role identity salience. Social Psychology Quarterly, 48, 203–215.Cancian, F. M., & Oliker, S. J. (2000). Caring and gender. Thousand Oaks, CA: Pine

Forge Press.Carroll, M. S., & Lee, R. G. (1990). Occupational community and identity among

Pacific Northwestern loggers: Implications for adapting to economic changes.In R. G. Lee, D. R. Field, & W. R. Burch (Eds.), Continuing in the sociology ofnatural resources (pp. 141–156). Boulder, CO: Westview Press.

Chen, J. S., & Hsu, C. H. C. (2001). Developing and validating a riverboat gamingimpact scale. Annals of Tourism Research, 28(2), 459–476.

Chhabra, D. (2007). Gendered social exchange theory: Variations across life spanin casino settings. Anatolia: An International Journal of Tourism and HospitalityResearch, 18, 145–153.

Crompton, R., & Sanderson, K. (1990). Gendered jobs and social change. London:Unwin Hyman.

Deccio, C., & Baloglu, S. (2002). Nonhost community resident reactions to the2002 Winter Olympics: The spillover impacts. Journal of Travel Research, 41,46–56.

Diekhoff, G. (1992). Statistics for the social and behavioral sciences: univariate, bivariate,multivariate. Dubuque, Iowa: Wm. C. Brown.

Douglas, C. H. (2006). Small island states and territories: Sustainable developmentissues and strategies: Challenges for changing islands in a changing world.Sustainable Development, 14, 75–80.

Dyer, P., Gursoy, D., Sharma, B., & Carter, J. (2007). Structural modeling ofresident perceptions of tourism and associated development on the SunshineCoast, Australia. Tourism Management, 28, 409–422.

Fisher, E., & Arnold, S. J. (1990). More than a labor of love: Gender roles attitudes,and consumer behavior. Psychology and Marketing, 11(2), 163–182.

Fischer, E., & Arnold, S. J. (1994). Sex, gender identity, gender role attitude, andconsumer behavior. Psychology and Marketing, 11(2), 163–182.

266 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

Franks, T. R. (1996). Managing sustainable development: Definitions, paradigmsand dimensions. Sustainable Development, 4, 53–60.

Gursoy, D., & Kendall, K. W. (2006). Hosting mega events: Modeling local’ssupport. Annals of Tourism Research, 33(3), 603–623.

Gursoy, D., & Rutherford, D. (2004). Host attitudes toward tourism: An improvedstructural modeling approach. Annals of Tourism Research, 31(3), 495–516.

Gursoy, D., Chi, C. G., & Dyer, P. (2010). Local’s attitudes toward mass andalternative tourism: The case of Sunshine Coast, Australia. Journal of TravelResearch, 49(3), 381–394.

Gursoy, D., Jurowski, C., & Uysal, M. (2002). Resident attitudes: A structuralmodeling approach. Annals of Tourism Research, 29(1), 79–105.

Hagger, M. S., Anderson, M., Kyriakaki, M., & Darkings, S. (2007). Aspects ofidentity and their influence of intentional behavior: Comparing effects forthree health behaviors. Personality and Individual Differences, 42, 355–367.

Hair, J., Anderson, R., Tathman, R., & Black, W. (1998). Multivariate data analysis(5th ed.). Upper Saddle River, NJ: Prentice Hall.

Haralambopoulous, N., & Pizam, A. (1996). Perceived impacts of tourism: The caseof Samos. Annals of Tourism Research, 23(3), 503–526.

Harrison, D. (1992). International tourism and the less developed countries. In D.Harrison (Ed.), The social consequences of tourism in less developed countries(pp. 1–18). London: Bellhaven.

Haukeland, J. V. (1984). Sociocultural impacts of tourism in Scandinavia: Studiesin three host communities. Tourism Management, 5, 207–214.

Hinds, J., & Sparks, P. (2008). Engaging with the natural environment: The role ofaffective connection and identity. Journal of Environmental Psychology, 28,109–120.

Ho, R. (2006). Handbook of univariate and multivariate data analysis and interpretationwith SPSS. Boca Raton: Taylor and Francis.

Huh, C., & Vogt, C. A. (2008). Changes in residents’ attitudes toward tourism overtime: A cohort analytical approach. Journal of Travel Research, 46, 446–455.

Jackson, M. S., & Inbarakan, R. J. (2006). Evaluating residents’ attitudes andintentions to act toward tourism development in Regional Victoria, Australia.International Journal of Tourism Research, 8, 355–366.

Johnson, J., Snepenger, D., & Akis, S. (1994). Residents’ perception of tourismdevelopment. Annals of Tourism Research, 21(3), 629–642.

Jones, D., Jurowski, C., & Uysal, M. (2000). Host community resident’s attitudes. Acomparison of environmental viewpoints. International Journal of Hospitality andTourism Research, 2, 174–189.

Kachigan, M. (1986). Advanced statistics. Ann Arbor: Michigan Press.Kwan, F. V. C., & McCartney, G. (2005). Mapping resident perceptions of gaming

impacts. Journal of Travel Research, 44, 177–187.Laws, E., Faulkner, B., & Moscardo, G. (1998). Embracing and managing change in

tourism. In E. Laws, B. Faulkner, & G. Moscardo (Eds.), Embracing andmanaging change in tourism: International case studies (pp. 1–10). New York:Routledge.

Leary, M. R., & Jones, J. L. (1993). The social psychology of tanning and sunscreenuse: Self-presentational motives as a predictor of risks. Journal of Applied SocialPsychology, 23, 1390–1406.

Lee, C., Kang, S. K., Long, P., & Reisinger, Y. (2010). Residents’ perceptions ofcasino impacts: A comparative study. Tourism Management, 31, 189–201.

Lim, C. C., & Cooper, C. (2009). Beyond sustainability: Optimizing island tourismdevelopment. International Journal of Tourism Research, 11, 89–103.

Lindberg, K., & Johnson, R. L. (1997). Modeling residents attitudes towardtourism. Annals of Tourism Research, 24(2), 402–424.

Liu, J., & Var, T. (1986). Residential attitudes toward tourism impact in Hawaii.Annals of Tourism Research, 13(2), 193–214.

Liu, J., Sheldon, P. J., & Var, T. (1987). Residents perceptions of the environmentalimpact of tourism. Annals of Tourism Research, 14(1), 17–37.

Mackay, K. J., & Campbell, J. M. (2004). An examination of residents’ support forhunting as a tourism product. Tourism Management, 25, 443–452.

R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268 267

Mannetti, L., Peirro, A., & Livi, S. (2004). Recycling: Planned and self expressivebehavior. Journal of Environmental Psychology, 24, 227–236.

Martin, A. S. B., & Gnoth, J. (2009). Is the Marlboro man the only alternative? Therole of gender identity and self-construal salience in evaluating male models.Market Letters, 20, 353–367.

Mason, P., & Cheyne, J. (2000). Residents’ attitudes to proposed tourismdevelopment. Annals of Tourism Research, 27(2), 391–411.

Mbaiwa, J. E. (2011). Changes on traditional livelihood activities and lifestylescaused by tourism development in the Okavango Delta, Botswana. TourismManagement, 32(5), 1050–1060.

McCall, G. J., & Simmons, J. L. (1978). Identities and interaction: An examination ofhuman associations in everyday life. New York: The Free Press.

Milman, A., & Pizam, A. (1988). Social impact of tourism on Central Florida.Annals of Tourism Research, 15(2), 208–220.

Musil, C., Jones, S., & Warner, C. (1998). Structural equation modeling and itsrelationship to multiple regression and factor analysis. Research in Nursing andHealth, 21, 271–281.

Nepal, S. (2008). Residents’ attitudes to tourism in Central British Columbia,Canada. Tourism Geographies, 10, 42–65.

Nunkoo, R., & Ramkissoon, H. (2010a). Gendered theory of planned behavior andresident support for tourism. Current Issues in Tourism, 13(6), 525–540.

Nunkoo, R., & Ramkissoon, H. (2010b). Modeling community support for aproposed integrated resort project. Journal of Sustainable Tourism, 18, 257–277.

Pedhazur, E. J. (1997). Multiple regression in behavioral research (3rd ed.). Orlando,FL: Harcourt Brace.

Perdue, R. R., Long, P. T., & Allen, L. (1990). Resident support for tourismdevelopment. Annals of Tourism Research, 17(4), 586–599.

Petrzelka, P., Krannich, R. S., Brehm, J., & Trentelman, C. K. (2005). Rural tourismand gendered nuances. Annals of Tourism Research, 32(4), 121–1137.

Petrzelka, P., Krannich, R. S., & Brehm, J. M. (2006). Identification with resourcebased occupations and desire for tourism: Are the two necessarily inconsis-tent?. Society and Natural Resources, 19, 693–707.

Pizam, A. (1978). Tourism’s impacts: The social costs of the destinationcommunity as perceived by its residents. Journal of Travel Research, 16, 8–12.

Potts, T. D., & Harril, R. (1998). Enhancing communities for sustainability: A travelecology approach. Tourism Analysis, 3, 133–142.

Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimatingindirect effects in simple mediation models. Behavior Research Methods,Instruments, & Computers, 36(4), 717–731.

Reed, M. G. (2003). Taking stands: Gender and the sustainability of rural communities.Vancouver, Canada: UBC Press.

Sheldon, P. J., & Var, T. (1984). Residents attitudes to tourism in North Wales.Tourism Management, 5, 40–47.

Sirakaya, E., Teye, V., & Sonmez, S. (2002). Understanding residents’ support fortourism development in the central region of Ghana. Journal of Travel Research,40, 151–155.

Skidmore, W. (1975). Theoretical thinking in sociology. London: Cambridge Univer-sity Press.

Smith, M., & Krannich, R. (1998). Tourism dependence and residents attitudes.Annals of Tourism Research, 25(4), 783–801.

Sobel, M. E. (1986). Some new results on indirect effects and their standard errorsin covariance structure models. In N. Tuma (Ed.), Sociological methodology(pp. 159–186). Washington, DC: American Sociological Association.

Sparks, P., & Shepherd, R. (1992). Self identity and the theory of plannedbehavior: Assessing the role of identification with ‘green consumerism’. SocialPsychology Quarterly, 55, 388–399.

Sparks, P., Shepherd, R., Wieringa, N., & Zimmermanns, N. (1995). Perceivedbehavioral control, unrealistic optimism and dietary change: An exploratorystudy. Appetite, 24, 243–255.

268 R. Nunkoo, D. Gursoy / Annals of Tourism Research 39 (2012) 243–268

Spence, J. T., & Helmreich, R. L. (1978). Masculinity and femininity: Theirpsychological dimensions. correlates and antecedents. Austin, TX: University ofTexas Press.

Steele, J., Bourke, L., Luloff, A. E., Liao, P., Theodori, G. L., & Krannich, R. S.(2001). The drop-off/pick-up method for household survey research. Journalof the Community Development Society, 32, 238–250.

Stets, J. E., & Biga, C. F. (2003). Bringing identity theory into environmentalsociology. Sociological Theory, 21, 398–423.

Stets, J. E., & Burke, P. J. (2000). Femininity/masculinity. In E. F. Borgatta & R. J.V. Montgomery (Eds.), Encyclopedia of sociology (pp. 997–1005). New York:Macmillan.

Stronza, A., & Gordillo, J. (2008). Community views of ecotourism. Annals ofTourism Research, 35(2), 448–468.

Stryker, S. (1968). Identity salience and role performance. The importance ofsymbolic interaction theory for family research. Journal of Marriage and theFamily, 30, 558–564.

Stryker, S. (1980). Symbolic interactionism: A social structural version. Benjamin/Cummings: Palo Alto, CA.

Stryker, S., & Burke, P. J. (2000). The past, present and future of identity theory.Social Psychology Quarterly, 63, 284–297.

Tovar, C., & Lockwood, M. (2008). Social impacts of tourism: An Australianregional case study. International Journal of Tourism Research, 10, 365–378.

Twining-Ward, L., & Butler, R. (2002). Implementing STD on a small island:Development and use of sustainable tourism development indicators inSamoa. Journal of Sustainable Tourism, 10, 363–387.

United Nations (1994). Report of the global conference on the sustainable development ofSmall Island Developing States: Bridgetown, Barbados. New York: United Nations.

Vargas-Sanchez, A., Plaza-Mejia, M., & Porras-Bueno, N. (2009). Understandingresidents’ attitudes toward the development of industrial tourism in a formermining community. Journal of Travel Research, 47, 373–387.

Weigert, A. J. (1997). Self, interaction and the natural environment: Refocusing oureyesight. New York: SUNY Press.

Yoon, Y., Gursoy, D., & Chen, J. S. (2001). Validating a tourism development theorywith structural equation modeling. Tourism Management, 22, 363–372.

Submitted 20 October 2010. Resubmitted 22 January 2011. Resubmitted 7 May 2011. Finalversion 19 May 2011. Accepted 25 May 2011. Refereed anonymously. Coordinating Editor:Jan Vidar Haukeland