《社会心理学》课程教学资源（文献资料）social self——Depression and Social Identity - An Integrative Review

Article logy Review Depression and Social ldentity:An Integrative Review 6831452383 SAGE Tegan Cruwys',S.Alexander Haslam',Genevieve A.Dingle', Catherine Haslam',and Jolanda Jetten' Abstract Social relationships play a key role in depression.This is apparent in its etiology,symptomatology,and effective treatment. as been little consensus about the best way to conceptualize the link bet en depression and social ore,the exten al-psy and in part oach i then used as a basisfor conceptualizing the role of social relationships in depression.operationalized in terms of six central hypotheses.Research relevant to the hypothese e preser an agenda for future research to depression,and to translate the nsigh oforetica ntity and y theory,self-categorization theory,mental health Dep he e menta health prob nent of der nn hae hed with at least 20%of people in develoned coun riencing it at some point in their lives.It is the leading 200d S e pronty area of the most ntations hins in D to treating health professionals and evidence indicates that lines the ample evidence that depression is a fundamentally social disorder,with reduced social connectednes sman, .and ta get for treatment chological).research suggests that only questions to be resoved a minority of people with depression receive adequate acute Why so al connectedness is so important in depression,how care (G nell.200 999 Simon,Fle it should be n and how it might be mos vel ap Wells 2001 Often this is due to the vides a social-n (Simon et)or the stigma associated with seeking eon group p esses that explains why social relation anti-depressant (Dwight-Johnso ships are critical for the functioning of the self.This is use 200% 1,Yang a basis for de oping six hype ses that relate to Ke that the a son with a histor of dent sion surement.mechanism.and inter ention.In the" irical expected to e ence four episodes across his or her lifes Evidence:Social Identity and Depression"section,the pan (Judd, 1997).Even among patients who receive the mpr ng a cot The University of Queensland.St Lucia.Australia Rafanelli,Grandi,Conti,&Belluardo,1998).For these rea- sons,ongoing research that contributes to our understanding 304

Personality and Social Psychology Review 2014, Vol. 18(3) 215–238 © 2014 by the Society for Personality and Social Psychology, Inc. Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1088868314523839 pspr.sagepub.com Article Depression is the second most common mental health prob￾lem, with at least 20% of people in developed countries experiencing it at some point in their lives. It is the leading cause of disability worldwide (Lopez, Mathers, Ezzati, Jamison, & Murray, 2006; World Health Organisation, 2012). Depression is one of the most common presentations to treating health professionals, and evidence indicates that its prevalence has been increasing for some time (Klerman & Weissman, 1989; Kruijshaar et al., 2005). Although evi￾dence-based treatments do exist for the condition (both phar￾macological and psychological), research suggests that only a minority of people with depression receive adequate acute care (Goldman, Nielsen, & Champion, 1999; Simon, Fleck, Lucas, & Bushnell, 2004; Young, Klap, Sherbourne, & Wells, 2001). Often, this is due to the expense of treatment (Simon et al., 2004) or the stigma associated with seeking anti-depressant medication or therapy (Dwight-Johnson, Sherbourne, Liao, & Wells, 2000; Phelan, Yang, & Cruz￾Rojas, 2006). Furthermore, relapse rates remain high, such that the average person with a history of depression is expected to experience four episodes across his or her lifes￾pan (Judd, 1997). Even among patients who receive the gold-standard treatment—comprising a combination of anti￾depressant medication and cognitive-behavioral therapy (CBT)—25% are expected to relapse within 2 years (Fava, Rafanelli, Grandi, Conti, & Belluardo, 1998). For these rea￾sons, ongoing research that contributes to our understanding of the etiology and treatment of depression has been priori￾tized by the World Health Orgnaisation (Lopez et al. 2006). In this article, we outline how a social identity approach (SIA) can address both of these priority areas. This review is divided into four broad sections. “The Important Role of Social Relationships in Depression” out￾lines the ample evidence that depression is a fundamentally social disorder, with reduced social connectedness1 impli￾cated as a cause, symptom, and target for treatment of depres￾sion. This section also draws attention to current gaps in knowledge and identifies three key questions to be resolved: Why social connectedness is so important in depression, how it should be measured, and how it might be most effectively enhanced through intervention. “The Social Identity Approach” section provides a social-psychological perspec￾tive on group processes that explains why social relation￾ships are critical for the functioning of the self. This is used as a basis for developing six hypotheses that relate to key aspects of depression and speak to these questions of mea￾surement, mechanism, and intervention. In the “Empirical Evidence: Social Identity and Depression” section, the 523839 PSRXXX10.1177/1088868314523839Personality and Social Psychology ReviewCruwys et al. research-article2014 1 The University of Queensland, St Lucia, Australia Corresponding Author: Tegan Cruwys, School of Psychology, The University of Queensland, St. Lucia, Queensland, 4072, Australia. Email: t.cruwys@uq.edu.au Depression and Social Identity: An Integrative Review Tegan Cruwys1 , S. Alexander Haslam1 , Genevieve A. Dingle1 , Catherine Haslam1 , and Jolanda Jetten1 Abstract Social relationships play a key role in depression. This is apparent in its etiology, symptomatology, and effective treatment. However, there has been little consensus about the best way to conceptualize the link between depression and social relationships. Furthermore, the extensive social-psychological literature on the nature of social relationships, and in particular, research on social identity, has not been integrated with depression research. This review presents evidence that social connectedness is key to understanding the development and resolution of clinical depression. The social identity approach is then used as a basis for conceptualizing the role of social relationships in depression, operationalized in terms of six central hypotheses. Research relevant to these hypotheses is then reviewed. Finally, we present an agenda for future research to advance theoretical and empirical understanding of the link between social identity and depression, and to translate the insights of this approach into clinical practice. Keywords social identity theory, self-categorization theory, depression, social capital, social isolation, mental health Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

216 Personality and Social Psychology Review 18(3) exist literature is reviewed to assess the degr sline for hatter)within a day of the str available evidence supports each of these It therefore appears that people are uniquely sensitive to Finally,"An Agenda for Research Into Social Identity and social forms of stress(such as rejection or conflict)relative Depression"highlights current gaps in the evidence and s to of e events theore cal und rstandin ific ep ness and depression but also to enhancing clinical interven (Tennant,2002).Most commonly,this is a social loss of tions that target its prevention and treatment.In this,our goa some kind such as the death of a loved one,but it may als s to present a novel anal the r le of social conne from othe venues of investigation and of informing clinical practice of depression but also appears to have a causal rolen t The Im ortant Role of Social development n Depression a third way inw hich s ocial connectedne Clinical depression is understood by nd p remission and recovery.Here,there is evidence that impaire tioners alike to be more than simply low mood.n addition social functioning often persists long after remission(C orvel tofeeling miserable,apathetic, and self- 1993;Kennedy,Foy,Sherazi,Mcdor ough,M ng an epis of major de pression als 1989:pa which is social withdrawal.Depression is typically charac Rassaby. 1980)Low social support also predicts terized by social isolation and reduced social connectedness response to treatment and early dropout(Trivedi etal,200 5 of thi has箱 dia gno of el ymntoms and is a core c ent of effective der essio treatment.More specifically CBT for depression (Beck 2011)acknowledges that social isolation isa central featur formally (e.g h pre RT cha impairment in this domain is significantly more common in condition than in other physical and mental illnesses n tha ngfu in her rities that also a sense of pleasure or success,particularly activities that strong risk actor for reppomoonhEn9ealgotevah e approac no ppo. social functioning relative to pharmacological .Kiecol-Gaser,)Fore treatments(Scott et al,2012). Hawkley,and Thisted (2010) found tha a broad foc lation was a goodn val o s on al nd date social connectedness specificallyea by helping a physical health.stress,and a number of objecti e indicator patient rejoin a sports team)rather than other kinds of acti ofs cial-relationship quality;Cacioppo et al ,2010).Lack (e.g.,by daily wall supp ial in but is rather th to be analysis.Bolger.DeLongis.Kessler. ncreased rate of positive reinforcement(Dimidjian,Martell (1989)found that interpersonal conflict was the most impor Addis,Herman-Dunn,2008).Lack of effective social for pre licting daily fluctuatic th nflict had individual (i.e.."me")rather than as a p roblem that i over a number of days.whereas for other kinds of stressor associated with the sense of the self derived from member there was evidence of habituation such that mood returned to ship in a social group ("us).For this reason,a range of

216 Personality and Social Psychology Review 18(3) existing literature is reviewed to assess the degree to which available evidence supports each of these hypotheses. Finally, “An Agenda for Research Into Social Identity and Depression” highlights current gaps in the evidence and sets out an agenda for future research, with a view not only to formulating a theoretical understanding of social connected￾ness and depression but also to enhancing clinical interven￾tions that target its prevention and treatment. In this, our goal is to present a novel analysis of the role of social connected￾ness in depression capable of stimulating new and fruitful avenues of investigation and of informing clinical practice. The Important Role of Social Relationships in Depression Clinical depression is understood by researchers and practi￾tioners alike to be more than simply low mood. In addition to feeling miserable, apathetic, and self-critical, a person experiencing an episode of major depression also exhibits a cluster of cognitive and behavioral changes, chief among which is social withdrawal. Depression is typically charac￾terized by social isolation and reduced social connectedness (Wade & Kendler, 2000). One of its core symptoms (which is as central to diagnosis as low mood) is anhedonia—loss of interest or pleasure in previously enjoyed activities (American Psychiatric Association, 2000). This most typi￾cally manifests as withdrawal from social relationships, both formally (e.g., quitting sporting groups) and informally (e.g., seeing friends less often). Reduced social connected￾ness is thus a key characteristic of depression, such that impairment in this domain is significantly more common in this condition than in other physical and mental illnesses (Hirschfeld et al., 2000). Critically, marked differences in social connectedness also emerge prior to the development of depression symp￾toms. Social isolation has therefore been observed to be a strong risk factor for the development and recurrence of depression (Cacioppo, Hughes, Waite, Hawkley, & Thisted, 2006; Glass, De Leon, Bassuk, & Berkman, 2006; Uchino, Cacioppo, & Kiecolt-Glaser, 1996). For example, in one study, Cacioppo, Hawkley, and Thisted (2010) found that perceived social isolation was a good longitudinal predictor of depression symptoms even after controlling for key candi￾date variables (demographic characteristics, personality, physical health, stress, and a number of objective indicators of social-relationship quality; Cacioppo et al., 2010). Lack of social support has also been found to predict suicidal ide￾ation 1 year later (Handley et al., 2012). In a more fine￾grained analysis, Bolger, DeLongis, Kessler, and Schilling (1989) found that interpersonal conflict was the most impor￾tant stressor for predicting daily fluctuations in negative mood. In addition, these researchers observed that interper￾sonal conflict had escalating effects on mood if it continued over a number of days, whereas for other kinds of stressors, there was evidence of habituation such that mood returned to baseline (or better) within a day of the stressor commencing. It therefore appears that people are uniquely sensitive to social forms of stress (such as rejection or conflict) relative to other stressful life events. Taken a step further, an episode of depression is often triggered by a specific negative event in the social sphere (Tennant, 2002). Most commonly, this is a social loss of some kind such as the death of a loved one, but it may also result from other factors such as family conflict, workplace bullying, or a relationship breakdown (Paykel, 1994). As a result, reduced social connectedness is not only symptomatic of depression but also appears to have a causal role in its development. There is also a third way in which social connectedness is implicated in depression, and this pertains to processes of remission and recovery. Here, there is evidence that impaired social functioning often persists long after remission (Coryell et al., 1993; Kennedy, Foy, Sherazi, Mcdonough, & Mckeon, 2007) and increases the risk of relapse (George, Blazer, Hughes, & Fowler, 1989; Paykel, Emms, Fletcher, & Rassaby, 1980). Low social support also predicts poor response to treatment and early dropout (Trivedi et al., 2005). As a corollary of this, social connectedness has also been found to play a role in the alleviation of depression symptoms and is a core component of effective depression treatment. More specifically, CBT for depression (Beck, 2011) acknowledges that social isolation is a central feature of presentation and often requires targeted intervention. In this regard, behavioral activation is a key CBT strategy that directly targets social connectedness (Cuijpers, Van Straten, & Warmerdam, 2007; Veale, 2008). This technique is pred￾icated on the assumption that withdrawal from meaningful activities maintains depressive symptoms. Patients are therefore encouraged to schedule activities that bring them a sense of pleasure or success, particularly activities that were previously important to them. Speaking to the value of this approach, although it has not been a focus for research attention, there is some evidence that CBT improves social functioning relative to pharmacological treatments (Scott et al., 2012). However, behavioral activation has a broad focus on all kinds of withdrawal or inactivity, and therefore each treat￾ment and practitioner varies in the degree to which they tar￾get social connectedness specifically (e.g., by helping a patient rejoin a sports team) rather than other kinds of activ￾ity (e.g., by helping a patient recommence daily walks). This is because the proposed mechanism of behavioral activation is not social in nature but is rather theorized to be an increased rate of positive reinforcement (Dimidjian, Martell, Addis, & Herman-Dunn, 2008). Lack of effective social functioning in depression is also typically conceptualized as an individual-level deficit—associated with the individual as an individual (i.e., “me”) rather than as a problem that is associated with the sense of the self derived from member￾ship in a social group (“us”). For this reason, a range of Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

Cruwys et a 217 individual-focused interventions-including social skills This section has outlined evidence that compromised training,assertivenes sor conflict resolution training.and social conn ness can precipitate,characterize,and main socia are linical depressio h ial rim,2000:ysa Kashdan.2009:Trivedi et al.2005) ceptualized and measured social connectedness.It is clea Although such inte entions are incr asingly pres for example,that an abundance of terms are used within the (Lewinsohn Clarke, currently,there is no relat the phenomena that we have collectively shins that might allow treating professionals to address these These include social supp ort (Cohen Wills 1985) cemns in a consistent and theoretically informed way loneliness(Cacioppo etal,26),social capital (Putnam only a han ful of studies directly me 、soc1 al net rks (F hrista i,2008) an oc nging (Ba ister itis unelear whether CBT might be enhanced through suring a person's living situation.their number of clo greater focus on social connectedness, or indeed,to wha friends,their employment status,their formal membership of such as mmunity and the in anhat the h about (Collins Do ig3008 ,Auerbach.Derubeis.01) Smith.&Layt tion.a wide variety of formal scales have been used to mea arly,it is unc whether the ett 1y group CBT sure the subjective quality of social exper such CRT of depre the nts of the treatn absence of a unifying frar ork or model that migh researchers have reported that patients atribute much of th lend coherence to the analysis of social connectedness in 9824 ers have rec epre the critical relat T Weissman Research Question 1(RQ1):How should social con- witz speaks R h o tion 2 RO2)Why and how does he outcor in at le of four int affect de that is the mecha role disputes. role transitions or inter nism of action)? Klerman, that Research Question3(RQ3):What types of socialco a in treating 100 addressing social factors in de it doe offer ally derived n del of social relationships in depres- It is with a view to providing a coherent and integrated erbal t was ed as a co e questions n th we the tha ive Res thnthe field of social psychology 1995),it is relatively weak in terms of theory.It espouse nedical "symptom' model of depressic and the goals o The Social Identity Approach re to symptoms ch (SIA)en saville Ch 1984.Weissma n et al.2000).Yet izatio IPT focuses on individual ties rather than group member- theory (Tumer.Hogg.Oakes.Reicher.Wetherell.1987 and no social factors are theorized o routinely mor r,Oakes,Haslam,McGarty,1994).O althoug ope to explanatory model that might account for the central role of 970).the a dominant social-psychological model of groun processe symptomatology,and that has been influential in the study of social and organiza effective tr s10n. ective action (e.g.,see 30

Cruwys et al. 217 individual-focused interventions—including social skills training, assertiveness or conflict resolution training, and increased social activity—are variously recommended as adjuncts to standard treatment (Nilsen, Karevold, Røysamb, Gustavson, & Mathiesen, 2012; Segrin, 2000; Steger & Kashdan, 2009; Trivedi et al., 2005). Although such interventions are increasingly prescribed (Lewinsohn & Clarke, 1999), currently, there is no coherent framework to understand these changes to social relation￾ships that might allow treating professionals to address these concerns in a consistent and theoretically informed way. Furthermore, only a handful of studies have directly mea￾sured social functioning in relation to CBT for depression (e.g., Evans & Connis, 1995; Luk et al., 1991). Consequently, it is unclear whether CBT might be enhanced through a greater focus on social connectedness, or indeed, to what extent the success of strategies such as behavioral activation is in any sense attributable to the improvements in connect￾edness that they may bring about (Collins & Dozois, 2008; Cuijpers et al., 2007; Webb, Auerbach, & Derubeis, 2012). Similarly, it is unclear whether the efficacy of group CBT (which is just as effective in the treatment of depression as individual CBT, see Oei & Dingle, 2008) is attributable to the social components of the treatment. Nevertheless, researchers have reported that patients attribute much of their improvement to group factors (Covi, Roth, & Lipman, 1982). Another key intervention for depression that has proven efficacy is interpersonal psychotherapy (IPT; Elkin et al., 1995). This approach places more emphasis on the critical role of social relationships than CBT (Weissman & Markowitz, 1994) and therefore speaks more directly to the evidence reviewed above. IPT proposes that depression is the outcome of problems in at least one of four interpersonal domains: grief, role disputes, role transitions, or interper￾sonal deficits (Weissman, Markowitz, & Klerman, 2000). However, we argue that although the efficacy of IPT (Weissman & Markowitz, 1994) speaks to the importance of addressing social factors in depression, it does not offer a theoretically derived model of social relationships in depres￾sion. In fact, because IPT was originally developed as a con￾trol verbal therapy condition for CBT in the Treatment of Depression Collaborative Research Project (Elkin et al., 1995), it is relatively weak in terms of theory. It espouses a medical “symptom” model of depression, and the goals of therapy are to alleviate symptoms and improve the social functioning of the individual (Klerman, Weissman, Rounsaville, & Chevron, 1984; Weissman et al., 2000). Yet, IPT focuses on individual ties rather than group member￾ships, and no social factors are theorized or routinely moni￾tored to show treatment outcome. Therefore, although IPT orients treatment toward social factors, it does not provide an explanatory model that might account for the central role of social relationships in the etiology, symptomatology, and effective treatment of depression. Below, we expand on what such a model might look like. This section has outlined evidence that compromised social connectedness can precipitate, characterize, and main￾tain clinical depression. The fact that the literature has con￾sistently found these effects is all the more surprising in light of the many different ways in which researchers have con￾ceptualized and measured social connectedness. It is clear, for example, that an abundance of terms are used within the literature to capture the phenomena that we have collectively referred to as “social connectedness” or “social relation￾ships.” These include social support (Cohen & Wills, 1985), loneliness (Cacioppo et al., 2006), social capital (Putnam, 2001), social networks (Fowler & Christakis, 2008), and belonging (Baumeister & Leary, 1995). Moreover, these constructs have been operationalized in ways as diverse as measuring a person’s living situation, their number of close friends, their employment status, their formal membership of community groups, and the frequency and intensity of their contact with family (Berry & Welsh, 2010; Holt-Lunstad, Smith, & Layton, 2010; Kikuchi & Coleman, 2012). In addi￾tion, a wide variety of formal scales have been used to mea￾sure the subjective quality of social experiences such as perceived support (Harpham, Grant, & Thomas, 2002; Heitzmann & Kaplan, 1988). This diversity in turn speaks to the absence of a unifying framework or model that might lend coherence to the analysis of social connectedness in depression. Indeed, although many researchers have recog￾nized the importance of this relationship, it is apparent that there has been little agreement regarding three central issues: Research Question 1 (RQ1): How should social con￾nectedness be measured? Research Question 2 (RQ2): Why and how does social connectedness affect depression (i.e., what is the mecha￾nism of action)? Research Question 3 (RQ3): What types of social con￾nectedness are likely to be the most beneficial in treating (or reducing the likelihood of) depression? It is with a view to providing a coherent and integrated answer to these questions that, in the following section, we outline a body of theory that has addressed similar issues within the field of social psychology. The Social Identity Approach The Social Identity Approach (SIA) encompasses both social identity theory (Tajfel & Turner, 1979) and self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987; Turner, Oakes, Haslam, & McGarty, 1994). Originally devel￾oped to explain intergroup phenomena, particularly discrimi￾nation and prejudice (Tajfel, 1970), the approach has become a dominant social-psychological model of group processes that has been influential in the study of social and organiza￾tional topics as diverse as leadership, communication, motiva￾tion, and collective action (e.g., see S. A. Haslam, Ellemers, Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

218 Personality and Social Psychology Review 18(3) Reicher Revnolds Schmitt 2010)It has also be From an SIA social relationshins ingly used as a framework for understanding health phenom fore not between ena(S.A.Haslam,Jetten,Postmes,&Haslam,2009:Jetten. individuals(e.g.friendships)that provide a pleasant accom Has Haslam 2012 ngoing personal act Instead,they hav ed in the SIA was not develoned to exnlain the nennd as a connectedness.and it did not emerge from a bio-medically above example.then.it is jane's relationshin with janet and oriented tradition (e.g.,psychiatry).Instead,the approach is and her capacity to define the three of them (and oth -psychological in social identity (as us Io ballers Th er.1999).For thisr on the n why identity-based relationships in the home,in the workplace,and in society at large are critical no in this regard,offers a well-established and long-standin only for self-definition but also for meaningful social 2010).The ga of this sction is to ouine the he。 an be se n that the sia is distin core tenets of the SIA to clarify the relevance of these for emphasis on the power of social group memberships to depression a ely,to formulate a se ructure a person self-conce and,through this, the this condition (). thatindividuals ceive themselves and their place in the world.Indeed.mon Key Premise I:Social Relationships Structure starkly,it suggests that it is social identities that give peopl Individuals'SelfConcebt and.Through This.Their plac in the wo with Behavior A key theoretical pre emise of the SIA is that people's sense of es et).A social identity is meaningful whenever self i comprised of both person and socio t has significance or importance to the individua hat .this e an group see this in o terms of interests attitudes and hehaviors that differ in ball tean meaning to her relationshin with mporta r individuals.On the othe a sense of common direc there are of context which we de This is clearly dan aligned with those of other members of the groups to which generally,it can be seen that in the world at large.social den tties (e.g.. ong (out-groups grou ne's neig A key idea here is that.to the extent that a given g activity that bind people together and allow for coordinated membership is contextually salient or provide s an ongoin goal-oriented endeavor. basis for les a basis for n light of the ising that s nd Jane)not only needs to be able to differentiate be ion it generally feels good to identify str onoly with a g Jill,say）and those s to bl people to defin hat is she ds to be able tos If and which s cial ide is a motiva usindeed,in this way,a sense of shared social identity ional preference (Tajfel&Tumer,1979:see also Elleme l fou ation for De Gil ,Haslam,2004).Here,social iden tification lit ng part o kes or (p.21;see also S.A. eations for self-esteem (Bettencourt Dorr,1997:S.A Haslam,Postmes,&Ellemers,2003). Haslam Reicher.2006:Ellemers.Kortekaas.Ouwerkerk

218 Personality and Social Psychology Review 18(3) Reicher, Reynolds, & Schmitt, 2010). It has also been increas￾ingly used as a framework for understanding health phenom￾ena (S. A. Haslam, Jetten, Postmes, & Haslam, 2009; Jetten, Haslam, & Haslam, 2012). Unlike the models reviewed in the previous section, the SIA was not developed to explain the health benefits of social connectedness, and it did not emerge from a bio-medically oriented tradition (e.g., psychiatry). Instead, the approach is social-psychological in origin and is first and foremost a the￾ory of social relationships grounded in a social model of self (Turner, 1999). For this reason, the approach, although not specific to depression, is relevant to its social dimensions and in this regard, offers a well-established and long-standing model supported by four decades of empirical research (and thousands of publications; for details, see Postmes & Branscombe, 2010). The goal of this section is to outline the core tenets of the SIA to clarify the relevance of these for depression and, ultimately, to formulate a series of testable predictions that might advance our understanding of the role of social connectedness in this condition (see Table 1). Key Premise 1: Social Relationships Structure Individuals’ Self-Concept and, Through This, Their Behavior A key theoretical premise of the SIA is that people’s sense of self is comprised of both personal and social identities. On one hand, this means that we can define and understand our￾selves in terms of our personal identity—seeing ourselves in terms of interests, attitudes, and behaviors that differ in important ways from those of other individuals. On the other hand, there are also a range of contexts in which we define and understand ourselves in terms of one or more social identities—seeing our interests, attitudes, and behaviors as aligned with those of other members of the groups to which we belong (i.e., in-groups) but as different from those of groups to which we do not belong (out-groups; Turner & Oakes, 1997). A key idea here is that, to the extent that a given group membership is contextually salient or provides an ongoing basis for social identification, it provides a basis for self￾categorization whereby the group becomes “self.” For example, to play a game of football, a woman (let us call her Jane) not only needs to be able to differentiate between those players who are on her team (Janet and Jill, say) and those who are not, but she also needs to be able to see her team￾mates as interchangeable representatives of a common in￾group; that is, she needs to be able to see herself and them as “us.” Indeed, in this way, a sense of shared social identity can be seen to provide the psychological foundation for most meaningful forms of social behavior. In simple terms, this is because, as Turner (1982) puts it, “social identity is what makes group behavior possible” (p. 21; see also S. A. Haslam, Postmes, & Ellemers, 2003). From an SIA perspective, social relationships are there￾fore conceptualized not only as bonds of affiliation between individuals (e.g., friendships) that provide a pleasant accom￾paniment to ongoing personal activity. Instead, they have a fundamental bearing on a person’s understanding of who they are and, as a result, on what they are able to do. In the above example, then, it is Jane’s relationship with Janet and Jill—and her capacity to define the three of them (and oth￾ers) in terms of a shared social identity (as “us footballers”)— that allows her to play and enjoy a game of football. The same logic explains why identity-based relationships in the home, in the workplace, and in society at large are critical not only for self-definition but also for meaningful social functioning. From this example, it can be seen that the SIA is distin￾guished from other models of social connectedness by its emphasis on the power of social group memberships to restructure a person’s self-concept and, through this, their behavioral repertoire. The approach argues that social identi￾fication fundamentally affects the way that individuals per￾ceive themselves and their place in the world. Indeed, more starkly, it suggests that it is social identities that give people a place in the world, and thereby also furnish them with a sense of purpose and meaning (Dingle, Brander, Ballantyne, & Baker, 2012; S. A. Haslam, Jetten, & Waghorn, 2009; Jones et al., 2011). A social identity is meaningful whenever it has significance or importance to the individual—that is, when he or she identifies with the group. We see this in our example, where it is Jane’s sense of herself as a member of a football team that gives meaning to her relationship with Janet and Jill and also gives them a sense of common direc￾tion and purpose—by virtue of the fact that this social iden￾tity specifies a constellation of shared norms, goals, and aspirations. This is clearly a mundane example, but more generally, it can be seen that in the world at large, social identities (e.g., where “us” encompasses one’s family, one’s workgroup, one’s church, one’s neighborhood community, etc.) provide the basis for networks of shared meaning and activity that bind people together and allow for coordinated goal-oriented endeavor. In light of the above points, it is not surprising that social identities have a profound impact on well-being. Indeed, pre￾cisely because they engender a sense of purpose and direc￾tion, it generally feels good to identify strongly with a group. This is particularly true, however, to the extent that social￾structural features of the world allow people to define in￾group identity as positive, distinct, and enduring—something for which social identity theory suggests there is a motiva￾tional preference (Tajfel & Turner, 1979; see also Ellemers, De Gilder, & Haslam, 2004). Here, social identification liter￾ally entails being part of something bigger and better, and a large body of research confirms that this has positive impli￾cations for self-esteem (Bettencourt & Dorr, 1997; S. A. Haslam & Reicher, 2006; Ellemers, Kortekaas, & Ouwerkerk, Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

Cruwys et a 219 Table I.Depression-Specific Hypotheses Derived From the Social ldentity Approach. Is of dep H3.The benefit of oup membership for depression symptoms will be moderated by relevant normative content ve in tors of social relations ships will be better pred of depr ymptoms than objective indicator H.Social inte tions for tent th cated in depression 1999:Phinney.Cantu.Kurtz.1997:Wann Branscomb Therefore.in addition to the a tion of how stronely 1990).This,then,is the basis for an initial hypothesis regard- pe rson identifies with a group,and how many groups they ing depression: identify with,to predict the implications of mental Hypothesis 1(H1):Social identification with meaningful ttend to the ups will predict lower levels of depression. the nature and basis of social identification).Accordingly third hypothesis is as follows: Fo ple,as well as being a member of a recreational football mative conten team,Jane may be a psychologist,a mother,a churchgoer and Ins r as each of these socia identitie Key Premise 2:Individuals'Self-Concept and se.and direction.cach can make a unigu and no Social Behavior Are Structured by Perceived contribution to mental health(tt,Hasam Social Relationships Has rch has sho n that,as well as having the i m, poor gene lack me Postmes.&Haslam.2009:Jetten.Haslam.Iyer.&Haslam. 2009).This therefore leads to a second hypothesis: and observation context incr ou con denc the link is real, important,and causa of the in which the life depression. connections,degree of contact,actual support)are represented qul,the process of s ing on John were a prof and only playing SIA is that,in the case of groups,other things are often no connected to his teammates.However,another player,James rs,Haslam,2010;Tajfel Tu outonaplace n the m very not the c identiti On the contrary.there is evidence that.at times.groups can ing one e might say that John(but not James)is amember of the be harmful and impede recovery(Crabtree,Haslam,Postmes football team.However,psycho logically speaking,it is James &Haslm,300nEelL20o:HeesoaCoh (but not hn)who ide the m,anc is this pro 2011.h des ntal healt vehicles for self-definition and social influence Turne outcomes associated with being a member of the team should 991),strong identification with a group that is negatively be more apparent for James than for John,despite the former's enined (e.g.,stigm ial group men heing a fo al tean member might might be apeer group with norms that drug-taking increase the likelihood that social identification is present or self-harm wher the shared it is only ever a crude indicator of individu actors'psy Hill,Borland.2001) on the soc he 30

Cruwys et al. 219 1999; Phinney, Cantu, & Kurtz, 1997; Wann & Branscombe, 1990). This, then, is the basis for an initial hypothesis regard￾ing depression: Hypothesis 1 (H1): Social identification with meaningful groups will predict lower levels of depression. Yet, although any particular social identity has the capac￾ity to be a valuable psychological resource, it is also the case that such identities are rarely mutually exclusive. For exam￾ple, as well as being a member of a recreational football team, Jane may be a psychologist, a mother, a churchgoer, and an Australian. Insofar as each of these social identities has the capacity to provide a person with a sense of meaning, purpose, and direction, each can make a unique and poten￾tially additive contribution to mental health (Jetten, Haslam, Haslam, Dingle, & Jones, 2014; Jones & Jetten, 2011; Ysseldyk, Haslam, & Haslam, 2013). This is particularly true if the identities are compatible (Iyer, Jetten, Tsivrikos, Postmes, & Haslam, 2009; Jetten, Haslam, Iyer, & Haslam, 2009). This therefore leads to a second hypothesis: Hypothesis 2 (H2): Social identification with a greater number of meaningful groups will predict lower levels of depression. Other things being equal, the process of seeing oneself as a member of a valued group (or groups) should generally be beneficial to health. However, a fundamental insight of the SIA is that, in the case of groups, other things are often not equal (S. Reicher, Spears, & Haslam, 2010; Tajfel & Turner, 1979). Accordingly, it is not the case that all social identities are beneficial as a basis for preventing or treating depression. On the contrary, there is evidence that, at times, groups can be harmful and impede recovery (Crabtree, Haslam, Postmes, & Haslam, 2010; Finfgeld, 2000; Helgeson, Cohen, Schulz, & Yasko, 2000; see also Molero, Fuster, Jetten, & Moriano, 2011). In particular, because social identities are powerful vehicles for self-definition and social influence (Turner, 1991), strong identification with a group that is negatively defined (e.g., stigmatized) or whose identity incorporates damaging norms and practices (e.g., anti-social behavior) has the potential to increase health vulnerability. Examples might be a peer group with norms that encourage drug-taking or self-harm—where the shared behavior on which group membership is based is itself deleterious to mental health (Schofield, Pattison, Hill, & Borland, 2001). Therefore, in addition to the question of how strongly a person identifies with a group, and how many groups they identify with, to predict the implications of internalized group memberships for a person’s mental health, it is also critical to attend to the content of those social identities (i.e., the nature and basis of social identification). Accordingly, a third hypothesis is as follows: Hypothesis 3 (H3): The benefit of group membership for depression symptoms will be moderated by relevant nor￾mative content. Key Premise 2: Individuals’ Self-Concept and Social Behavior Are Structured by Perceived Social Relationships A large body of research has shown that, as well as having poor general health, individuals who lack meaningful social relationships are far more prone to depression. The fact that this relationship is robust to differences in researcher perspec￾tive, sample population, and observation context increases our confidence that the link is real, important, and causal. What the SIA offers that goes beyond other models is a specification of the way in which the realities of social life (e.g., number of connections, degree of contact, actual support) are represented psychologically and internalized by perceivers. For example, if John were a professional footballer and only playing with his teammates for financial incentive, he might play on his football team without particularly valuing the team or feeling connected to his teammates. However, another player, James, might miss out on a place in the team but attend training every week, avidly supporting the team, and valuing both the foot￾ball community and his place within it. Sociologically speak￾ing, one might say that John (but not James) is a member of the football team. However, psychologically speaking, it is James (but not John) who identifies with the team, and it is this pro￾cess of identification that is critical in shaping behavior, atti￾tudes, and self-concept. As a result, any mental health outcomes associated with being a member of the team should be more apparent for James than for John, despite the former’s lack of official group member status. The point here, then, is that although objective group membership (e.g., being a formal team member) might increase the likelihood that social identification is present, it is only ever a crude indicator of individual actors’ psy￾chological perspective on the social world. To ascertain individuals’ social connectedness, the most relevant social Table 1. Depression-Specific Hypotheses Derived From the Social Identity Approach. H1. Social identification with meaningful groups will predict lower levels of depression. H2. Social identification with a greater number of meaningful groups will predict lower levels of depression. H3. The benefit of group membership for depression symptoms will be moderated by relevant normative content. H4. Subjective indicators of social relationships will be better predictors of depressive symptoms than objective indicators. H5. Social identification will determine the impact of the various social factors (e.g., social support) that are implicated in depression. H6. Social interventions for depression will be more effective to the extent that they increase social identification. Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

220 Personality and Social Psychology Review 18(3) indicator is therefore their r ion of shared social iden nds on whethe tification in impo rtant and valued domains (Eggins with (Doosie.Haslam.Spears,Oakes O'Brien,Reynolds,Haslam,&Crocker,2008).The fact &Koomen,1998:Hopkins&Murdoch,1999);what it means that it is the subjective experience of belonging to a socia to be compas depends o on compares on mann（0 norato 2004:Rev 2000 In this way,the SIA conceptualizes the self not as a set of Hypothesis 4 (4):Subiective indicators of social rela stable traits but as a potentially fluid process.That said,in nshins will he better predictors ofdepre essive symptoms y con d as stable b than objective indicators. stable CTumer et al.1994).Indeed.it is worth noting tha The foregoing claim that objective reality influence psychologists and psychological methods often go to great huma ony indirectly through th engths to ensure this stability (e.g.. the yar which hun tion transforms reality has long been a dominant focus of ing to personal and social identity formation over the short (particularly in the cognitive tradi term and long term (see Blanz,1999:Oakes,Haslam, psychological rese ing that thi 06 Haslam, wever,It Is won :de en to than for most other models previouslyused tocon the role of social relationships in depression. perceivers to use them)andf(Oakes,Bruner I hes tignmCmticadonasectionhtsocialidcnif6 1957).This means that people will be inclined to define an the in term th that othe of soc ness(e.g.instrumental or emotional support)cannot be ben cample.Jane ma eficia I for health.Nevertheless,it does suggest that socia be more likely to define herself as the supporter of a partic oup n Thi ar football team if she regularly attends the games of tha ng on as a gr ber)wi for ple that whether or not is beneficia ese two contexts for health will depend,among other things,on the degree to workingasa psychologist. who she is asa persor which the source of that support and who she sees herself to be group mer ific group mem feel that the groun matters to them and that they matter to ol w ed and tious (Tumer,Reynolds,Haslam,&Veenstra,2006). ion is tha below Hypothesis 5 (H5):Social identification will det er Hl and H2)should he conditioned by both the ace the impact of the various social factors (e.g.,social sup- port)that are implicated in depression man (say de i Key Pre mise 3:Social ldentific ion Is a Dyr th hi a cess Tha sponds to Me mic ningful Variation in seen as relevant to the issue at hand (e.g dealing with the the Social World stress of parenthood rathe than with s of work;Sani This point Is pa Within the SIA.individuals are understood to define both nd oth as it allows us to son.That is,"self"is defined,in part,through contrast to what when such interventions will affect the self-concept and is"ot self."At a group level,this mea ns that who are is of an individua nd thus,be ng or More specifi lly,it leads to th wing hypothesi defined by who"youare (and what I am not).How Hypothesis 6 (H6):Social interventions for depression the comparative context within which the self is understood will be more effective to the extent that they increase changes,so too will the meaning of self.What it means to be social identification

220 Personality and Social Psychology Review 18(3) indicator is therefore their perception of shared social iden￾tification in important and valued domains (Eggins, O’Brien, Reynolds, Haslam, & Crocker, 2008). The fact that it is the subjective experience of belonging to a social category that underpins meaningful group behavior (Turner & Oakes, 1997) leads to our fourth hypothesis: Hypothesis 4 (H4): Subjective indicators of social rela￾tionships will be better predictors of depressive symptoms than objective indicators. The foregoing claim that objective reality influences human experience and behavior only indirectly through the lens of perception should not be a controversial proposition. On the contrary, the variety of ways in which human percep￾tion transforms reality has long been a dominant focus of psychological research (particularly in the cognitive tradi￾tion; Kruglanski, 1989). However, it is worth noting that this theme provides a stronger focus for social identity research than for most other models previously used to conceptualize the role of social relationships in depression. These arguments lead to an assertion that social identifi￾cation is the “active ingredient” of social connectedness. This does not mean that other aspects of social connected￾ness (e.g., instrumental or emotional support) cannot be ben￾eficial for health. Nevertheless, it does suggest that social identification (i.e., seeing oneself as a group member) will largely determine the impact of these processes. This means, for example, that whether or not social support is beneficial for health will depend, among other things, on the degree to which the source of that support is perceived to be an in￾group member. In other words, the benefits of social rela￾tionships for health should be most apparent when individuals feel that the group matters to them, and that they matter to the group. Both these points have empirical support in pre￾dicting depression, as will be outlined further below. Hypothesis 5 (H5): Social identification will determine the impact of the various social factors (e.g., social sup￾port) that are implicated in depression. Key Premise 3: Social Identification Is a Dynamic Process That Responds to Meaningful Variation in the Social World Within the SIA, individuals are understood to define both themselves and others through a process of social compari￾son. That is, “self” is defined, in part, through contrast to what is “not self.” At a group level, this means that who “we” are is defined partly by our understanding of “them” (and what we are not); whereas at a personal level, who “I” am is partly defined by who “you” are (and what I am not). However, as the comparative context within which the self is understood changes, so too will the meaning of self. What it means to be a psychologist depends on whether one compares oneself with physicists or historians (Doosje, Haslam, Spears, Oakes, & Koomen, 1998; Hopkins & Murdoch, 1999); what it means to be compassionate depends on whether one compares one￾self with Mother Theresa or Adolf Eichmann (Onorato & Turner, 2004; Reynolds & Oakes, 2000). In this way, the SIA conceptualizes the self not as a set of stable traits but as a potentially fluid process. That said, in many contexts, the self will be experienced as stable because the contexts in which it is located (or studied) are relatively stable (Turner et al., 1994). Indeed, it is worth noting that psychologists and psychological methods often go to great lengths to ensure this stability (e.g., through the standardiza￾tion of testing regimes; Reynolds et al., 2010). There are a range of factors that the SIA sees as contribut￾ing to personal and social identity formation over the short term and long term (see Blanz, 1999; Oakes, Haslam, & Turner, 1994; Postmes & Jetten, 2006; Postmes, Haslam, & Swaab, 2005). Broadly, though, the salience of particular identities is seen to reflect their accessibility (the readiness of perceivers to use them) and fit (Oakes, 1987; after Bruner, 1957). This means that people will be inclined to define and understand themselves in terms of categories that have proved to be useful in the past, and that allow them to make sense of their current circumstances. For example, Jane may be more likely to define herself as the supporter of a particu￾lar football team if she regularly attends the games of that team (high accessibility) and if she is in a conversation about football with a supporter of a rival team (high fit) rather than working as a psychologist. Moreover, in these two contexts, who she is as a person—and who she sees herself to be—is likely to reflect the norms of the specific group membership that informs her identity: So that at the game, she is loud and emotional, whereas at work, she is reserved and conscien￾tious (Turner, Reynolds, Haslam, & Veenstra, 2006). One key implication of this analysis for depression is that the ability of an individual to benefit from the social connect￾edness that flows from internalized group membership (as per H1 and H2) should be conditioned by both the accessibil￾ity of a suitable group membership and its contextual fit. So, for example, it should be easier for a man (say) to benefit from the social support that his family can provide if he has a history of strong ties with his family and if his family is seen as relevant to the issue at hand (e.g., dealing with the stress of parenthood rather than with stress of work; Sani, Magrin, Scrignaro, & McCollum, 2010). This point is par￾ticularly important for interventions that seek to improve social connectedness in depression, as it allows us to specify when such interventions will affect the self-concept and social identities of an individual and thus, be beneficial. More specifically, it leads to the following hypothesis: Hypothesis 6 (H6): Social interventions for depression will be more effective to the extent that they increase social identification. Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

Cruwys et a 221 In light of this hypothesis,there is a final point to be made confrontation(S.A.Haslam&Reicher,2006).and discrimi about the utility of the SIA for depression.Unlike most clini- nation (Branscombe,Schmitt,&Harvey,1999).In large. cal models, the approach conceptualize self-concept and able and contex The physical and mental health benefits of social identifi because social identification is a more fluid-and hence. cation are therefore well-established and not specific to potentially treatment-responsive orted cognitionso ack or so these dieamenlo least as important in this domain.In this section.we review Segrin,2000).social identity is generally highly responsive evidence that speaks to this possibility and to each of the six to changes in a person's social or environmental context hypotheses proposed in the previous section.From this fore,social identity interventions that target an indi review.two the nes are apparer +l- First,that current evidenc that much cho full ment (HelliwellBarrington-Leigh.2012).All this sug and the material int Evidence for HI:Social identification with meaningful groups will predict lower levels of depression. depres on. Only a handful of previous studies have included measures mpirical Evidence:Social Identity and or manipulations of social identification along with a depen The sia.on one hand.and the clinical literature on de tities (e. )hat have a validated clinical m onon the ther.rta owith e idicaoofepr (e.g.negative mood)that have validated social identification works.They are es ons"for me and depression has barely commenced construction. here is,however,evidence tha l-being mas s is pro es h eediesaicnion,nabudhnceofempcaleidence high social identification with a valued group predicts fewer depression symptoms.This negative correlation persists For e across divers rang Reynolds,Turner,Bromhead,&Subasic,2009),and when measuring identification with diverse groups including fam ily (Sani et al.,2 0),tertiary inst n that build social identity have been shown to improve well decline (Gleibs, change in depression symptoms over time more strongly H vice versa,although the effect remains significant in et al.,2012: Knight Social identification has also been found to buffer individu mately half used non-diagnostic measures of depr als from the negative impact of a range of stressors,includ they nevertheless tell a consistent story that speaks to the 30

Cruwys et al. 221 In light of this hypothesis, there is a final point to be made about the utility of the SIA for depression. Unlike most clini￾cal models, the approach conceptualizes self-concept and social identity as fundamentally malleable and context￾dependent. From the perspective of developing effective interventions, this malleability has considerable potential, because social identification is a more fluid—and hence, potentially treatment-responsive—construct than, say, dis￾torted cognitions or lack of social skills. Although these lat￾ter constructs might be altered through extended therapeutic work over months or even years (Kovacs & Beck, 1978; Segrin, 2000), social identity is generally highly responsive to changes in a person’s social or environmental context. Therefore, social identity interventions that target an indi￾vidual’s community or environment are likely to have ongo￾ing therapeutic benefits over and above those that can be achieved in brief one-on-one medical or psychological treat￾ment (Helliwell & Barrington-Leigh, 2012). All this sug￾gests that social identification—and the material and psychological factors that feed into it—may be particularly suitable as a target for therapeutic intervention to counteract depression. Empirical Evidence: Social Identity and Depression The SIA, on one hand, and the clinical literature on depres￾sion, on the other, each represent substantial research disci￾plines comprising hundreds of researchers and thousands of published works. They are established fields—“pylons” for the bridge that we propose to build between the two. However, the bridge itself between social identity and depression has barely commenced construction. There is, however, evidence that social identification is a powerful predictor of mental health and well-being more generally. Therefore, although clinical depression has received less attention, an abundance of empirical evidence indicates that social identity is implicated in a range of related health phenomena. For example, along the lines of H2, there is evidence that the number of social identities that people have prior to a stroke is a good predictor of their recovery and well-being 6 months following the event (C. Haslam et al., 2008). Acquiring new group memberships is similarly protective following trauma (Jones et al., 2011; Jones et al., 2012). Among older adults, group interventions that build social identity have been shown to improve well￾being, reduce falls, and slow cognitive decline (Gleibs, Haslam, Haslam, & Jones, 2011; Gleibs, Haslam, Jones, et al., 2011; C. Haslam, Haslam et al., 2010; C. Haslam, Haslam, et al., 2012; Knight, Haslam, & Haslam, 2010). Social identification has also been found to buffer individu￾als from the negative impact of a range of stressors, includ￾ing illness (S. A. Haslam, Jetten, & Waghorn, 2009), memory loss (Jetten, Haslam, Pugliese, Tonks, & Haslam, 2010), confrontation (S. A. Haslam & Reicher, 2006), and discrimi￾nation (Branscombe, Schmitt, & Harvey, 1999). In large, representative community samples social identification has also been found to predict life satisfaction and general well￾being (Helliwell & Barrington-Leigh, 2012). The physical and mental health benefits of social identifi￾cation are therefore well-established and not specific to depression. However, given the centrality of social relation￾ships to the etiology, symptomatology, and treatment of depression, it seems plausible that social identification is at least as important in this domain. In this section, we review evidence that speaks to this possibility and to each of the six hypotheses proposed in the previous section. From this review, two themes are apparent: First, that current evidence is predominantly supportive of these hypotheses; and sec￾ond, that much remains to be done to test these hypotheses fully. Evidence for H1: Social identification with meaningful groups will predict lower levels of depression. Only a handful of previous studies have included measures or manipulations of social identification along with a depen￾dent measure of depression. For the most part, the literature consists either of studies with crude indicators of social iden￾tities (e.g., ethnicity) that have a validated clinical measure of depression, or studies with crude indicators of depression (e.g., negative mood) that have validated social identification measures. Nevertheless, we identified 16 relevant studies that have directly examined the relationship between degree of social identification with valued groups and depression symptoms, with a total of more than 2,700 participants. More detail of each of the studies is provided in Table 2. All studies report a negative relationship between these variables, such that high social identification with a valued group predicts fewer depression symptoms. This negative correlation persists across diverse populations ranging from Norwegian heart surgery patients (S. A. Haslam, O’Brien, Jetten, Vormedal, & Penna, 2005) to Australian school students (Bizumic, Reynolds, Turner, Bromhead, & Subasic, 2009), and when measuring identification with diverse groups including fam￾ily (Sani et al., 2010), tertiary institution (Cameron, 1999) and ethnic group (Branscombe et al., 1999). Some of this work also demonstrates that social identification predicts change in depression symptoms over time more strongly than vice versa, although the effect remains significant in both directions (Iyer et al., 2009; Cruwys et al., in press). Although these studies were correlational and approxi￾mately half used non-diagnostic measures of depression, they nevertheless tell a consistent story that speaks to the protective role of group memberships in preventing depres￾sive symptoms. This story is also consistent with evidence Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

川停育房府香吾意多常寺的含》月 夏胃育育复夏育 言司器囂 N 222

222 Table 2. The Relationship Between Social Identification and Depression Symptoms, as Reported in 14 Published Studies. No. Authors Year Journal N Population Social group ID measure Depression measure r 1 Cruwys, Haslam, Dingle, Jetten, Hornsey, Chong & Oei in press JAffectDisord 52 Disadvantaged community Recreational group 4 items Doosje, Ellemers, and Spears (1995) DASS-21 (21 items) −0.18* 2 92 Psychotherapy patients with depression or anxiety Therapy group 11 items Leach et al., (2008) ZSRDS (20 items) −0.33* 3 Wakefield, Bickley, and Sani 2013 JPsychSom 152 People with MS MS support group 4 items Doosje, et al. (1995) HADS (7 item) −0.31* 4 Sani, Herrera, Wakefield, Boroch, and Gulyas 2012 BJSP Study 1 194 Polish people Family 4 items Doosje et al. (1995) CES-D (20 items) −0.46* 5 Study 2 150 Eastern European Army Unit Army 14 item Leach et al. (2008) BDI-II (21 items) −0.18* 6 Sani, Magrin, Scrignaro, and McCollum 2010 BJSP 113 Adult Scottish community Family 5 items from various scales BDI-II (21 items) −0.32* 7 Bizumic, Reynolds, Turner, Bromhead, and Subasic 2009 APIR Study 1 113 Australian school staff School 4 items Haslam (2001) Mental Health Inventory (adapted) 5 items −0.29* 8 Study 2 693 Australian school students (12-17 years) School 4 items Haslam (2001) DASS-21 (7 items) −0.19* 9 Iyer, Jetten, Tsivrikos, Postmes, and Haslam 2009 BJSP Study 1 105 British students starting university University students 3 items Doosje et al. (1995) 9-item scale Branscombe, Schmitt, and Harvey (1999) −0.30* 10 Study 2 264 British students starting university University students 3 items Doosje et al. (1995) 6-item scale Branscombe et al. (1999) −0.40* 11 S. A. Haslam, O’Brien, Jetten, Vormedal, and Penna 2005 BJSP Study 1 34 Norwegian heart surgery patients Family and friends 2 items Doosje et al. (1995) 6-item negative emotion scale −0.12 12 Study 2 40 British workers (20 bomb disposal officers, 20 bar staff) Work colleagues 2 items Doosje et al. (1995) 4-item negative emotion scale −0.34* 13 Cameron 1999 GD: TRP 167 American university students Mt Allison University 3 factors (Cameron, 1999) BDI-II (21 items) −0.37* −0.11* −0.23* 14 Branscombe et al. 1999 JPSP 139 African Americans Ethnic group 14 item Multi-group Ethnic Identity Measure 6-item negative emotion scale −0.17* 15 Branscombe and Wann 1991 JSSI Study 1 187 American undergraduates Sports team Wann and Branscombe (1990) 1 frequency question −0.16* 16 Study 3 332 American undergraduates Sports team Wann and Branscombe (1990) 1 frequency question −0.10* Note. JAffectDisord = Journal of Affective Disorders; JPsychSom = Journal of Psychosomatic Research; MS = multiple sclerosis; APIR = Applied Psychology: An International Review; BJSP = British Journal of Social Psychology; JPSP = Journal of Personality and Social Psychology; GD: TRP = Group Dynamics: Theory, Research, and Practice; JSSI = Journal of Sport & Social Issues; HADS = Hospital Anxiety and Depression Scale; CES-D = Centre for Epidemiologic Studies – Depression; BDI-II = Beck Depression Inventory 2nd Edition; DASS-21 = Depression, Anxiety Stress Scales (short form); ZSRDS = Zung Self Rated Depression Scale. *p < .05. Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

223 that anomie.social fragmentation.and lack of community identities are less likely to develop depression in the context associated with th of life events that present severe challenges to well-being. More ive still are the ntal data shov ,2011. that manipulations targeting individuals social identifica Speaking further about the importance of the psychologi ) ship for health,evidenc example.in an im estudy.Reicher ows t ers or guards in a simulated prison(revisiting the of the Stanford Prison Experiment;Haney,Banks,& Zimbardo 1973).Over the course of th 9-day experime which hey be nged (rath than one or three)perfor pressor tas more quickl 1070 increase the sense of shared social identity among the prison- Finally Cruwys et al (2013)report findings from an epi demiological study with a repr ntative sample of more ipants that explore ated by the of on-site clinicians the n group of social In a further example,Gleibs,Haslam,Jones,et al.(2011) which respondents belonged was a strong predictor of invited care hom re gender-based socia dep concurrently and ngitudin ally in f 4 year oreove this e ntoms deere Depressed resp ndents with no shins who at a 12-week follow -up.A significant improvement was seen joined one group reduced their risk of depression relapse by among men who ha higher levels of cial isolation and 4 610 hey joined thre e groups their risk re ce b owing ial and that have pn e properti in times of psychological vulne rability o in th depre ssion (C.Ha 3015 t the cally entails ha and there of the findings.Nevertheless,our confidence in H1 is rein fore.conveys greater resilience. ored by the fact Eviden ce for H3:The benefit of group but also to be both reliable and robust mbe depressior will be moderated by relevant normative content Evidence for H2:Social identification witha e a great deal of information to greate of med ingful groups will predict their members about appropriate ways to think,feel,and act lower levels of depression Such information(in the form of"norms";Cialdini&Trost me has be o influ Existing literatur laim tha and varied self-con cept (i.e.,those who are high in self-complexity:Linville, (2012)found that women were influenced to eat either more ile.I987 Ale group me er set a norm con ualized in terms of a person's possession of multiple sumption (Johnston White.2003:Reed.Lange social identities,several authors have argued that these con Ketchie,&Clapp,2007)and smoking (Schofield et al are compatibl Ha am et al.,20 lyer et a Even simply reminding people o hat rt has s he Ithy beha 304

Cruwys et al. 223 that anomie, social fragmentation, and lack of community are associated with the increased prevalence of suicide (Durkheim, 1897/1951; Hawton, Harriss, Hodder, Simkin, & Gunnell, 2001). More persuasive still are the experimental data showing that manipulations targeting individuals’ social identifica￾tion have consequent effects on depression symptoms. For example, in an immersive study, Reicher and Haslam (2006) randomly assigned 15 well-adjusted men to be either prison￾ers or guards in a simulated prison (revisiting the paradigm of the Stanford Prison Experiment; Haney, Banks, & Zimbardo, 1973). Over the course of the 9-day experiment, manipulations (in particular, of group boundary permeabil￾ity; Tajfel & Turner, 1979, see also Ellemers, 1993) served to increase the sense of shared social identity among the prison￾ers and this led to a gradual reduction in depression symp￾toms. However, guards’ sense of shared social identity declined and this led to a significant increase in depression (corroborated by the observations of on-site clinicians). In a further example, Gleibs, Haslam, Jones, et al. (2011) invited care home residents to join gender-based social groups (gentlemen’s and ladies’ clubs) that took part in fort￾nightly social activities. Compared with baseline, social identification increased and depression symptoms decreased at a 12-week follow-up. A significant improvement was seen among men, who had higher levels of social isolation and depression at baseline. Importantly, these experimental studies provide initial evidence for the causal role of social identification in shap￾ing mood and clinical outcomes in vulnerable populations. As with most of the correlational studies in this area, depres￾sion was not the researchers’ primary focus and hence, there are residual questions about the precise clinical implications of the findings. Nevertheless, our confidence in H1 is rein￾forced by the fact that across both surveys and experimental studies, the negative association between social identifica￾tion and depression appears not only to be moderately strong but also to be both reliable and robust. Evidence for H2: Social identification with a greater number of meaningful groups will predict lower levels of depression. Existing literature provides reliable support for the claim that individuals who have a more complex and varied self-con￾cept (i.e., those who are high in self-complexity; Linville, 1987) are buffered against failure in any one domain and are less likely to become depressed (Koch & Shepperd, 2004; Linville, 1987). Although self-complexity was not originally conceptualized in terms of a person’s possession of multiple social identities, several authors have argued that these con￾cepts are compatible (C. Haslam et al., 2008; Iyer et al., 2009). Consistent with this claim, recent studies have found that individuals who report having a greater number of social identities are less likely to develop depression in the context of life events that present severe challenges to well-being, such as starting university (Iyer et al., 2009), having a stroke (C. Haslam et al., 2008), or experiencing brain trauma (Jones et al., 2011). Speaking further about the importance of the psychologi￾cal dimensions of group membership for health, evidence shows that individuals’ objectively measured resilience is increased by experimental manipulations that make multiple group memberships salient. In one study (Jones & Jetten, 2011), participants who were asked to list five groups to which they belonged (rather than one or three) performed better on a cold-pressor task and recovered more quickly from strenuous exercise. Finally, Cruwys et al. (2013) report findings from an epi￾demiological study with a representative sample of more than 4,000 British participants that explored (among other things) the relationship between group memberships and mental health. In this study, the number of social groups to which respondents belonged was a strong predictor of depression symptoms both concurrently and longitudinally (4 years later). Moreover, this effect was more than 3 times as strong among respondents with a history of depression. Depressed respondents with no group memberships who joined one group reduced their risk of depression relapse by 24%; if they joined three groups their risk of relapse reduced by 63%. Such research is part of a growing body of evidence that supports H2 and suggests that social identities are resources that individuals can draw on—and that have pro￾tective properties—in times of psychological vulnerability of the form typically associated with depression (C. Haslam, Jetten, & Haslam, 2012). In these contexts, having more identities typically entails having more resources and there￾fore, conveys greater resilience. Evidence for H3: The benefit of group membership for depression symptoms will be moderated by relevant normative content. Social groups communicate a great deal of information to their members about appropriate ways to think, feel, and act. Such information (in the form of “norms”; Cialdini & Trost, 1998; Turner, 1991) has been shown to influence health via at least two routes. First, behavioral conformity can have sig￾nificant implications for health. For instance, Cruwys et al. (2012) found that women were influenced to eat either more or less popcorn than those in a control condition when an in￾group member set a norm for high or low popcorn consump￾tion, respectively. Similar effects have been found for alcohol consumption (Johnston & White, 2003; Reed, Lange, Ketchie, & Clapp, 2007) and smoking (Schofield et al., 2001). Even simply reminding people of a particular social identity that they hold (i.e., making it salient) is enough for them to endorse less healthy behaviors that are in line with Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015

224 Personality and Social Psychology Review 18(3) that groun's norms fes involving high salt or fat inta 2003).Similarly and directly Oyserman.Fryberg.Yoder.2007:Tarrant Butler.2011). depression.Crabtree et al.(2010)found that social identifi- Along these lines,there is some evidence that behavior cation with a mental health support group was associated mplicated in depre cic sim social infl with greater percei ed social support,rejection of r 2012).Lonelin If has also and of to spread through social networks (Cacioppo,Fowler,& group is not necessarily a positive experience,and it can be Christakis 2009),and suicide and self-harm are so clearly difficult for individuals to leave these groups,particularly le to the tries Stac pport (How ing of these nhe (Stack 2005) found that those who identified more st ongly with a mental The influence of social groups,however,goes beyond health support group had lower self-esteem mere beh Because In this way,the content of social i and the spe- ally defining t i meaning groups unhealth behaviors or is stigmatizing this ha itself.Within the health domain,this has been de nonst the capacity to moderate the capacity of social identification instance St.Claire and He 009 o ameliorate depre In particula in the event tha elderl ribe the manipu Hasla Morion.() with som empirical worl ts reached th em ing th .D condition where their olde oidentity was made salicn 1999:Heilbron Prinstcin 2008)It is not ve and they were told tha ever,whether the potential harm associated with these group ed cogn contras s more potent the various benefits t rell-being tha pa sted th stigm health r m levine and reicher (1996)found that female tive content can result in a non significant relationshi between identification and well-being (Molero et al,2011 lik whe an d re t ive the gro e.g mitt. when thei mbe Gome2 s Morales 2012) the e Evidence for H4:Subjective indicators of social relationshibs will be suberior bredictors of 1 example,that social isolation would prove to be far more depressive symptoms than obje likely to engender depression in t thre ene Among researchers interested in social determinants of tha me ningful social ide health,in recent years,there has been an emphasis on"objec lcators of al con dent parti beh avior social g puns also English i ongitudinal study of age eing:the Ho the relative value or and labour Dynamics in australia)However.a Sia to tif n generally to well- depression would predict that,although indicators such a y and a ervice that redu isk (via identif cation:H1as well as factors that potentially increase depre ience ultimately it is the nsychological sense ofconnection risk (via normative content H3).For ample,on on to the group that is key to predicting mental health up (e.g., s HIV ossible t sitive)can buffer the stress of exnerier ran ing diserimin (Branscombe et al.,1999;Molero et al,2011;Outten well they approximate social identification and therefore Schmitt,Garcia,Branscombe,2009;Schmitt,Spears, how well.theoretically.they should predict related health

224 Personality and Social Psychology Review 18(3) that group’s norms (e.g., involving high salt or fat intake; Oyserman, Fryberg, & Yoder, 2007; Tarrant & Butler, 2011). Along these lines, there is some evidence that behaviors implicated in depression are similarly subject to social influ￾ence—in particular, social withdrawal and suicidality (Handley et al., 2012). Loneliness itself has also been found to spread through social networks (Cacioppo, Fowler, & Christakis, 2009), and suicide and self-harm are so clearly vulnerable to the influence of others (Motto, 1970; Stack, 2003) that many countries have legal limits on media report￾ing of these phenomena (Stack, 2005). The influence of social groups, however, goes beyond mere behavioral conformity. Because identification with a group entails psychologically defining the self as similar to, and part of, the group, the normative content of group iden￾tity also influences thoughts, feelings, and even perception itself. Within the health domain, this has been demonstrated in several studies. For instance, St. Claire and He (2009) found that when participants were made to think of them￾selves as “elderly,” they were more likely to describe them￾selves as having hearing problems. In a study with a similar manipulation, C. Haslam, Morton, et al. (2012) found that 72% of healthy older adults reached threshold for dementia on a diagnostic test when they were randomly assigned to a condition where their older-person identity was made salient and they were told that aging was associated with general￾ized cognitive deficits. By contrast, only 14% of the partici￾pants met the same threshold when their older-person identity was not made salient. More relevant to issues of mental health, R. M. Levine and Reicher (1996) found that female sports science students were more likely to be distressed by a facial scar than a knee injury when their identity as women was made salient but more likely to be distressed by the knee injury when their sports science identity was salient. Although these studies do not speak specifically to depres￾sion, they demonstrate that the experience and expression of mental health symptoms are profoundly influenced by the content of salient social identities. It seems highly plausible, for example, that social isolation would prove to be far more likely to engender depression in contexts where it threatened rather than affirmed a contextually meaningful social iden￾tity (e.g., as a football fan vs. a mountaineer). Moreover, as well as shaping thoughts, feelings, and behavior, social groups also communicate information about the relative value or worth of group members. Although social identification generally provides a boost to well-being, identifying with a stigmatized group is more complex, as it encompasses factors that reduce depression risk (via identifi￾cation; H1) as well as factors that potentially increase depres￾sion risk (via normative content; H3). For example, on one hand, identifying with a stigmatized group (e.g., as African American, as a foreign student, or as someone who is HIV￾positive) can buffer the stress of experiencing discrimination (Branscombe et al., 1999; Molero et al., 2011; Outten, Schmitt, Garcia, & Branscombe, 2009; Schmitt, Spears, & Branscombe, 2003). Similarly, and directly relevant to depression, Crabtree et al. (2010) found that social identifi￾cation with a mental health support group was associated with greater perceived social support, rejection of mental ill￾ness stereotypes, and resistance to stigma. On the other hand, in and of itself, identifying as a member of a stigmatized group is not necessarily a positive experience, and it can be difficult for individuals to leave these groups, particularly when they are also a source of meaning and support (Howard, 2008; Link et al., 1997). Consistent with this, Crabtree et al. found that those who identified more strongly with a mental health support group had lower self-esteem. In this way, the content of social identities and the spe￾cific meaning of groups for their members both have a pow￾erful impact on psychological health. When this content encourages unhealthy behaviors or is stigmatizing, this has the capacity to moderate the capacity of social identification to ameliorate depression. In particular, in the event that a valued social group had normative content encouraging self￾criticism and self-harm, the social identity model would pre￾dict that group members might be at elevated risk of depression. This is consistent with some empirical work demonstrating that these behaviors are subject to social influ￾ence in adolescent peer groups (Dishion, McCord, & Poulin, 1999; Heilbron & Prinstein, 2008). It is not yet clear, how￾ever, whether the potential harm associated with these groups is more potent than the various benefits to well-being that result from social identification. Several studies with stigma￾tized groups have suggested that, at worst, negative norma￾tive content can result in a non-significant relationship between identification and well-being (Molero et al., 2011) and an increased desire to leave the group (e.g., Garstka, Schmitt, Branscombe, & Hummert, 2004; Fernández, Branscombe, Gómez, & Morales, 2012). Evidence for H4: Subjective indicators of social relationships will be superior predictors of depressive symptoms than objective indicators. Among researchers interested in social determinants of health, in recent years, there has been an emphasis on “objec￾tive” indicators of social connectedness, evident particularly in the structure of large-scale population surveys (e.g., the English Longitudinal Study of Ageing; the Housing Income and Labour Dynamics in Australia). However, a SIA to depression would predict that, although indicators such as quality and amount of social contact or access to services might generally be correlated with individuals’ actual expe￾rience, ultimately, it is the psychological sense of connection to the group that is key to predicting mental health outcomes. To interrogate this hypothesis, it is possible to rank vari￾ous measures of social connectedness as a function of how well they approximate social identification and therefore, how well, theoretically, they should predict related health Downloaded from psr.sagepub.com at Remen University of China on September 6, 2015