Group commitment in the face of discrimination: The role of legitimacy appraisals
In contrast to everyday use of the term discrimination, we propose that discrimination can be appraised as either illegitimate or legitimate, and a comprehensive analysis of responses to discrimination needs to account for both ways of experiencing discrimination. We examine how perceived pervasiveness of discrimination and legitimacy appraisals affect group commitment among women in academia (Study 1) and tobacco smokers facing an upcoming smoking ban (Study 2). We found support for our hypothesis that legitimacy of discrimination appraisals moderates the effect of pervasiveness of discrimination. In both studies, group identification and collective action intentions were undermined most when the ingroup claimed that discrimination against them was legitimate and discrimination was perceived as pervasive. In both studies, group identification mediated the effects on collective action intentions. The results highlight the important role of legitimacy appraisals in understanding disadvantaged group members' responses to discriminatory treatment. Copyright © 2010 John Wiley & Sons, Ltd.
When people face discrimination based on their membership in a particular social group they can respond as individuals and attempt to protect their personal identity, or they can respond as group members and protect their collective identity (Branscombe & Ellemers, 1998; Tajfel & Turner, 1979). That is, when threat to a group identity is experienced in the form of discrimination from an outgroup to an ingroup member, people can respond by either distancing from that stigmatized group identity or by increasing their commitment to the ingroup and solidarity with others who suffer the same fate. Prior research with disadvantaged groups has found that when discrimination is perceived as pervasive, group members are likely to display collective responses, reflected in greater commitment to the ingroup (Branscombe, Schmitt, & Harvey, 1999; Foster, 2001; Gurin & Townsend, 1986; Jetten, Branscombe, Schmitt, & Spears, 2001; Schmitt, Branscombe, Kobrynowicz, & Owen, 2002).
Pervasiveness of discrimination reflects the extent to which disadvantaged groups face consistent discrimination across social contexts and the extent to which their experiences with discrimination are diagnostic of future negative treatment. Even though we gain a better understanding of responses to discrimination when taking into account whether discrimination is perceived as pervasive or rare, it is also clear that pervasiveness is not the sole determinant of responses to discrimination. A key question that still needs to be resolved is whether victims of discrimination respond to their treatment in the same way when they perceive their exclusion as justifiable and even appropriate rather than unjustifiable and unfair. Discrimination is typically defined as negative and differential treatment on the basis of group membership (e.g. when a person is not hired or promoted on the basis of race, religion or gender). Such discrimination is often unlawful and perceived to be illegitimate because it is based on people's group membership rather than their merit (i.e. skills or qualifications). When the group appears to be irrelevant to the form of exclusion (e.g. a person who is not hired for an administrative position because they are Jewish), such treatment is likely to be perceived as illegitimate and thus be named ‘discrimination’.
People do not, however, label every instance of exclusion or differential treatment as ‘discrimination’. Contrary to the way, the term discrimination is often used in social discourse, the judgment about whether a group is discriminated against or faces prejudice does not imply by definition that such treatment will be also appraised as illegitimate (see Crandall, Eshleman, & O'Brien, 2002, for empirical evidence of this point). In fact, one does not have to search very far for examples of discrimination that are appraised as legitimate, sometimes even by the targets of discrimination (see Crandall et al., 2002; Crocker & Major, 1994; DePaulo, 2006; Oakes, Haslam, & Turner, 1994). Consider a few examples where people are likely to appraise exclusion based on category membership as legitimate: Exclusion of those who are HIV positive from blood donation, the restriction of certain jobs to able-bodied people, excluding prisoners from voting in national elections or prevention of children from obtaining a driver's license. Other examples include singlism (e.g. favouring married couples over single applicants for rental properties; see DePaulo, 2006), or the exclusion of short people from the army or police force.
Previous social psychological research has considered discrimination that is perceived to be legitimate (Crandall et al., 2002; Crocker & Major, 1994). However, this literature typically examines how advantaged groups perpetuate such discrimination (see Gaertner & Dovidio, 1977). Rather surprisingly, no research (to our knowledge) has examined how disadvantaged groups respond to discrimination when they perceive it to be legitimate. Analysis is needed to better understand the consequences of legitimacy appraisals for disadvantaged group behaviour. We propose that group-based discrimination may have different, even opposite, consequences for its victims when it is appraised as legitimate rather than illegitimate.
Therefore, we address the consequences of the ingroup's claim that discriminatory treatment is legitimate or illegitimate for group commitment. Accordingly, we use the term discrimination to refer to differential group-based treatment, and argue that whether it is appraised as legitimate or illegitimate is a separate judgment. Indeed, the perceived legitimacy of the same form of discrimination can shift across time. For example, in the United States, prior to the Supreme Court decision of Brown v. Board of Education in 1954, exclusion of African Americans from White schools was widely deemed to be legitimate, but after this desegregation decision, it became perceived as illegitimate (see Adams, Biernat, Branscombe, Crandall, & Wrightsman, 2008). Thus, we adopt a definition of discrimination that does not assume that discrimination is appraised as illegitimate per se. Consistent with Brown's (1995) definition of discrimination as ‘the expression of negative affect, (…) towards members of a group on account of their membership of that group’ (p. 8), we argue that such group-based differential treatment is separable from whether it is appraised as legitimate or illegitimate. Doing so allows for a more comprehensive account of how people respond to discrimination, taking into account forms of discrimination and responses that have been previously relatively unexplored.
Drawing on social identity (Tajfel & Turner, 1979) and self-categorization theories (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), we propose that whether pervasive discrimination instigates collective responses depends largely on the extent to which discrimination is appraised as illegitimate. The prevalence of collective responses should be higher when discrimination is perceived as pervasive and illegitimate compared to rare and illegitimate. In contrast, when discrimination against the ingroup is appraised as legitimate, perceiving such discrimination as pervasive is likely to reduce group commitment.
PERVASIVENESS OF DISCRIMINATION AND GROUP COMMITMENT
Social identity theory argues that responses to membership in a disadvantaged group depend on subjective understandings of the intergroup context (Tajfel, 1978; Tajfel & Turner, 1979). Responses to a threatened social identity are likely to be individualistic if people believe that their place in the social hierarchy is determined by their individual deservingness and is not driven by group membership. In contrast, a social change belief system ‘implies that the nature and structure of the relations between social groups in the society is perceived as being characterized by marked stratification, making it impossible or very difficult for individuals, as individuals, to divest themselves of an unsatisfactory, underprivileged or stigmatized group membership’ (p. 35). Pervasive discrimination implies such social stratification—shared discriminatory experiences with other ingroup members, and alienation from those who are seen as the perpetrators of discrimination (i.e. members of the dominant group). Intergroup categorization is strengthened by increased perception of intragroup similarity (Drury & Reicher, 2000; Turner et al., 1987). That is, perceptions of ‘us’ (i.e. we who experience discrimination) are homogenized and ‘we’ are further differentiated from ‘them’ (i.e. those who discriminate against us). Thus, perceptions of pervasive discrimination result in an intergroup understanding of the world (‘they are discriminating against us’) (see Schmitt & Branscombe, 2002).
In contrast, when discrimination is seen as rare—either because it is restricted to a few outgroup perpetrators or because it is limited to a few isolated contexts—it is less likely to severely limit the ability of group members to gain acceptance and status as individuals. Thus, rare discrimination is consistent with social identity theory's social mobility beliefs, in which people are capable of engaging in individual mobility and avoiding the negative consequences of discrimination. When people perceive group-based treatment as rare, they are less likely to define themselves in terms of their group membership because it is mostly irrelevant to understanding the outcomes received from the social world. Rare discriminatory treatment is likely to lead to individual-level responses precisely because exclusion experiences are not highly correlated with group membership.
Research has found that perceptions of pervasive discrimination are positively related to group identification in disadvantaged groups (Branscombe et al., 1999; Leonardelli & Tormala, 2003; Schmitt et al., 2002). Furthermore, experimental research has provided evidence that perceptions of pervasive discrimination have a causal effect on group commitment. Jetten and colleagues (2001, Exp 2) found that people with body piercings who were led to believe they could expect pervasive discrimination from the mainstream identified more strongly with others who have body-piercings compared to those who were led to expect rare instances of discrimination from the mainstream.
APPRAISING DISCRIMINATION AS LEGITIMATE VERSUS ILLEGITIMATE
Importantly though, the existing studies concerning the relationship between the pervasiveness of discrimination and group commitment have been limited to contexts in which discrimination is appraised as illegitimate by those who endure it (Branscombe et al., 1999; Wright, 1997; see also Crandall et al., 2002). Even though prior work has examined targets' perceptions of the fairness of discriminatory treatment, it has typically examined the way context affects fairness ratings (e.g. when discriminatory treatment is ambiguous), or the extent to which disadvantaged group members endorse negative stereotypes (e.g. Mlicki & Ellemers, 1996), rather than whether the discriminatory treatment itself is perceived to be fair (i.e. appraised legitimacy of discrimination as an independent variable). In other words, despite the observation that some forms of group-based discrimination are appraised as legitimate (even by targets), so far, we have rarely studied these forms of discrimination (for an exception see Wright, 1997). This is probably because these forms of discrimination are not perceived to be pressing social problems at present. There is perhaps another reason why researchers may have focused more on forms of discrimination that are appraised as illegitimate rather than legitimate. There has been a tendency in social psychology to view treating individuals on the basis of their group membership by definition as illegitimate (see Oakes et al., 1994; Spears, in press, for a discussion of this point). In recognition of this, the term discrimination is often distinguished from other terms such as differentiation (see Doise, 1986). From this perspective, the idea that negative group-based treatment could be considered legitimate by targets of discrimination is defined out of existence and, so far, is thereby seen as not worthy of investigation.1
For example, social identity theory has addressed the legitimacy of existing status relations (Tajfel & Turner, 1979), but to date has not examined responses to discriminatory treatment that differs in the extent to which it is appraised as legitimate or illegitimate. It has been found that when group members perceive their group's low status as legitimate, they are more accepting of the status-quo than when their group's low status position is seen as illegitimate (Turner & Brown, 1978; see also Livingstone, Spears, Manstead, & Bruder, in press; Mummendey, Mielke, Wenzel, & Kanning, 1996). Furthermore, in minimal groups, when the group's low status is legitimate, group identification and willingness to engage in collective action are lower than when the group's low status is perceived to be illegitimate (Ellemers, Wilke, & van Knippenberg, 1993).
Consistent with these past studies, we predict that when discrimination is appraised as illegitimate, group members have a basis on which to challenge the dominant group and rally ingroup members together in solidarity to address the threat to identity posed by pervasive discrimination. These conditions provide the possibility of successful attempts at social change and the creation of a positively distinct group identity (Reicher & Haslam, 2006; Simon & Klandermans, 2001; Tajfel, 1978). In particular when discrimination is perceived as pervasive (compared to rare), the salience of the intergroup comparison enhances the pressure to change the status quo and the disadvantaged group is more likely to make use of opportunities to directly challenge the illegitimate treatment by the majority group (Simon & Klandermans, 2001; Tajfel & Turner, 1979). In contrast, discrimination that is appraised as legitimate takes away the cause for action by undermining the validity of the collective grievances of the group (Bylsma, Major, & Cozzarelli, 1995; Major & Testa, 1989; Tajfel & Turner, 1979). We predict that this is particularly likely to occur when discrimination is pervasive (compared to when it is rare), because group identification is likely to wane in such situations and the group will lose its capacity to mobilize to engage in collective action (Drury, Cocking, Beale, Hanson, & Rapley, 2005).
Our rationale that group members do indeed sometimes go along with others' discriminatory views touches upon processes suggested by System Justification Theory (SJT; Jost, Banaji, & Nosek, 2004). SJT assumes that people, disadvantaged group members included, are motivated to justify and uphold existing systems of inequality. However, even though SJT has considered the processes whereby groups legitimize status inequality and status hierarchies, this framework has not yet considered the possibility that prejudice and discriminatory treatment can be appraised as legitimate by those suffering from such negative treatment. Importantly too, and in contrast to social identity theory, SJT does not specify the conditions under which the legitimacy of discrimination affects group commitment. In line with social identity theory, we argue that responses to discrimination are conditional on the socio-structural context including how pervasive the discrimination by the dominant group is perceived to be (Tajfel & Turner, 1979).
THE PRESENT RESEARCH
In Study 1, we tested our predictions among women in the work place. Study 2 was conducted among tobacco smokers considering an upcoming smoking ban. We employed two different operationalizations of pervasiveness of discrimination. In Study 1, the pervasiveness of discrimination against the ingroup was manipulated by providing bogus survey results from female employees concerning the extent of discrimination (see Schmitt, Branscombe, & Postmes, 2003). In Study 2, we manipulated whether smokers could expect new limitations from a national smoking ban to affect their group in only a limited number of contexts (rare) or a wide variety of contexts (pervasive).
To our knowledge only one prior study has examined group commitment in response to ingroup claims regarding the degree to which discrimination is appraised as illegitimate. Wright (1997; Exp. 1) found that participants endorsed a more collective response to discrimination if they received information from an ingroup member who expressed anger towards the outgroup and described discrimination as illegitimate. Our studies differ from Wright's because rather than considering legitimacy appraisals as exerting a main effect on group commitment, we consider legitimacy appraisals as a moderator of the effects of pervasive discrimination. We manipulated appraisals of the legitimacy of discrimination by providing false feedback about whether other ingroup members perceived the discrimination they face as legitimate or illegitimate. In many group contexts, other ingroup members' views and claims are an important reference point for appraising the legitimacy of discrimination because ingroup members are perceived as valid sources of social influence (Turner et al., 1987).
In both studies, we examined how perceived pervasiveness and appraised legitimacy of discrimination affect group commitment indicators including group identification and willingness to engage in collective action to counteract the discriminatory treatment. Group commitment is generally perceived as a prerequisite for collective action, but it is also clear that group commitment is not always a sufficient predictor of willingness to engage in collective action (Klandermans, Sabudeco, Rodriguez, & de Weerd, 2002; Simon & Klandermans, 2001). We therefore examined whether perceptions of pervasiveness and legitimacy of discrimination appraisals would affect intentions to engage in actions aimed at counteracting the discriminatory treatment.
We predicted that legitimacy appraisals will moderate the effects of pervasive discrimination on group commitment. Specifically, we predicted that when discrimination is appraised as illegitimate, group commitment should be higher the more that discrimination is perceived as pervasive. When that discrimination is appraised to be illegitimate, we predict that an intergroup interpretation of discrimination (as a pervasive rather than rare phenomenon) will encourage ingroup commitment. Under these conditions, group members are consciously engaged in a power struggle with the group that unfairly treats them (Simon & Klandermans, 2001) and pervasive discrimination can serve as a basis for group commitment (e.g. Branscombe et al., 1999; Jetten et al., 2001). In contrast, when ingroup members appraise discrimination as legitimate, pervasive discrimination should lower group commitment because opportunities for contesting the group's low status are undermined and distancing from the group identity becomes a more attractive option than when discrimination is rare. Therefore, when discrimination is appraised as legitimate, we predicted that group commitment should be lower the more that discrimination is perceived as pervasive.
Participants and Design
Participants were 152 women from three British universities. They were approached via email postings that were distributed by administrators in each of these universities to all women in academic roles that were on the payroll. Participants were invited to take part in a study on gender in the workplace. Age ranged from 21 to over 60 years, with a modal age between 31 and 40 (29.5%). When asked to describe how senior their position in the university was, 3.8% reported being very junior, 14.7% reported being junior, 46.8% were of intermediate seniority, 29.5% described themselves as senior and 4.5% as very senior. The design was a 2 (Pervasiveness of discrimination: rare versus pervasive) × 2 (Appraised legitimacy of discrimination: illegitimate versus legitimate) between-subjects design, with random assignment to conditions.
Procedure and Manipulations
Pervasiveness of discrimination was manipulated first by providing participants with information about gender discrimination in society, substantiated with relevant statistics (adapted from Schmitt et al., 2003). In the pervasive condition, discrimination was described as being widespread and a frequent occurrence in all areas of life (e.g., ‘Women routinely face discrimination and inequality in employment, salary, education, politics, the courtroom and in everyday interpersonal interactions’). In contrast, in the rare discrimination condition, it was emphasized that ‘women now face relatively infrequent discrimination’. We checked the success of this manipulation by assessing how widespread gender discrimination is perceived to be.
Legitimacy appraisals of gender discrimination were manipulated by providing participants with false feedback concerning the responses of female employees in 10 different organizations that we had ostensibly obtained in a prior study where employees were asked to comment on ‘whether they thought it is understandable or not when women are not given the same opportunities as men’. We then presented participants with four responses that represented the overall consensus of opinion among those female employees. The four statements participants received either concurred that gender discrimination was legitimate or illegitimate (see Appendix A). Participants then completed the legitimacy appraisal manipulation check and the dependent measures, after which they were thanked for their participation and debriefed.
All responses were made on 7-point scales ranging from 1, strongly disagree, to 7, strongly agree.
We checked the pervasiveness of discrimination manipulation with three items. The items were: ‘At present, many men discriminate against women’, ‘At present, discrimination is still pervasive in many areas of life’ and ‘At present, discrimination is still widespread in many areas of life’ (α = .77). The legitimacy appraisal manipulation was checked using an open-ended question asking participants to describe in a few words whether they believe the way women are generally treated in the workplace is fair or not. Responses were coded by two independent raters as representing beliefs about whether the treatment of women in the workplace is fair or not, regardless of the extent to which discrimination is perceived as pervasive. Inter-rater reliability was high (Cohen's κ was .78) and inconsistencies were resolved by discussion.
Five items, adapted from Schmitt et al. (2002), were included to measure gender identification: ‘I feel a sense of solidarity with other women’, ‘I like being with other women’, ‘I value other women’, ‘Being a woman reflects something central about who I am’ and ‘Overall, being a woman has very little impact on how I feel about myself’ (reverse-scored). The reliability of the scale was high (α = .80) and the items were averaged with higher scores indicating stronger identification with other women.
Willingness to stand up for women's rights was measured using three items: ‘Thinking about how women are treated makes me want to fight for women's rights,’ ‘As a result of how women are treated I want to stand up for women's issues’, and ‘I want to directly confront those responsible for treating women badly’. The items were averaged (α = .85) with higher scores indicating greater willingness to engage in collective action.
Results and Discussion
Analysis of the pervasiveness of discrimination manipulation check revealed a significant main effect of pervasiveness of discrimination, F(1, 149) = 17.15, p < .001, η = .103. In line with the manipulation, discrimination was perceived as more pervasive in the pervasive condition (M = 5.13, SD = 1.24) compared to the rare condition (M = 4.19, SD = 1.54). Both the legitimacy of discrimination main effect and the interaction between pervasiveness and legitimacy were not significant, F(1, 149) = .08, p = .784, η = .001, and F(1, 149) = .23, p = .630, η = .002, respectively.
Frequencies of irrelevant, neutral or missing responses on the legitimacy appraisal check did not differ across the two legitimacy conditions. Further analyses revealed that, even though participants in the legitimate condition were just as likely to express views that women are treated unfairly in the workplace (n = 26, 52.0%) compared to fairly (n = 24, 48.0%), participants in the illegitimate appraisal condition were more likely to perceive the treatment of women in the workplace as unfair (n = 42, 76.4%) than fair (n = 13, 23.6%), χ2(2) = 6.82, p = .032. When we only analysed responses coded as unfair treatment, we found that the frequency of these responses was higher in the illegitimate than in the legitimate appraisal condition, χ2(1) = 3.77, p = .052. Responses did not vary as a function of the pervasiveness of discrimination manipulation, χ2(1) = .18, p = .671.
In sum, even though women were more likely to perceive discrimination as unfair than as fair, in line with the manipulation, participants were more likely to perceive the treatment of women as unfair in the illegitimate appraisal condition than in the legitimate appraisal condition.
Analysis of the group identification measure revealed only a significant interaction between perceived legitimacy and pervasiveness of discrimination, F(1, 149) = 4.11, p = .044, η = .027. The means and standard deviations are displayed in Table 1. Identification tended to be higher when discrimination was appraised as pervasive and illegitimate than when it was appraised as rare and illegitimate, F(1, 149) = 3.46, p = .065, but pervasiveness did not affect identification within the legitimacy condition, F(1, 149) = 1.06, p = .304.
Table 1. The effect of appraised legitimacy and perceived pervasiveness of discrimination on gender identification and collective action intentions among female employees: Study 1
|Gender identification||5.31a (1.04)||5.75b (.92)||5.46a (1.09)||5.20ab (1.21)|
|Collective action intentions||4.21ab (1.29)||4.56a (1.45)||4.62a (1.40)||3.79b (1.35)|
Analysis of the willingness to engage in collective action scale revealed the predicted interaction, F(1, 148) = 7.00, p = .009, η = .045 (see Table 1). Simple main effect analyses showed that when discrimination was appraised as illegitimate, willingness to engage in collective action did not differ in the pervasive condition compared to the rare condition, F(1, 149) = 1.33, p = .250. When discrimination was appraised as legitimate, collective action intentions were lower when discrimination was pervasive than when it was rare, F(1, 148) = 6.47, p = .012.
Although the pattern of significant simple effects differed somewhat for group identification and collective action, the overall interaction was the same. Therefore, we sought to determine whether group identification mediated the effects of the two manipulations on collective action. As recommended for small samples, we used bootstrapping analyses (see Preacher, Rucker, & Hayes, 2007) to test a model whereby identification mediates the interactive effect of pervasiveness and legitimacy on willingness to engage in collective action. In these analyses, mediation is significant if the 95% Bias Corrected confidence intervals for the indirect (mediated) effect do not include 0 (Preacher et al., 2007). Results based on 1000 bootstrapped samples indicated that zero is not included in the 95% confidence intervals (lower CI = −.805, upper CI = −.002). This provides evidence for mediated moderation, where identification mediated the interactive effect of legitimacy appraisals and pervasiveness of discrimination on willingness to engage in collective action.
In sum, we found that effects of legitimacy appraisals were most pronounced when discrimination was pervasive. However, the effects were slightly different on measures tapping gender identification and measures assessing willingness to engage in collective action. We found that when discrimination in the workplace was appraised as legitimate, perceptions of pervasive gender discrimination led to lower collective action intentions than when discrimination was perceived as rare. When gender discrimination was appraised as illegitimate, gender identification was marginally more pronounced when discrimination was pervasive than when discrimination was rare. Moreover, despite the fact that the simple main effect comparisons on gender identification and collective action intentions were somewhat different, we found support for the hypothesis that the interaction between pervasiveness and legitimacy of discrimination on collective action intentions was mediated by group identification.
These results provide initial evidence for the detrimental effect on collective action intentions when discrimination is perceived as pervasive and legitimate. Under such conditions, ingroup members are less interested in engaging in action aimed at bringing discrimination to the attention of others. The findings in the legitimacy condition suggest that pervasive discrimination will be more likely than rare discrimination to lead to withdrawal of support for the ingroup and encourage self-definition in individual terms. A trend in the opposite direction was found when discrimination was appraised as illegitimate and pervasive. Women were marginally more likely to align themselves with their gender group and support collective action under these conditions than when discrimination was rare.
We examined whether the effects obtained with gender in Study 1 would generalize to a rather different group—tobacco smokers. Examining our predictions among smokers permits us to manipulate pervasiveness of discrimination differently than in Study 1. At the time of the study, there were discussions in the United Kingdom about the introduction of a smoking ban in public places (to take effect July 2007). In an online questionnaire, we varied perceptions of whether this ban would either restrict the rights of smokers in many new contexts (pervasive discrimination condition) or the ban would not increase the exclusion of smokers in additional contexts (rare discrimination condition). We again manipulated legitimacy of discrimination by having participants witness ingroup members expressing the opinion that group-based negative treatment is fair or unfair.
We predicted that the manipulations of legitimacy appraisals and pervasiveness of discrimination would affect identification with other smokers and willingness to engage in collective action. As in Study 1, we also predicted that group identification would mediate the effect of the manipulations on collective action intentions. We again expected an interaction between legitimacy appraisals and pervasiveness of discrimination whereby pervasive and illegitimate discrimination should enhance group commitment and willingness to engage in collective action compared to when discrimination is illegitimate and rare. We again tested whether appraised legitimate and pervasive discrimination lowers responses on these measures compared to when discrimination was appraised as legitimate and rare.
In addition, we assessed willingness to quit smoking as a form of exiting the group. We predicted the opposite pattern of findings on this measure as on group identification and willingness to engage in collective action. That is, willingness to quit smoking should be lower when discrimination is appraised as pervasive and illegitimate than when it is appraised as rare and illegitimate. In contrast, willingness to quit smoking should be higher when discrimination is appraised as pervasive and legitimate than when it is appraised as rare and legitimate. We also examined whether group identification mediated the intention to quit smoking effect to determine whether intention to quit smoking can be conceived of as an individual mobility strategy (Tajfel & Turner, 1979).
Participants and Design
Participants were 95 self-identified smokers, taking part in an online study (59 males; 36 females; M age = 42 years). The design was a 2 (Pervasiveness of discrimination: rare versus pervasive) × 2 (Appraised legitimacy of discrimination: illegitimate versus legitimate) between-subjects factorial design, with random assignment to conditions.
An online questionnaire invited smokers to take part in a study assessing their views and opinions. At the end of the questionnaire, participants were asked to indicate their gender, age, the number of years they had been smoking, and to respond to nine questions assessing the extent to which they were addicted (e.g. ‘overall, how addicted are you to smoking’, α = .75) on 7-point scales. At the end of the study, participants were thanked for their participation and debriefed.
Pervasiveness of discrimination was manipulated by providing participants with a fictitious BBC news report on how the planned smoking ban was expected to affect the lives of smokers in the UK. In the pervasive discrimination condition, smokers read that the ban would severely restrict smoking in public places (e.g. ‘The White Paper on Public Health plans to make most enclosed public areas, including offices and factories, smoke-free’ and that ‘This proposal for a smoking ban goes further than had been expected—up to 90% of bars could be smoke-free within a few years’). In the rare discrimination condition, smokers read that many public places would not be affected by the ban (e.g. ‘Smoking will still be allowed in pubs that limit their food sales to snacks, rather than prepared meals’ and that it was believed that ‘most people would be happy with more no-smoking areas and better ventilation’).
We manipulated perceptions of legitimacy of discrimination in a similar way as in Study 1 by presenting smokers with statements from ingroup members arguing for or against the legitimacy of discriminatory treatment of smokers by non-smokers. Participants were told that we had conducted a study among smokers on whether they felt it was understandable when others treated them negatively on the basis of their being a smoker. Participants were then presented with excerpts of responses that ostensibly came from six smokers (see Appendix B).
One item was included to assess the effectiveness of the pervasiveness manipulation: ‘Smokers should expect to be discriminated against by non-smokers’. The legitimacy of discrimination manipulation was checked with three items assessing to what extent smokers felt it was understandable, reasonable, and justifiable when others excluded them because they are a smoker (1 ‘strongly disagree’ to 7 ‘strongly agree’, α = .90).
The same five items as in Study 1 were adapted to assess identification with other smokers (e.g. ‘I feel a sense of solidarity with other smokers’) on a scale ranging from 1 ‘strongly disagree’ to 7 ‘strongly agree’ (α = .69, after reverse scoring the last item).
Willingness to engage in collective action was assessed by asking participants to what extent they would engage in the following behaviours if they were treated negatively for being a smoker: ‘Stop smoking to reduce your chances of being discriminated against in the future’, ‘Not worry too much and just accept it as it is’ and ‘Bring the negative treatment of smokers to the attention of anti-discrimination organizations’ (ranging from 1 ‘not at all’ to 7 ‘very much’). The reliability of the three items was satisfactory (α = .63; after recoding the first two items). These items were averaged, with higher scores indicating higher willingness to engage in collective action.
Intentions to Quit Smoking
One item was included assessing desire to quit smoking (‘I would like to quit smoking’) on a scale ranging from 1, strongly disagree to 7, strongly agree.
Results and Discussion
Participants reported having been a smoker on average for 25 years (ranging from 2 to 60 years). Preliminary analyses revealed that number of years as smoker and self-perceived addiction did covary with the dependent variables. Therefore we controlled for these variables in the analyses reported below. While we had explicitly solicited smokers in Britain to take part in the study (as the ban was directly relevant to their future), there were 10 non-British residents who completed the questionnaire (i.e. residents of the United States, Germany and Ireland). Preliminary analyses revealed, however, that including country of residence in the analyses did not affect the results in any of the analyses and we therefore included the non-British participants in the analyses reported below.
Analysis of the pervasiveness of discrimination measure revealed a marginal effect of the pervasiveness manipulation, F(1, 91) = 3.33, p = .071, η = .035. In line with the manipulation, there was a tendency for smokers to expect more discrimination from non-smokers in the pervasive condition (M = 4.09, SD = 2.40) than in the rare discrimination condition (M = 3.19, SD = 2.33). The main effect for legitimacy of discrimination and the interaction term were not significant, F(1, 91) = 1.85, p = .177, η = .02, and F(1, 91) = .19, p = .665, η = .002, respectively.
We found a main effect of the legitimacy of discrimination manipulation on the legitimacy check, F(1, 89) = 4.25, p = .042, η = .046. Discrimination was appraised as more legitimate in the legitimate discrimination condition (M = 2.70, SD = 2.02) than in the illegitimate discrimination condition (M = 1.92, SD = 1.57). The main effect for pervasiveness of discrimination and the interaction term were not significant, F(1, 89) = .01, p = .984, η = .000, and F(1, 89) = .94, p = .334, η = .010, respectively. In sum, results on the manipulation checks of legitimacy and pervasiveness of discrimination provide evidence that these constructs were manipulated orthogonally.
Analysis of the group identification measure revealed that self-rated addiction was a significant covariate, F(1, 89) = 6.84, p = .01, η = .071, with higher identification the more participants perceived themselves as addicted. The only other effect was a significant interaction between perceived legitimacy and pervasiveness of discrimination, F(1, 89) = 5.72, p = .019, η = .060. The adjusted means and standard deviations are displayed in Table 2. Identification was lower when discrimination was appraised as pervasive and legitimate than when it was appraised as rare and legitimate, F(1, 89) = 6.90, p = .010. Pervasiveness did not affect identification in the illegitimate discrimination condition, F(1, 89) = 1.89, p = .479.
Table 2. The effect of appraised legitimacy and perceived pervasiveness of discrimination on group identification, collective action intentions, and intention to quit smoking, adjusted means: Study 2
|Group identification||4.57ab (.96)||4.82ab (1.73)||5.18a (1.00)||4.23b (1.20)|
|Collective action intentions||5.38ab (1.52)||5.70a (1.53)||5.62a (1.06)||4.72b (1.76)|
|Intention to quit||3.06ab (2.19)||2.23a (2.12)||2.16a (1.68)||3.93b (2.47)|
The analysis of covariance assessing effects of pervasiveness and legitimacy of discrimination on collective action intentions revealed that number of years smoking was a significant covariate, F(1, 89) = 13.49, p < .001, η = .132. Willingness to engage in collective action was higher the longer participants had been smoking. In addition, the predicted interaction was found, F(1, 89) = 4.18, p = .044, η = .045 (see Table 2). Simple main effect analyses revealed that when discrimination was appraised as legitimate, collective action intentions were lower if discrimination was pervasive than if it was rare, F(1, 89) = 4.37, p = .039. Collective action intentions did not differ as a function of pervasiveness in the illegitimate discrimination condition, F(1, 89) = .59, p = .443.
As in Study 2, we assessed whether identification with other smokers mediated the effect of the manipulations on willingness to engage in collective action. Results based on 1000 bootstrapped samples indicated that zero was not included in the 95% confidence interval for the indirect effect (lower 95%CI = −1.071, upper 95%CI = −.017). This supports our prediction of mediated moderation—the interactive effect of appraised legitimacy and pervasiveness of discrimination on collective action intentions is mediated by group identification.
Intention to Quit Smoking
Analysis of variance on the intention to quit smoking measure revealed a significant effect of number of years smoking, F(1, 89) = 4.50, p = .037, η = .048, with greater intention to stop smoking the fewer years participants had been smoking. The only other effect was a significant interaction between legitimacy and pervasiveness of discrimination, F(1, 89) = 8.66, p = .004, η = .089 (see Table 2). Simple main effects revealed, in line with predictions, when discrimination was appraised as legitimate, intention to quit smoking was higher in the pervasive than the rare condition, F(1, 89) = 7.69, p = .007. When discrimination was appraised as illegitimate, intention to quit did not differ in the rare compared to the pervasive condition, F(1, 89) = .98, p = .326.2
In sum, the interactive effect of legitimacy and pervasiveness of discrimination was replicated among tobacco smokers. We found that when discrimination was appraised as legitimate, pervasive discrimination lowered group identification and collective action intentions compared to when discrimination was perceived as rare. However, when discrimination was appraised as illegitimate, no differences as a function of pervasiveness were observed on any of the three dependent measures. Mediational analyses provided evidence that the interactive effect of pervasiveness and appraised legitimacy on collective action intentions is mediated by group identification.
We found this same interactive effect of the manipulations on measures assessing intentions to quit smoking. When discrimination was appraised as more legitimate, perceiving discrimination as more pervasive increased intentions to quit smoking. From an applied perspective, this suggests that smoking bans in contexts in which exclusion of smokers is appraised as legitimate could be effective in reducing smoking, but when such bans are appraised as illegitimate they may actually be counterproductive in the sense that smokers are less likely to want to quit smoking.
By disentangling effects of appraised legitimate versus illegitimate discrimination, it became clear that their consequences for group commitment depend on pervasiveness of discrimination. Because we adopted a definition of discrimination that is less restrictive and does not assume that discrimination is by definition only those forms of group-based exclusion that are appraised as illegitimate, we were better able to examine a broader range of discriminatory behaviours (see Brown, 1995; Crandall et al., 2002).
Our findings show that pervasive discriminations can undermine group commitment when discrimination is perceived as legitimate. Specifically, group identification (Study 2) and collective action intentions (Studies 1 and 2) were lower when discrimination was perceived as pervasive and legitimate compared to rare and legitimate. This finding is novel and demonstrates that the effect observed in previous research where pervasive discrimination enhances group commitment (Jetten et al., 2001; Leonardelli & Tormala, 2003; McCoy & Major, 2003; Schmitt et al., 2002) is absent when discrimination is appraised as more legitimate. Our findings suggest that even though pervasiveness of discrimination enhances ‘Us’ versus ‘Them’ distinctions, in the context of legitimate discrimination this does not result in collective responses. Instead, legitimate discrimination weakens the basis for challenging the discriminatory treatment (Bylsma et al., 1995; Major & Testa, 1989). Because the ingroup appears to be deserving of discriminatory treatment, this effectively blocks the possibility of achieving a more positive social identity through collective action. Because the disadvantaged group does not have the means to counteract the negative consequences of discrimination collectively, disadvantaged group members are likely to seek out more individualistic identity management strategies to cope with the exclusion they face (Tajfel & Turner, 1979). That is, in particular when discrimination is pervasive, members distance themselves psychologically from the disadvantaged group under such conditions, compared to when discrimination is rare. The lower collective action intentions in such contexts is likely to further reinforce the perceived legitimacy of discrimination and lend support for the existing status quo.
We only found a marginally significant effect on the group identification measure in Study 1 such that when discrimination was appraised as illegitimate, group identification was higher when discrimination was pervasive rather than rare. This finding concerning the responses of disadvantaged group members to illegitimate discrimination confirms and expands on previous research (see also Jetten et al., 2001; Leonardelli & Tormala, 2003; McCoy & Major, 2003; Schmitt et al., 2002). This suggests that when the outgroup engages in socially contestable exclusion, the pervasiveness of discrimination is not a deterrent to maintaining group membership, but fuels commitment to the group.
However, it is also fair to say that support for the prediction in relation to illegitimate discrimination was rather weak (despite the fact that the pattern across both studies on the group identification and collective action intention measure was in the predicted direction). Although it is unclear why we did not find stronger support for this prediction here, it is important to be cautious in interpreting the lack of significance. It is difficult to directly compare the present studies to previous research precisely because the manipulations of illegitimacy and pervasiveness were not orthogonal in prior work (as they were in the present studies).
Implications and Future Research
Our research adds to existing research on the importance of legitimacy appraisals. Previous social identity research has typically examined the role of appraised legitimacy on group commitment in minimal groups (e.g. Ellemers et al., 1993; Wright, 1997) or groups that are not faced with disadvantage in society at large (e.g. arts versus science students, Turner & Brown, 1978). In our studies, we examined the role of legitimacy appraisals ingroups that are disadvantaged in ways that are not limited to a temporary laboratory context. The few studies that have examined the relationship between legitimacy appraisals and group commitment among real disadvantaged groups have been correlational in nature (see Livingstone et al., in press; Mummendey et al., 1996). By manipulating ingroup claims that discrimination is legitimate or illegitimate, we were able to provide greater certainty about the direction of causality in the relationship between legitimacy appraisals and group commitment. Our findings confirm the important role of legitimacy appraisals, and their interaction with the social context, as determinants of group commitment (Tajfel & Turner, 1979). It is important to note that our manipulation checks show that perceptions of pervasiveness and legitimacy appraisals were manipulated orthogonally. That is, in our participants' minds, it was quite conceivable to consider discrimination to be rare and appraise it as illegitimate or to perceive discrimination as pervasive but to appraise it as legitimate. That is, the perceived legitimacy of discrimination and its pervasiveness (or vice versa) were independent judgments.
In both studies, we manipulated legitimacy of discrimination by presenting participants with other ingroup members' claims that their group's treatment was legitimate or illegitimate. Our reasoning was that group members typically turn to other ingroup members as valid sources of information on how to appraise the situation (Turner et al., 1987; Wright, 1997). Our legitimacy manipulation checks in both studies provide good evidence that such claims by the ingroup did affect participants' appraisals of the legitimacy of discrimination. However, future research should explore this process in more detail by manipulating legitimacy appraisals in other ways in order to assess the effects of ingroup claims and perceived legitimacy on collective responses. Future research should also examine whether our effects are only observed when the ingroup claims that discrimination is legitimate. We would predict that similar claims by outgroup members would not have the same impact, if only because ingroup members may have a priori expectations that outgroup members will appraise discrimination as legitimate. When discrimination is pervasive, and outgroup members appraise ingroup discrimination as legitimate, we predict that intergroup dynamics should become increasingly hostile and commitment to the minority group should be enhanced (rather than attenuated as we found here).
Our findings corroborate and highlight the central role that Tajfel and Turner (1979) credited to the socio-structural context in determining responses to low status and discriminatory treatment. Whether groups will simply give up their collective struggle when faced with discrimination depends on the pervasiveness of group-based exclusion and the appraisal of the legitimacy of the discriminatory treatment they receive. For female employees and smokers, there was only weak support that facing pervasive discrimination in combination with the perception that discriminatory treatment is illegitimate strengthened group members' identification (Study 1). Our data revealed more support for the prediction that pervasive discrimination should lead to the opposite response when such discriminatory treatment is perceived as legitimate by those affected by it. In such contexts, pervasive discrimination undermined the extent to which disadvantaged group members were likely to engage in collective level identity management strategies and it weakened their willingness to fight for the collective cause.
We thank Mette Hersby, Catherine Philpot and Anna Rabinovich for their assistance with the data collection and coding. This research was supported by an ESRC grant (R000223981) and an ARC grant (DP0984576).
For example, in contrast to Brown's (1995) definition adopted here, Major, Quinton, and McCoy (2002) refer to legitimacy/justice in their definition of discrimination. They define discrimination as ‘an unjustified negative or harmful action toward the members of a group, simply because of their membership in that group’ (p. 264). The Major et al. (2002) definitions precludes the possibility that discrimination can be appraised as legitimate by those who are negatively affected by it. Our research empirically separates these two components—appraised legitimacy and perceived pervasiveness of discrimination.
Further analyses examined whether intentions to quit were mediated by group identification. The bootstrapping analyses revealed that zero was just included in the 95% confidence interval for the indirect effect (lower CI = −.007, upper CI = 1.19). This suggests that rather than group identification being a mediator of the intention to quit effect, intention to quit may be better conceived of as an indicator of actual disidentification (rather than simply low group identification as measured here). That is, our group identification measure may not have assessed disidentification; as a result, this could be why our identification measure did not serve as a reliable mediator.
Illegitimate Discrimination Condition:
‘…I don't think it is right at all that women are kept out of top-manager positions. It is too often (wrongly) assumed that women are not willing to work the 60+ hours or so that are needed at that level, just because they have children’ (29-year-old female, sales assistant).
‘…Yes, women *are* treated differently in my organization, and most of the time I think it's totally unjustified’ (36-year-old female, manager).
‘…Women are more concerned with a healthy work-life balance, but that does not mean they should not get promoted when they clearly are ready for a more senior position’ (46-year-old female, administrator).
‘… In my organization, I'm not always so sure that people get promoted on the basis of performance. I do think there is plenty of unfairness towards women’ (58-year-old female, academic).
Legitimate Discrimination Condition:
‘…I think it is understandable that women are kept out of top-manager positions. Many women have children and they are just not willing to work the 60+ hours or so that are needed at that level’ (29-year-old female, sales assistant).
‘…Yes, women *are* treated differently in my organization, and most of the time I think it's totally justified’ (36-year-old female, manager).
‘…Women are more concerned with a healthy work-life balance. So it is understandable that you don't find that many women at the top’ (46-year-old female, administrator).
‘… In my organization, it is clear that people get promoted on the basis of performance and there is no unfairness in that’ (58-year-old female, academic).
Illegitimate Discrimination Condition:
‘…I don't think it is justifiable when smokers are not hired for certain jobs. I just think you should judge people on their qualities and not on whether they smoke or not’ (21-year-old, male smoker).
‘…A lot of people don't like being close to a smoker, but that doesn't mean that it is then also OK to treat them negatively because they smoke’ (18-year-old, female smoker).
‘…If people don't appreciate smokers, that's their problem’ (36-year-old, male smoker).
‘… Just as smokers have to be considerate of non-smokers, so do non-smokers have to respect smokers needs. That's why it is not OK just to be treated badly because I'm a smoker’ (28-year-old, female smoker).
‘…Not understandable. Smoking is an addiction and we are the victims of it but that doesn't mean others should be allowed to judge us’ (19-year-old, female smoker).
‘…I think it's not fair to stay away from us just because we smoke’ (22-year-old, female smoker).
Legitimate Discrimination Condition:
‘…I think it is understandable when smokers are not hired for jobs where being healthy and looking fit is important. I'm thinking of instructor jobs in health-clubs or something like that’ (21-year-old, male smoker).
‘…I can understand non-smokers who are disgusted by having to stand in other people's smoke and just don't want to put up with it’ (18-year-old, female smoker).
‘…You know when you smoke that not everyone will appreciate it and that is just something you have to accept’ (36-year-old, male smoker).
‘… As a smoker, you just can't imagine how off-putting it is for a non-smoker to see someone light up’ (28-year-old, female smoker).
‘…In a way it is understandable. Smoking is an addiction and we are the victims of it’ (19-year-old, female smoker).
‘…I think it's fair when people don't like to smoke passively that they want to stay away from smokers’ (22-year-old, female smoker).