The Importance of In-group Morality

Although previous research has focused on competence and sociability as the characteristics most important to positive group evaluation, we suggest that morality is more important. Studies with pre-existing and experimentally-created in-groups showed a set of positive traits to constitute distinct factors of morality, competence, and sociability. When asked directly, Study 1 participants reported that their in-group’s morality was more important than its competence or sociability. An unobtrusive Factor Analytic method also showed morality to be a more important explanation of positive in-group evaluation than competence or sociability. Experimental manipulations of morality and competence (Study 4) and morality and sociability (Study 5), showed that only in-group morality affected aspects of the grouplevel self-concept related to positive evaluation (i.e., pride in, and distancing from, the ingroup). Consistent with this, identification with experimentally-created (Study 2b) and preexisting (Studies 4 and 5) in-groups predicted the ascription of morality to the in-group, but not competence or sociability. The importance of in-group morality 3 Group virtue: The importance of morality (vs. competence and sociability) in the positive evaluation of in-groups Both Western and Eastern metaphysics specify the characteristics that constitute people’s greatest virtues and vices (for discussions, see MacIntyre, 1984; Nietzsche, 1887/1967; Shweder, Much, Mahapatra, & Park, 1997). For example, Aristotle listed 14 human virtues in his Ethics, including “righteous indignation,” “the just,” “sincerity,” “wisdom” and “friendliness” (Robinson, 1989). According to Aristotle, people who display these characteristics should be evaluated positively, as good and virtuous people. Contemporary research on the characteristics used to positively evaluate groups has tended to focus on two of the virtues mentioned by Aristotle -competence (i.e., wisdom) and sociability (i.e., friendliness). This emphasis of competence and sociability has meant that little attention has been given to the characteristic that Aristotle, and most metaphysics, considers most important to the evaluation of human virtue – morality (see MacIntyre, 1984; Nietzsche, 1887/1967). Metaphysics is not the only branch of thought that considers morality the most important of virtues. De Waal’s (1996) comparative studies of primates have led him to argue that the complex social interaction required for survival within the highly interdependent groups in which humans (as well as Chimpanzees and Bonobos) live requires that individual behavior be guided by a sense of morality. In quite different work, Schwartz (1992) has shown that people across the world consider moral values among the most important “guiding principles” in their lives. Thus, this paper extends previous work on the role of competence and sociability in individuals’ positive evaluation of in-groups by also examining the role of morality. In five studies, we use different methods to compare the importance of morality, competence, and sociability in the positive evaluation of pre-existing (Studies 1, 3, 4 and 5) and experimentally created (Studies 2a and 2b) in-groups. The importance of in-group morality 4 The Basis of Positive Evaluation A wide variety of research has shown that the single traits people ascribe to themselves and others (e.g., intelligent, skilled, competent) tend to cluster together to describe a specific human characteristic (i.e., competence; Anderson & Sedikides, 1991; Osgood, Suci, and Tannenbaum, 1957; White, 1980; Williams & Best, 1982). In turn, these characteristics (e.g., competence, sociability, morality, strength, prestige) tend to fall along two more general dimensions of positive evaluation: power (i.e., agency, dominance, dynamism) and benevolence (i.e., communion, nurturance, favorability; for reviews, see Leach, 2006; Osgood, 1971; White, 1980; Williams & Best, 1982). For example, Rosenberg, Nelson, and Vivekananthan (1968) showed that the single personality traits ascribed to individuals tended to coalesce into specific clusters indicating the more abstract human characteristics of competence (i.e., “skillful,” “industrious,” “intelligent”), sociability (i.e., “good-natured,” “happy,” “warm,” “sociable”), and morality (i.e., “honest,” “sincere”), among others. The competence cluster fell along one, more general, dimension described as “intellectual goodness,” whereas the sociability and morality clusters fell along a second, more general, dimension described as “social goodness.” Competence and Sociability Although a wide variety of human characteristics may be used to evaluate groups (Brewer & Campbell, 1979; Osgood et al., 1957; Williams & Best, 1982; for a review, see Leach, 2006), most contemporary research tends to examine only competence and sociability. For example, a wide variety of research shows that groups that are more successful than others tend to be ascribed high competence (e.g., Conway et al., 1996; Eagly & Steffen, 1984; Fiske et al., 2002; Judd et al., 2005; Poppe & Linssen, 1999; Ridgeway, Boyle, Kuipers, & Robinson, 1998). In contrast, groups that are less successful tend to be ascribed moderate sociability. Theories of social role (Eagly & Steffen, 1984), status-expectation The importance of in-group morality 5 (Ridgeway et al., 1998), interdependence (e.g., Alexander, Brewer, & Herrmann, 1999), social dominance (Sidanius & Pratto, 1999), and system justification (Kay & Jost, 2003), all suggest that this differential ascription of characteristics serves to legitimate differential success by providing each group a distinct basis of positive evaluation (see also Fiske et al., 2002). More specifically, the functionalist perspective at the heart of these theories suggests that competence and sociability enable more and less successful groups to evaluate themselves positively if they characterize themselves in a way that legitimates their level of success. It is argued that a more successful in-group can feel good about itself because of its competence whereas a less successful in-group can feel good about itself because of its sociability. Despite the popularity of the functionalist perspective, few studies have directly asked people how important competence and sociability are for their in-groups (as we do in Study 1). More critically, little work has empirically examined the importance of competence and sociability in individuals’ positive evaluation of their in-groups. It has simply been presumed that those characteristics that individuals most ascribe to their in-group are most important to individuals’ positive evaluation of the in-group and to individuals’ (group-level) self-concept. However, the degree to which traits are ascribed to an in-group may be influenced by a number of contextual concerns that have little to do with the psychological importance these traits have for in-group membership (e.g., Judd et al., 2005; van Knippenberg, 1978). For example, success at a task may lead individuals to ascribe more competence to their in-group without necessarily making competence more important to the positive evaluation of the ingroup (for a discussion, see Spears, Jetten, & Doosje, 2001). Thus, rather than assuming that the degree to which characteristics are ascribed to an in-group is an indication of their importance, the present studies use both direct and unobtrusive methods to actually assess the importance of competence and sociability in individuals’ positive evaluation of their inThe importance of in-group morality 6 group. As the vast majority of research on the ascription of characteristics to groups has not examined morality, we know very little about its importance to positive evaluation. In addition, the importance of morality is unclear because the few studies that aimed to address morality conflated it with other characteristics, such as dominance (Phallet and Poppe, 1997) or sociability (Fiske et al., 2002; Jost & Kay, 2005; see also Wojciszke, 1994). However, a group may be moral (i.e., honest and trustworthy) without necessarily being sociable (i.e., warm and friendly), and vice versa. For example, a moral in-group may be honest about an out-group’s failings, even where this means that they are being less warm and friendly toward the out-group. And, a sociable in-group may be warm and friendly in a disingenuous (and thus immoral) attempt to win friends and influence people. Although morality and sociability can be seen as falling along the same general dimension of evaluation (i.e., benevolence), they are conceptually distinct characteristics. Indeed, research at the individual level shows people to treat personality traits related to morality as distinct from traits related to sociability (e.g., Anderson & Sedikides, 1991; Rosenberg et al, 1968; White, 1980). There is no reason this should not also be the case at the group level. Thus, in several studies, we establish the validity of distinct, trait-based, measures of in-group morality, competence, and sociability. If morality is a distinct in-group characteristic, it should play a distinct role in the positive evaluation of in-groups. The Importance of Morality The social identity tradition assumes that any characteristic ascribed to an in-group can provide a basis for positive evaluation (see Oakes, Haslam, & Turner, 1994; Turner, Hogg, Oakes, Reicher, & Wetherell, 1987). Thus, at a theoretical level, the social identity tradition allows for morality to be the most important characteristic in the positive evaluation of in-groups. However, most empirical work within the social identity tradition has examined The importance of in-group morality 7 competence as the basis of positive in-group evaluation (for reviews, see Bettencourt, Dorr, Charlton, & Hume, 2001; Ellemers, 1993; Spears et al., 2001). Thus, research in the social identity tradition tends to share the functionalist approach in other work on competence and sociability in group evaluation. More specifically, in the social identity tradition, the sense of competence thought to follow from an in-group’s intellectual or economic success is presumed to enable both the positive evaluation of the in-group (i.e., “in-group favoritism”) and the negative evaluation of a less successful out-group (i.e., “out-group derogation”). As such, in-group competence is widely considered “status-defining” – the basis of a status distinction that favors the in-group over an out-group (for reviews, see Bettencourt et al., 2001; Spears et al., 2001). Given that competence is presumed to be the most important basis of positive ingroup evaluation, the social identity tradition suggests that in-groups lacking in competence must use “alternative” characteristics to evaluate themselves (Lemaine, 1974). For example, Tajfel and Turner (1979) argued that low status in-groups may evaluate themselves positively by emphasizing their artistry, rather than their (low) competence. In this way, low status groups are expected to exercise “social creativity” to replace competence as a basis of their collective virtue. Thus, some studies in the social identity tradition have interpreted unsuccessful groups ascribing relatively high morality to their in-group as the use of morality as an “alternative” basis of positive evaluation (e.g., Blanz, Mummendey, & Otten, 1995; Ellemers & Van Rijswijk, 1997). However, these studies did not directly ask participants how much importance they gave to their in-group’s morality, competence, or sociability. Neither did these studies actually examine the importance of in-group morality, competence, and sociability in the positive evaluation of the in-group. Thus, there is little evidence that competence is especially important to the positive evaluation of in-groups, or that morality is an “alternative” characteristic that only becomes important when it is difficult to positively The importance of in-group morality 8 evaluate the in-group based on its competence. Despite the theoretical possibility in the social identity tradition that morality can be most important to positive in-group evaluation, most research is based on the presumption that competence is most important. Although we undertook the present studies because there has been little examination of the importance of in-group morality, some work is broadly consistent with the notion that morality is of special importance to people. For example, at the individual level, Rodriguez Mosquera, Manstead, and Fischer (2002) found university students in Spain and the Netherlands to both show a very strong belief that being dishonest and untrustworthy would make them feel bad about themselves. Consistent with this, Schwartz and colleague’s series of cross-national studies shows individuals to consider moral values among the most important “guiding principle” in their lives (for a review, see Schwartz, 1992). Even the most success-oriented sample, university students in the United States, rated values related to morality as more important than values related to competence (Schwartz & Bardi, 2001). At the group level, De Waal’s (1996) comparative studies of primate behavior led him to argue that a sense of morality is essential to the group life of primates such as humans. Without a sense of morality in the group, he argued, it would be difficult for members to coordinate their behavior in ways that maximize benefits for themselves and the group as a whole. Consistent with this, recent models of procedural justice suggest that the perceived “trustworthiness” of an in-group is a major determinant of individuals’ positive evaluation of the group and their willingness to cooperate with other members (for reviews, see Skitka, 2003; Tyler & Blader, 2003). This may be why Campbell and colleague’s large-scale, multimethod studies of ethnocentrism throughout the world found morality (i.e., honesty, trustworthiness) to be central to how individuals evaluated their ethnic in-group relative to out-groups (Brewer & Campbell, 1979; Levine & Campbell, 1972). Thus, there appears to be good reason to expect that morality is of central importance to the positive evaluation of inThe importance of in-group morality 9 groups and to the group-level self-concept. Although morality has been given little place in contemporary Western psychology, the broader tradition of Western thought has long viewed morality as the most important of human virtues (for discussions, see MacIntyre, 1984; Nietzsche, 1887/1967; Shweder et al., 1997). The classical view, perhaps best represented by Aristotle’s Ethics, sees morality as the most important basis upon which people should evaluate themselves, and others, positively (MacIntyre, 1984). Indeed, Aristotle argued that all other virtues were, to some degree, dependent on morality. According to Aristotle, competence and sociability are only virtuous if they are expressed through a moral character that makes them sincere and trustworthy (MacIntyre, 1984). This suggests that morality is most important to positive self-evaluation because determines the meaning of every other human characteristic (see Osgood et al., 1957). For example, a political party that is competent in administration or sociable in campaigning is most assured of its goodness if it is also trustworthy. Competence or sociability in an immoral party gives little basis for positive self-evaluation. Indeed, immoral competence is dangerous and immoral sociability is disingenuous (Osgood et al., 1957; Wojciszke, 1994). Overview No prior research of which we are aware has examined the importance of morality to individuals’ positive evaluation of their in-group. Although a great deal of work has examined the mean-level ascription of competence and sociability to groups, this work has tended to either exclude morality or to assume that the importance of a characteristic is indicated by the degree to which it is ascribed to a group. Rarely has the actual importance of morality, competence, and sociability to the positive evaluation of in-groups been examined. In an important first step, Studies 1 and 2 use Confirmatory Factor Analysis (CFA) to validate trait measures of morality, sociability, and competence as distinct in-group The importance of in-group morality 10 characteristics. Study 1 directly asks individuals to explicitly indicate whether morality, competence, or sociability is the characteristic most important for their in-group to possess. To complement this approach, Studies 1-3 use Exploratory Factor Analysis (EFA) to unobtrusively examine the importance of morality, competence, and sociability in participant’s positive evaluation of their in-groups. EFA shows which of the three characteristics explains the most variance common to the positive traits ascribed to the ingroup. In Study 3, we also examine relative group success as a potential moderator of the importance of morality, competence, and sociability to in-group evaluation. If morality is most important, it should remain so irrespective of the in-group’s success relative to an outgroup. In addition, morality should not be most important to the positive evaluation of an outgroup. Thus, we also examine the importance of morality, competence, and sociability to outgroup evaluation. Studies 4 and 5 take an experimental approach. Here we manipulate the morality, competence, and sociability of a pre-existing in-group to examine effects on two aspects of the group-level self-concept related to positive evaluation: group pride and perceived group variability. If morality is most important to individual’s positive evaluation of their in-group, then high morality should lead to greater group pride. Conversely, low morality should lead individuals to distance themselves from the in-group. Study 4 compares the effects of ingroup morality and competence whereas Study 5 compares the effects of in-group morality and sociability. To offer yet another approach of examining the importance of in-group morality, competence, and sociability, Studies 2b, 4 and 5 assess participants’ identification with their in-group. If morality is central to the group-level self-concept, then in-group identification should lead individuals to ascribe morality to their in-group. In contrast, in-group identification should be less predictive of ascriptions of competence or sociability to the inThe importance of in-group morality 11 group. Study 1 We use this study to offer evidence of the construct validity of trait measures of ingroup morality, competence, and sociability. First, we used CFA to examine the latent structure of the positive traits designed to indicate these three group characteristics. Second, we included a number of traits to indicate the general positive valence (“virtue”, “value”, “worth”) of the in-group. By correlating this measure of valence to the three in-group characteristics, we aimed to establish that morality, competence and sociability are equally positive characteristics. Third, we correlated each of the three group characteristics to a set of closely related traits. By showing that in-group morality, competence, and sociability are only correlated to those traits that are conceptually similar we aimed to demonstrate the convergent and divergent validity of the three group characteristics. As we conceptualize ingroup morality as an in-group’s collective “correctness,” we expect morality to be most associated with the evaluation of the in-group as “good” and “correct,” rather than “bad” and “wrong” (see Osgood et al., 1957). As suggested by Eagly and Steffen (1984), Conway et al. (1987), and Poppe and Linnsen (1999), we expect in-group competence to be most associated with achievement (e.g., “successful,” “prestigious”). And, based in Rosenberg et al. (1968) and Osgood et al. (1957), we expected in-group sociability to be most associated with communality (e.g., “kind,” “helpful”). This study uses both a direct and an unobtrusive method to assess the importance of in-group morality, competence, and sociability. In a direct method, we ask participants to explicitly indicate how important it is for their in-group to possess traits indicating morality, competence, and sociability. If morality is as important as we suggest, then individuals should recognize it as such. We also used the unobtrusive method of EFA to assess the importance given morality, competence, and sociability in the positive evaluation of the inThe importance of in-group morality 12 group. EFA is based on the common factor model (for discussions, see Fabrigar, Wegener, MacAllum, & Strahan, 1999; Tabachnich & Fidell, 1996). As such, EFA examines the variance that the positive in-group traits have in common with each other in search of factors that account for the covariance between these traits. Each factor explains a proportion of the common variance in the positive traits (see Tabachnich & Fidell, 1996). Thus, we use EFA to examine to what degree morality, competence, and sociability actually explain the common variance of participants’ positive evaluation of their in-group. Famously, Osgood et al. (1957) used a similar approach to show that favorability tended to be a more important factor than strength and activity in the evaluation of people and other entities. Here, we expect that morality explains more of the common variance in positive in-group evaluation than competence or sociability. EFA serves as an unobtrusive method of examining the importance of morality, competence, and sociability in positive in-group evaluation because (1) it does not rely on the mean level at which traits are ascribed to the in-group, and (2) it does not require us to directly ask individuals to explicitly state the importance they give to in-group traits. Method Participants and Procedure Eighty-four male and female students at Leiden University in the Netherlands received partial course credit for their participation (Mage = 21, SD = 4.29). Participants were brought into a “lab” to complete a multi-part questionnaire. In the first part of the questionnaire, we asked participants to indicate how important it was for the groups to which they belonged to possess a list of traits. Measures Drawing on the literature summarized above, and the more extensive review by Leach (2006), we provided participants with nine traits designed to indicate group morality, The importance of in-group morality 13 competence, and sociability. These traits were interspersed with the other traits described below. Responses were given on a seven-point scale, anchored by 1 (not very important) and 7 (very important). In an improvement on previous research at the group level, morality (α = .80) was assessed with the traits “honest,” “sincere,” and “trustworthy” (see also Anderson & Sedikides, 1991; Osgood et al., 1957; Rosenberg et al., 1968). As in a good deal of stereotype research, sociability was assessed with the traits “likeable,” “warm,” and “friendly” (α = .69) whereas competence was assessed with the traits “competent,” “intelligent,” and “skilled” (α = .86). Based on Leach (2006), we used 16 other traits to assess valence (virtuous, valuable, constructive, worthless[reversed], α = .57) as well as correctness (good, right, bad[reversed], wrong[reversed], α = .66); communality (benevolent, kind, helpful, safe, α = .69); and achievement (superior, successful, prestigious, ambitious, α = .79). These characteristics were used to validate our measures of in-group morality, competence, and sociability. Results Confirmatory Factor Analysis As a first step, we examined our suggestion that morality, competence, and sociability are distinct in-group characteristics. Thus, we submitted the nine items assessing participant’s explicit statements of the importance of these in-group characteristics to a CFA using maximum likelihood estimation. Each item was specified to indicate only its hypothesized factor and no errors were allowed to correlate. However, the three hypothesized latent factors of morality, competence, and sociability were allowed to correlate with each other. The hypothesized three-factor measurement model fit the data well (for a review of the indices used below, see Hu & Bentler, 1999). The model adequately reproduced the observed covariance matrix, as the Chi-square statistic was small and non-reliable, χ (24) = 28.59, p = .24. However, this is not surprising given the small sample size. More telling is The importance of in-group morality 14 that the χ/df ratio (= 1.19) fell below 2.5. Good model fit was also suggested by a wide variety of fit (NNFI = .975, CFI = .983, IFI = 984, GFI = .940) and residual (SRMR = .064, RMSEA = .048) indices. In addition, the factor loadings for morality (.51-.90), competence (.76-.93), and sociability (.60-.77) were all sizeable and statistically reliable (all p < .05). The inter-correlations between factors were low (i.e., .02, .29, .30). Further analysis showed the hypothesized three-factor measurement model to fit better than more parsimonious alternatives. Thus, a two-factor model that specified a competence factor and a combined morality-sociability factor fit less well than the threefactor model, Δχ (2) = 32.89, p < .005. This two-factor model also fit poorly in an absolute sense: NNFI = .821, CFI = .871, IFI = .875, GFI = .873 and SRMR = .107, RMSEA = .127. In addition, a two-factor model that specified a morality factor and a combined competencesociability factor fit less well than the three-factor model, Δχ (2) = 113.49, p < .005. This model also fit poorly in an absolute sense: NNFI = .414, CFI = .577, IFI = .592, GFI = .739 and SRMR = .169, RMSEA = .231. Lastly, a one-factor model of general positive in-group evaluation fit less well than the hypothesized three-factor model, Δχ (3) = 146.65, p < .005. This one-factor model also fit poorly in an absolute sense: NNFI = .279, CFI = .459, IFI = .477, GFI = .695 and SRMR = .189, RMSEA = .256. Construct Validity To further examine the construct validity of our measures of in-group morality, competence, and sociability, we computed (zero-order and partial) correlations between these characteristics and the four others we included. Results are shown in Table 1. Consistent with our conceptualization of morality, competence, and sociability as equally positive characteristics, they were similarly correlated to the individual items as well as the composite scale of valence. As expected, in-group morality was uniquely correlated with correctness whereas in-group competence was uniquely correlated with achievement. Also as expected, The importance of in-group morality 15 in-group sociability was uniquely correlated with communality. The Importance of Morality, Competence, and Sociability We examined the importance of morality, competence, and sociability with two different methods. To unobtrusively examine the importance of the morality, competence, and sociability factors, we submitted the nine traits to a Principal-Axis FA (with maximum likelihood estimation and Oblimin rotation). As oblique rotation tends to equalize the common variance explained by each factor, we report the percent of variance each factor explains upon initial extraction (see Tabachnich & Fidell, 1996). Examination of the common variance each factor explained upon initial extraction showed morality to be most important (i.e., 32.98%). Competence explained 15.64% of the common variance whereas sociability explained 6.95%. This three-factor solution fit the data well -χ (12) = 6.13, p = .91, χ/df = .511, and accounted for a majority of the common variance in the items (55.57%). We also used a direct method to examine the importance participants explicitly gave to the morality, competence, and sociability of their in-groups. We asked participants to indicate the importance of each group trait in an absolute fashion. Paired samples t-tests showed that participants viewed the traits constituting morality (M = 6.39, SD = .65) as more important for their in-group than competence (M = 5.10, SD = .99), t (84) = 11.28, p < .001, d = 1.22, or sociability (M = 6.11, SD = .69), t (84) = 3.19, p = .002, d = .35. In addition, ingroup competence was deemed less important than sociability, t (84) = 7.86, p < .001, d = .85. Discussion This study offered initial support for our suggestion that morality, competence, and sociability are distinct in-group characteristics. We used CFA to show that morality, competence, and sociability factors account well for the associations between a set of positive traits designed to indicate these characteristics. More parsimonious alternatives did not The importance of in-group morality 16 account for the latent structure in the in-group traits as well as these three distinct factors, which were only modestly correlated. The validity of these morality, competence, and sociability factors was further suggested by their correlations with other in-group traits. Morality was uniquely associated with correctness, competence was uniquely associated with achievement, and sociability was uniquely associated with communality. However, in-group morality, competence, and sociability, were equally positive in valence. This study used two different approaches to show that in-group morality is more important to individuals than is competence or sociability. When individuals were directly asked about the importance of traits indicating morality, competence, and sociability, they reported that morality traits were the most important for their in-groups to possess. In contrast to what is suggested by previous theory, competence was deemed the least important characteristic for participant’s in-groups to possess. This direct method of asking participants to report the importance of morality, competence, and sociability was complemented by an unobtrusive approach. EFA showed the morality factor to explain twice the common variance explained by competence and over four times the common variance explained by sociability. Thus, morality was the characteristic most important to the participants’ positive evaluation of their in-groups. Although a promising first step, Study 1 did not force participants to focus on one specific in-group. Indeed, participants were likely to have thought of quite different in-groups when indicating the importance of morality, competence, and sociability. Thus, Studies 2a and 2b sought to corroborate and extend the present results by assigning participants to an experimentally-created in-group with which participants had no prior experience. This allowed us a greater degree of methodological control by having participants judge a single in-group that we represented in the same way to all participants. In addition, Studies 2a and 2b sought to extend Study 1 by having participants ascribe traits indicating morality, The importance of in-group morality 17 competence, and sociability to the experimentally-created in-group. As this kind of trait ascription is typical of previous studies of group competence and sociability, Studies 2a and 2b aimed to be more comparable to previous work. However, unlike previous work, we actually gauge the importance of morality, competence, and sociability in the positive evaluation of experimentally-created in-groups. As in Study 1, we utilize the unobtrusive method of EFA to accomplish this. Study 2 Two, highly similar, studies were designed to pursue three main aims. First, we use CFA to show that the nine traits we designed to indicate group morality, sociability, and competence constitute three distinct factors. Unlike Study 1, these studies ask participants to ascribe positive traits to an experimentally created in-group with which they had no prior experience. Second, we use the unobtrusive method of EFA to examine the actual importance of the morality, sociability, and competence factors in explaining the variance common to the positive traits ascribed to the in-group. As in Study 1, we expect morality to be more important than competence or sociability to the positive evaluation of the in-group. The third aim of Study 2 is to provide yet another method for assessing the importance of morality, competence, and sociability to in-group membership. Thus, we assess the degree to which identification with an in-group predicts the degree to which individuals ascribe morality, competence, and sociability to this in-group. As group identification is strongly associated with the positive evaluation of the in-group, and with the incorporation of the in-group into the self-concept (e.g., Leach et al., 2006; for reviews, see Ashmore, Deaux, & McLaughlin-Volpe, 2004; Ellemers et al., 1999), we expect in-group identification to be an especially good predictor of participant’s ascription of morality to their in-group. In-group identification should be less predictive of the ascription of competence or sociability to the in-group, as these characteristics should be less important to positive inThe importance of in-group morality 18 group evaluation and the group-level self-concept. With Study 2b we also aim to further corroborate the results of Studies 1 and 2a by working in a different national context. Method Participants Study 2a. Participants were 105 (72 women, 33 men) psychology students at the University of Amsterdam, the Netherlands (Mage = 21, SD = 4.29). They participated for course credit. Study 2b. Participants were 158 (119 women, 39 men) students at the University of Sussex in England (Mage = 21, SD = 4.49). They participated for course credit or payment (of 5 British pounds). Procedure Both studies were embedded within experiments focused on social comparison within an experimentally created in-group (i.e., Vliek, Leach, & Spears, 2006). Participants were brought to a classroom and seated in separate cubicles for a computer-administered study. The computer introduced the study as an “exploration into different kinds of perception.” Participants were told that research had revealed that people could be characterized as having one of two perceptual styles: “global” and “detailed.” The ostensible goal of the experiment was to discover whether these styles of visual perception mattered for performance on the kind of perception-based tasks often used to measure intelligence. Admittedly, this oft-used cover story is more relevant to in-group competence than morality or sociability. Never the less, we still expected morality to be most important to participants’ positive evaluation of the in-group. In order to determine their perceptual style, participants completed several trials of two different tasks: a dot-estimation task and a dice-recognition task. Upon completion, the computer ostensibly calculated the participants’ score on the tasks. Presumably based on this The importance of in-group morality 19 score, the computer categorized each participant as a “detailed perceiver.” Participants were then presented with a list of positive traits and asked to indicate to what degree each characterized detailed perceivers. Responses were given on a seven-point scale, anchored by 1 (not all) and 7 (very much). The traits, including the nine designed to indicate morality, competence, and sociability, were presented in a random order. In Study 2b, participant’s group identification was measured immediately after in-group categorization. Measures In Study 2a, the scales of morality (α = .79; M = 4.85, SD = .89), sociability (α = .76; M = 4.57, SD = .89), and competence (α = .72; M = 5.35, SD = .72) used in Study 1 all proved reliable. The scales of in-group morality (α = .90; M = 4.81, SD = .96) competence (α = .88; M = 5.15, SD = .91), and sociability (α = .85; M = 4.57, SD = .86) also proved reliable in Study 2b. Using a seven-point Likert-type response scale, Study 2b participants were asked to what degree they identified as a “detailed perceiver” with 4 items (α = .75) from the scale developed by Ellemers et al. (1999). Thus, we asked participants to what degree they “identified with” detailed perceivers, whether they were “glad to be categorized as a detailed perceiver,” whether they felt “strong ties” with detailed perceivers, and whether they “saw themselves as” a detailed perceiver. Results Confirmatory Factor Analysis The nine traits designed to indicate in-group morality, competence, and sociability were submitted to a Confirmatory Factor Analysis with maximum likelihood estimation parallel to that of Study 1. As Studies 2a and 2b were highly similar in method, we examined the combined data in a multi-group analysis. In essence, this approach estimated one measurement model that constrained each item’s loading on its hypothesized factor to be The importance of in-group morality 20 equal across the two studies. As in Study 1, items were specified as loading on only on factor and no errors were allowed to correlate. However, the three latent factors corresponding to morality, competence, and sociability were allowed to correlate. The three-factor model fit the data well, χ (54) = 110.88, p < .001, χ/df = 2.05. This was also shown in a variety of fit (NNFI = .963, CFI = .972, IFI = .972, GFI = .919) and residual (SRMR = .042, RMSEA = .064) indices. A model specifying a combined moralitysociability factor and a competence factor, fit less well than the hypothesized model, Δχ (4) = 176.76, p < .001. This alternative model also fit poorly in an absolute sense, as shown in a variety of fit (NNFI = .860, CFI = .887, IFI = .888, GFI = .789) and residual (SRMR = .072, RMSEA = .123) indices. The Importance of Morality, Competence, and Sociability As in Study 1, we used an EFA to unobtrusively examine the importance of morality, competence, and sociability in the positive traits ascribed to participant’s in-group. In Study 2a, the three-factor solution fit the data well, χ (12) = 19.79, p = .07, χ/df = 1.65, and accounted for a majority of the common variance (58%). As expected, morality was the first factor extracted before rotation. Thus, group morality explained the greatest amount of the common variance in the positive traits participant’s ascribed to their in-group (i.e., 33.34%). Sociability was the second factor extracted (explaining 15.45% of the common variance) and competence was the third factor extracted (explaining 9.62% of the common variance). In Study 2b, the three-factor solution fit the data satisfactorily, χ (12) = 28.74, p = .004, χ/df = 2.40, and accounted for a majority of the common variance (i.e., 72.46%). Morality was the first factor extracted before rotation and thus explained the greatest amount of the common variance (i.e., 55.29%). Competence (8.61%) and sociability (8.57%). explained near equal amounts of variance. In-group Identification The importance of in-group morality 21 In Study 2b, we performed a MANOVA, with in-group identification as a continuous predictor of participants’ ascription of morality, competence, and sociability to their in-group. We also included gender as a categorical predictor. As gender had no main or interaction effects (both p > .93, both partial η < .003) it is not discussed further. In-group identification had a multivariate effect on the ascription of the three characteristics to the in-group, F (3, 152) = 8.98, p < .001, partial η = .15. At the univariate level, in-group identification predicted the ascription of competence -F (1, 158) = 15.64, p < .001, partial η = .09, as well as sociability -F (1, 158) = 17.19, p < .001, partial η = .10. However, as predicted, in-group identification had the largest effect on participants ascription of morality to their in-group -F (1, 158) = 23.02, p < .001, partial η = .13. Analysis of partial correlations, accounting for all three group characteristics, showed in-group identification to uniquely predict in-group morality, partial r = .19, p = .02. However, partial correlations revealed that in-group identification did not uniquely predict in-group sociability (r = .12, p = .15) or competence (r = .06, p = .43). Discussion Studies 2a and 2b categorized participants as members of a quasi-minimal in-group. We then asked them to indicate to what degree a list of positive traits characterized this ingroup. As participants had no pre-conceived notion of this in-group or its characteristics, they were free to imbue the in-group with whichever traits they preferred. A CFA showed these traits to constitute distinct factors of morality, sociability, and competence. We used the unobtrusive method of EFA to show that the morality factor explained much more of the common variance in these positive traits than did competence or sociability. In these two ways, Studies 2a and 2b corroborate Study 1. That morality is the most important factor in the positive evaluation of the in-groups in these two studies is all the more impressive given that in-group membership was defined in terms of a perceptual style presumably linked to The importance of in-group morality 22 intelligence. Clearly the competence-oriented context of this study, and participants’ status as students, made morality fairly irrelevant to the in-group membership examined. Thus, Study 2 may be considered a conservative test of our hypotheses. Additionally, Study 2b showed that the more individuals identified with the in-group to which they were assigned, the more they ascribed morality to this in-group. Indeed, ingroup identification uniquely predicted in-group morality; In-group identification did not uniquely predict in-group competence or sociability. Thus, identification with an experimentally created in-group determined positive evaluation of this in-group only in terms of its morality. Despite these encouraging results, we thought it important to show that our hypotheses also apply to the ascription of traits to a real, pre-existing in-group. And, because much previous work suggests that relative group success affects the importance given to morality, competence, and sociability in the positive evaluation of in-groups, we thought it important to manipulate relative group success. These concerns guided the design of Study 3. Study 3 As outlined in the introduction, a functionalist perspective suggests that competence should be the characteristic most important to the positive evaluation of an in-group that is more successful than an out-group. In addition, a functionalist perspective suggests that either sociability or morality should be the characteristic most important to the positive evaluation of a less successful in-group, as competence is presumed unavailable as a basis of positive evaluation. However, as stated in the introduction, little previous work has examined the actual importance of morality, competence, and sociability in the positive evaluation of more or less successful in-groups. Thus, as shown in Studies 1 and 2, we expect that where ingroup morality is examined as a characteristic distinct from competence and sociability, morality should be most important to the positive evaluation of an in-group. In contrast to work that presumes that morality is an alternative way of positively evaluating a less The importance of in-group morality 23 successful in-group, we expect that morality will be most important to the positive evaluation of the in-group irrespective of the in-group’s success relative to an out-group. Study 3 also aimed to extend the previous two by having individuals ascribe positive traits to an out-group (relative to which the in-group was more or less successful). If morality is most important to the positive evaluation of in-groups because morality is central to the group-level self-concept, then morality must necessarily be less important in the positive evaluation of out-groups (as they are not part of the self). Thus, we suggest that competence and sociability should be more important than morality in the positive evaluation of outgroups. Our reasoning is consistent with what is implied in some research on group stereotypes. Studies that have utilized measures of sociability free of morality suggest that sociability is most important to the positive evaluation of less successful out-groups (e.g., Conway et al., 1996; Eagly & Steffen, 1984; Judd et al., 2005). When an out-group is more successful than the in-group, competence appears to be more central to the positive evaluation of the out-group (e.g., Conway et al., 1996; Eagly & Steffen, 1984; Fiske et al., 2002; Poppe & Linnsen, 1999; Ridgeway et al., 1998). However, these studies did not examine the actual importance of competence or sociability in the positive evaluation of more or less successful out-groups as we do here. Method Participants and Procedure One-hundred and twenty male and female students at Leiden University in the Netherlands received partial course credit for their participation. As in numerous studies of in-group bias, we manipulated relative in-group success by providing participants with false feedback about students at their university. Ostensibly on the basis of their reputation among employers, the cover page of the questionnaire reported that students at participants’ university had better (or worse) “general ability,” “social skills,” “international orientation,” The importance of in-group morality 24 and “communication skills,” than students at a relevant nearby university (i.e., the University of Amsterdam). This manipulation was designed to provide feedback about the in-group’s general success by referring to a number of behavioral domains relevant to student’s future career prospects. We included information regarding these different domains so that the ingroup’s relative success would be seen in general terms, rather than in the more specific terms of morality, competence, or sociability. However, these domains of group performance appear more relevant to competence and sociability than to morality, implying a stringent test for our contention that morality is most important in the positive evaluation of the in-group, regardless of its relative success. Immediately after the manipulation of relative group success, participants were asked to indicate to what degree a list of traits characterized students at their university. Responses were given on a seven-point scale, anchored by 1 (not all) and 7 (very much). The 9 items used in the preceding studies were included here to measure in-group morality (α = .86; M = 4.76, SD = .92), competence (α = .79; M = 5.36, SD = .76), and sociability (α = .77; M = 4.91, SD = .83). Participants were then asked to ascribe these same traits to the out-group: morality (α = .80; M = 4.60, SD = .73), competence (α = .82; M = 5.10, SD = .78), and sociability (α = .66; M = 4.69, SD = .70).