Back to Journal Cover Page


CURRENT RESEARCH IN SOCIAL PSYCHOLOGY


Volume10, Number 2
Submitted: July 24, 2004
First Revision: August 5, 2004
Second Revision: September 30, 2004

Accepted: October 1, 2004
Publication Date: October 1, 2004

SELF-DIRECTED NEGATIVE AFFECT: THE DISTINCT ROLES OF INGROUP IDENTIFICATION AND OUTGROUP DEROGATION

Sandro Costarelli
Rose Marie Callà
University of Trento, Italy

ABSTRACT

Prior research has shown that negative affect directed at the self follows from intergroup evaluation. Building on social identity theory, we hypothesized that ingroup identification impacts on felt self-directed negative affect after making intergroup differentiating evaluations. Additionally, prior work has demonstrated that people downgrade the outgroup to the extent that they identify with the ingroup. Accordingly, building on previous work on the so-called positive-negative asymmetry in intergroup discrimination, we expected that outgroup derogation but not ingroup favoritism would mediate the hypothesized affective impact of ingroup identification. Results indicated support for both predictions. These findings add to previous work by tracing the sources of self-directed negative affect in striving of group members for positive social identity through the process of establishing intergroup differentiation.

[13]
---------------
[14]

INTRODUCTION

Prior research has shown that expressing prejudice and discrimination (i.e., intergroup differentiation) is likely to result in self-directed negative affect to the extent that these responses conflict with non-prejudicial and -discriminatory personal standards (e.g., Devine, Monteith, Zuwerink, & Elliot, 1991; Monteith, 1993, 1996; Monteith, Devine, & Zuwerink, 1993). However, Mackie and Smith (1998) as well as Ho and Driscoll (1998) have recently argued that many of our personal standards may actually be very closely intertwined with more socially-based entities (e.g., motives, norms, roles; cf. Costarelli & Palmonari, 2003)). Accordingly, these authors have suggested that future research in this area should take into account the interplay between personal standards and these more socially-based entities that are conceptually relevant to intergroup evaluation. In keeping with this suggestion, the present investigation focuses on the potential moderating role played in this process by such an important social motive as ingroup identification (cf. Brown, 1995; Tajfel & Turner, 1986).

Self-Directed Negative Affect and Ingroup Identification

On the one hand, according to social identity theory (SIT: Tajfel, 1981; Tajfel & Turner, 1986), one of the key motives for individuals is that of maintaining a positive view of their own group memberships. As a consequence, group members higher in identification with the ingroup (henceforth, higher identifiers) are more likely to be motivated to establish ingroup favoring differentiation, compared to lower identifiers (Hogg & Abrams, 1988; but see Brown, 2000; Hinkle & Brown, 1990). On the other hand, according to self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), when individuals categorize themselves as group members, they internalize, and conform to, ingroup social norms. Indeed, any group-level response can be conceptualized as mediated by conformity to the content of relevant ingroup norms (Spears, Doosje, & Ellemers, 1997). Accordingly, violation of any of those standards that orient behavior such as currently enforced principles of non-discrimination should affect the self (Breckler & Greenwald, 1986). Support for this argument would be revealed, for example, by finding own violation of relevant group norms linked to the subsequent experience of negative affect. Indeed, being relatively temperate in the public expression of pro-ingroup biased intergroup evaluation and behavior has become the content of important normative prescriptions towards the so-called intergroup fairness. Consistent with this notion, van den Bos, Wilke, and Lind (1998) have found that fairness is a particularly salient issue precisely when people are concerned about potential problems that are associated with socially based identity processes.

Based on this reasoning, the strong motivation of higher identifiers to establish pro-ingroup differentiation should lead them to perceive their transgression of personal standards towards intergroup fairness as highly salient. In turn, for this group of participants, this should pose a threat to the overall integrity of self. Accordingly, in a first study, we predicted that higher identifiers would feel more negatively than lower identifiers after expressing pro-ingroup biased evaluations (Hypothesis 1).

Self-Directed Negative Affect and Outgroup Derogation

Recent research by Mummendey and her co-workers has demonstrated that group members display more often and overtly ingroup favorable (i.e. ingroup favoritism) rather than outgroup unfavorable evaluations and behaviors (i.e. outgroup derogation; for a review, see

[14]
---------------
[15]

Mummendey & Otten, 1998). These researchers report evidence that points to a normative explanation for this asymmetry effect in intergroup differentiation. Negative information is more salient than positive information (Fiske, 1980). Evaluation or behavior on negative dimensions is therefore more visible than on positive ones. As a result, this renders it less socially acceptable to differentiate the ingroup from a relevant outgroup on negative dimensions compared to positive ones (Hewstone, Fincham, & Jaspars, 1981; Mummendey & Otten, 1998). On these conceptual bases, in our experimental scenario outgroup derogation should be viewed as socially undesirable. In turn, this should elicit self-directed negative affect. However, in line with the predictions of SIT, recent research has demonstrated that people downgrade the outgroup to the extent that they identify with the ingroup (e.g., Branscombe & Wann, 1992, 1994; Mummendey, Brown, & Klink, 2001; Wann & Branscombe, 1995). Accordingly, we predicted that outgroup derogation but not ingroup favoritism would be the specific locus of intergroup differentiation (Brewer, 1979) mediating the effect of ingroup identification on the self-directed negative affect felt by the participants subsequent to making intergroup evaluations (Hypothesis 2).

In the current research, we investigated these processes in South-Tyrol, a multi-ethnic territory in northern-Italy. In this area, the German-speaking population group was discriminated against with respect to access to jobs in the public administration during the Fascist dictatorship. Since the end of the latter, such jobs have been allocated equally by law to members of the German- and Italian-speaking population groups. Over the years, among the members of both of these two groups this has created a strong feeling of negative interdependence. In turn, for Italian- and German-speaking South-Tyroleans, this has increased the identification with the respective groups of belongingness (see Capozza & Manganelli Rattazzi, 1999; Costarelli & Colloca, 2004 ; Kirchler & Zani, 1989). Therefore, it is for such cogent reasons that it seemed particularly appropriate to examine the joint affective impact of ingroup identification and intergroup differentiation on experienced negative affect in this specific intergroup context.

STUDY 1

METHOD

Participants and Procedure

We carried out Study 1 in the South-Tyrolean town of Bozen, Italy, in 2002. The participants were seventy-one higher school students (58 women, 13 men). All belonged to the Italian-speaking population group.

We asked the participants to give their perception of two linguistic groups, their own (i.e., Italian-speaking) and one other group (i.e., the German-speaking group) on a number of trait adjective items that appeared in random order on a list. We counterbalanced rating order of the target groups. We then asked the participants to answer some questions about their identification with the ingroup. Finally, they were asked to rate their affect based on how they felt after completing the intergroup evaluation task.

[15]
---------------
[16]

Design and Dependent Measures

The design was a 2 (Ingroup Identification: Lower vs. Higher) X 2 (Social Categorization of the target group: Ingroup vs. Outgroup) mixed factorial with the last factor as within-participants. (See Table 1.)

One important aim of the study was to assess whether ingroup favoritism and/or outgroup derogation (i.e. the loci of intergroup differentiation) would potentially mediate the effect of ingroup identification on the post-evaluation self-directed negative affect (see H2). To this end, evaluations of the ingroup and the outgroup were assessed on positive and negative attitude dimensions. We then operationalized the loci of intergroup differentiation as the difference score between the positive evaluation of the ingroup and the outgroup (ingroup favoritism) and as the difference score between the negative evaluation of the outgroup and the ingroup (outgroup derogation). Following prior relevant research by Kaplan (1972), we obtained the positive and negative evaluation ratings of the ingroup and the outgroup by splitting each of six typically bipolar semantic differential scales taken from a study by Mac Donald and Zanna (1998) into the respective two unipolar items. One unipolar item assessed the endorsement of the positive pole of the bipolar scale (e.g., Germans: are not at all likeable/extremely likeable). The other unipolar item measured the endorsement of the negative pole of the bipolar scale (e.g., Germans: are not at all dislikable/extremely dislikable). This operation resulted in twelve unipolar items differing in valence. The positively-valenced traits were: likeable, attractive, admirable. The negatively-valenced traits were: dislikeable, repulsive, contemptible. The participants rated all items on 6-point scales (0 = not at all, 5 = extremely) with no neutral point.

We assessed ingroup identification by administering to the participants a standardized ingroup identification scale (Doosje, Ellemers, & Spears, 1995, see Appendix). We derived measures of self-directed negative affect from ratings of their self-directed negative feelings after evaluating the target groups as assessed on 6-point scales (0 = does not apply at all, 5 = applies very much) with no neutral point. The negative affect measure we used comprised some of those negative affect items which in a study by Devine et al. (1991) loaded uniquely on the factor that they interpreted as reflecting negative feelings directed at the self. The items we employed were angry at myself, guilty, disappointed with myself, disgusted with myself, ashamed.

Table 1. Descriptive Statistics

M

SD

Cronbach Alpha

Ingroup identification

3.61

0.57

.83

Self-directed negative affect

3.07

0.63

.73

Ingroup positive evaluation

4.39

0.71

.82

Outgroup positive evaluation

3.66

0.66

.72

Ingroup negative evaluation

0.56

0.50

.72

Outgroup negative evaluation

1.58

0.59

.73

 

RESULTS AND DISCUSSION

First, we tested for main and interaction effects of both gender and target group rating order on the dependent variables used in the following analyses. We found these factors to produce no significant main or interactive effects. Therefore, we collapsed data across these factors in subsequent analyses.

[16]
---------------
[17]

Preliminary Analyses

For each target group, we constructed separate positive and negative evaluation indexes: A six-item Positive evaluation index, and a six-item Negative evaluation index. For each target group, we found positive and negative evaluations to be negatively correlated (rs: Ingroup = - .56; Outgroup = - .61). However, they were not completely reciprocal. Therefore, the separate exploration of ratings of the target groups as assessed on the positive and the negative items appeared methodologically justified.

Subsequently, we divided respondents by means of a median split (preliminary use of other division methods such as modal and tertile split resulted in uneven distributions of the participants). Accordingly, we classified those participants scoring above the median (= 4) as higher, and the rest as lower, on identification with the ingroup. A subsequent one-way analysis of variance (ANOVA) showed that these two groups of participants differed significantly on the aforementioned measure, F(1, 70) = 31.81, p < .0001, (Lower identifiers: M = 3.03; SD = 0.50; Higher Identifiers: M = 4.43; SD = 0.71).

Affect

First, we constructed a self-directed negative affect index. To this end, we conducted a Principal Components analysis (with Varimax rotation) to check whether the five affect items on which the participants rated their own affect could be clustered into a unique affect category. From this analysis a one-factorial solution emerged (total percentage variance explained = 82%; our loading criterion was .50 or higher). This supported our initial selection of the specific items that we subsequently employed to construct a self-directed negative affect index by averaging the ratings provided by the participants for all items.

A subsequent one-way ANOVA revealed that ingroup identification (Higher vs. Lower) impacted on the self-directed negative affect experienced by the participants after being involved with the intergroup evaluation task, F(1, 70) = 4.91, p < .031. As predicted under Hypothesis 1, higher identifiers experienced stronger self-directed negative affect than did lower identifiers (see Table 2).

Table 2. Self-Directed Negative Affect and Ingroup Favoritism as a Function of Ingroup Identification

Lower Identification

Higher identification

Self-directed negative affect

2.83
(0.60)

3.29
(0.72)

Ingroup positive evaluation

4.23
(062)

4.64
(067)

Outgroup positive evaluation

3.93
(0.60)

3.35
(0.62)

Ingroup negative evaluation

0.23
(0.62)

0.97
(0.40)

Outgroup negative evaluation

1.53
(0.60)

1.64
(0.76)

Note: Standard deviations are shown in parenthesis

[17]
---------------
[18]

Intergroup Differentiation

We investigated the effect of ingroup identification on the loci of intergroup differentiation by using a 2 X 2 X 2 mixed-model ANOVA: Ingroup Identification (Higher vs. Lower) was a between-participants variable, whereas Social Categorization of the target group (Ingroup vs. Outgroup) and Trait Valence (Positive vs. Negative) were within-participants variables. This analysis yielded a main effect of social categorization of the target group, F(1, 70) = 284.85, p < .001, a main effect of trait valence, F(1, 70) = 27.07, p < .001, a social categorization of the target group by trait valence interaction, F(3, 70) = 27.97, p < .001, and an ingroup identification by social categorization of the target group by trait valence interaction, F(7, 70) = 18.07, p < .001 (which is the effect of interest). Subsequent contrast analyses indicated that on the positively-valenced trait items both lower (t(35) = 2.34, p = .027) and higher identifiers (t(36) = 6.84, p < .001) established intergroup differentiation by evaluating the ingroup more favorably than the outgroup (i.e., through ingroup favoritism)). However, as shown in Table 2, simple-interaction effect F tests (Jaccard, 1976) revealed that the on the positively-valenced trait items higher identifiers showed significantly more ingroup favoritism than did lower identifiers, F(1, 70) = 12.51, p < .001.

As shown in Table 2, on the negatively-valenced trait items higher identifiers established intergroup differentiation by derogating the outgroup more than the ingroup (i.e., through outgroup derogation; t(36) = -3.88, p < .01). By contrast, this analysis revealed that lower identifiers expressed no differential negative ratings of the two target groups, t(35) < 1, p > 1.

Mediational Analysis

We tested whether ingroup favoritism and outgroup derogation mediated the impact of ingroup identification on self-directed negative affect. The previous analyses indicated that higher identifiers expressed both greater ingroup favoritism and outgroup derogation as well as experienced stronger self-directed negative affect, compared with lower identifiers. It would seem likely that those participants higher in ingroup identification upgraded the ingroup or downgraded most greatly the outgroup. In turn, this would generate the stronger self-directed negative affect they experienced, relative to lower identifiers. Therefore, next we tested whether this effect of ingroup identification on self-directed negative affect is indeed mediated by ingroup favoritism- and/or outgroup derogation-based intergroup differentiation. To this end, first, we computed two distinct indexes of intergroup differentiation based on the positively-valenced rating items (i.e., ingroup favoritism) and on the negatively-valenced rating items (i.e., outgroup derogation) by subtracting the positive evaluations of the outgroup from those of the ingroup, and the negative evaluations of the ingroup from those of the outgroup, respectively.

Then, following Baron and Kenny (1986) and Judd & Kenny (1981), to test whether the effect of ingroup identification is mediated by ingroup favoritism- and/or outgroup derogation-based intergroup differentiation, we conducted two distinct mediational analyses. In one mediational analysis, first, we regressed self-directed negative affect on ingroup identification. Secondly, we regressed ingroup favoritism on ingroup identification. Thirdly, we regressed self-directed negative affect on ingroup identification and ingroup favoritism. Contrary to what would be expected if there is mediation, the effects of ingroup identification on self-directed negative affect, (Beta = .26, t(71) = 2.21, p < .05) and of ingroup identification on ingroup favoritism, (Beta = .42, t(71) = 3.20, p < .05) were significant in the first two regression equations, but the effect of ingroup favoritism was not significant in the

[18]
---------------
[19]

third regression equation, (Beta = .07, t(71) = .07, p > 1, n.s.). Indeed, the effect of ingroup identification on self-directed negative affect was significant in the third regression equation, (Beta = .25, t(71) = 2.32, p < .05, Rsquared = .28), where we controlled for the effect of ingroup favoritism by entering it simultaneously in the regression model as a predictor. Thus, this analysis showed that ingroup favoritism-based intergroup differentiation did not mediate the effect of ingroup identification on self-directed negative affect.

In a second mediational analysis we conducted, first, we regressed self-directed negative affect on ingroup identification. Secondly, we regressed outgroup derogation on ingroup identification. Thirdly, we regressed self-directed negative affect on both ingroup identification and outgroup derogation. As would be expected if there is mediation, the effects of ingroup identification on self-directed negative affect, (Beta = .26, t(71) = 2.21, p < .05) and of ingroup identification on outgroup derogation, (Beta = .33, t(71) = 2.55, p < .05) were significant in the first two regression equations, and the effect of outgroup derogation was significant in the third regression equation, (Beta = .29, t(71) = 2.41, p < .05, Rsquared = .30, Rsquared change: F(1, 71) = 4.00, p = .042). Indeed, the effect of ingroup identification on self-directed negative affect was not significant in the third regression equation (Beta = .14, t(71) = 0.97, p > 1, n.s.), where we controlled for the effect of outgroup derogation by entering it simultaneously in the regression model as a predictor.

Thus, in line with evidence from recent research (e.g., Costarelli & Colloca, 2004), this analysis showed that outgroup derogation- but not ingroup favoritism-based intergroup differentiation mediated the effect of ingroup identification on self-directed negative affect, thus confirming our Hypothesis 2.

STUDY 2

We conducted a second study to attempt to replicate the aforementioned findings while addressing three methodological ambiguities and limitations of the first study. In keeping with a suggestion by Ho and Driscoll (1998), a first refinement concerns the explicit measurement of perceptions regarding the proposed underlying process for the increase in self-directed negative affect observed in Study 1: A discrepancy perceived between the prescriptions of principles towards intergroup fairness and pro-ingroup biased intergroup ratings (henceforth, FN discrepancy) drives those affective effects. Based on a median split on the discrepancy measure, in Study 2 we test this line of reasoning by assessing the affective impact of ingroup identification under conditions of higher vs. lower perceived FN discrepancy. Besides, we include a direct assessment of the assumption that FN discrepancy is indeed perceived as a violation of the intergroup fairness norm. To this end, in Study 2 we measure perceived salience of the intergroup fairness norm after the participants have expressed ingroup-outgroup distinctions. Finally, in this study we assess ingroup identification prior to the measurement of intergroup bias. By contrast, in Study 1 we measured ingroup identification after the participants expressed intergroup ratings. Although intergroup ratings did not influence ingroup identification, it cannot be ruled out that in Study 1 group identification was a result rather than a cause of the comparison process.

Thus, the only difference between Study 1 and 2 was the inclusion of the above two measures. Accordingly, in Study 2 the predictions were similar to those formulated in Study 1. In line with our rationale, we expected that when perceiving own intergroup evaluation as being highly discrepant with the prescriptions of the intergroup fairness norm, higher ingroup identifiers would experience greater negative affect, compared to lower identifiers.

[19]
---------------
[20]

Concerning the process underlying this effect, additionally, as in Study 1 we based our predictions, on the one hand, on evidence showing that it less socially acceptable to differentiate the ingroup from a relevant outgroup on negative dimensions compared to positive ones (Hewstone et al., 1981; Mummendey & Otten, 1998). On the other hand, we also based our hypotheses on evidence showing that people downgrade the outgroup to the extent that they identify with their own group (e.g., Branscombe & Wann, 1992, 1994; Mummendey et al., 2001; Wann & Branscombe, 1995). Accordingly, we also hypothesized that outgroup derogation but not ingroup favoritism would mediate the effect of ingroup identification on felt negative affect. However, in line with our rationale, we expected to observe this effect only for those participants that perceived themselves as having evaluated the target groups in a way that was highly discrepant with the prescriptions of the intergroup fairness norm. Finally, we also posited that a discrepancy between prescriptions of intergroup fairness and own intergroup evaluations is perceived as a violation of the fairness norm. In line with this assumption, thus, we predicted that the participants would self-report their intergroup evaluations to be discrepant with the norm of intergroup fairness to the extent that they had perceived this latter norm as being salient while they were evaluating the target groups.

METHOD

Participants and Procedure

We carried out the study in Bozen in 2002. The participants were ninety-four higher-school students (56 women, 38 men). All belonged to the Italian-speaking population group.

Except for the fact that we measured ingroup identification prior to the expression of intergroup evaluation, Study 2 followed a procedure similar to that of the preceding study. First, we measured ingroup identification. We then asked the participants to give their perception of the same two targets as in Study 1 (i.e., members of the Italian- and the German-speaking population group) on a number of trait adjective items that appeared in random order on a list. We counterbalanced rating order for target group. Next, we then asked the participants to answer (a) one question about the extent to which they perceived their evaluation of the target groups as being discrepant with the prescriptions of principles towards intergroup non-discrimination, and (b) one question about the extent to which they had perceived the norm of intergroup non-discrimination as being salient while they were evaluating the target groups. Finally, we then asked the participants to rate their affect based on how they felt after completing the intergroup evaluation prior task.

Design and Dependent Measures

The design was a (Ingroup Identification: Lower vs. Higher) X 2 (Perceived FN Discrepancy: Lower vs. Higher) between-participants factorial. (See Table 3.)

[20]
---------------
[21]

Table 3. Descriptive Statistics

M

SD

Cronbach Alpha

Ingroup Identification

2.88

0.65

0.80

Self-directed negative affect

2.57

0.81

0.86

Ingroup positive evaluation

4.97

0.65

0.70

Outgroup positive evaluation

3.82

0.70

0.87

Ingroup negative evaluation

1.49

0.62

0.71

Outgroup positive evaluation

1.81

0.66

0.80

Fairness norm salience

2.61

0.80

(single-item scale)

Fairness norm discrepancy

2.04

0.74

(single-item scale)

We assessed all measures on 6-point scales (0 = not at all, 5 = extremely) with no neutral point. The respective mean, standard deviation, and scale reliability can be found in Table 3. By administering to the participants a standardized scale (Doosje et al., 1995), we calculated mean scores for ingroup identification. We measured evaluation of the target groups in the same way as in Study 1. Again, for each target group, we constructed separate positive and negative evaluation indices by averaging the mean scores of the positive and negative adjective traits: a six-item Positive evaluation index, and a six-item Negative evaluation index. As in Study 1, both for the ingroup and the outgroup, we found ratings expressed on the positively- and the negatively-valenced items, although highly correlated, to be not completely reciprocal (rs = - .48 and - .53, respectively). Therefore, the computation of ingroup favoritism and outgroup derogation scores appeared methodologically justified. We then computed scores for intergroup differentiation by subtracting positive ratings of the outgroup from those of the ingroup (ingroup favoritism), and by subtracting negative ratings of the ingroup from those of the outgroup (outgroup derogation).

In Study 2, however, we included two new measures. By means of two questions (Think about the evaluation of the two groups you have just provided: To what extent is it in line with the societal principles towards non-discrimination against people belonging to social groups other than the group one belongs to?, and In this questionnaire, as you were providing your evaluation of the two groups, to what extent did you feel the moral obligation not to discriminate the other group?), we calculated mean scores for perceived salience of the intergroup fairness norm and FN discrepancy, respectively. Uneven distributions of the participants resulted by preliminary divisions using modal and tertile split. As a consequence, we split on the median of the participant scores obtained for ingroup identification and FN discrepancy. Accordingly, we classified the participants scoring above the median (ingroup identification = 2.83; FN discrepancy = 2.00) as higher and the rest as lower in ingroup identification and FN discrepancy, respectively. Two subsequent separate one-way ANOVAs showed that that these groups of participants differed significantly on the respective measures (ingroup identification: F(1, 93) = 209.21, p < .001 (Lower: M = 2.00; Higher: M = 3.71); FN discrepancy: F(1, 93) = 81.83, p < .001 (Lower: M = 1.28; Higher: M = 2.87).

As in Study 1, we derived the measure of self-directed negative affect from ratings of self-directed negative feelings experienced after evaluating the target groups as assessed on 6-point scales (0 = does not apply at all, 5 = applies very much) with no neutral point. The items were the same as in Study 1 (angry at myself, guilty, disappointed with myself, disgusted with myself, ashamed).

[21]
---------------
[22]

RESULTS

Preliminarily, we tested for the effects of gender and target group rating order on the dependent variables used in the following analyses. No main or interaction effects emerged for these variables. As a consequence, we collapsed data across these factors in subsequent analyses, from which we consequently eliminated the former.

Subsequently, a 2 (FN discrepancy: Higher vs. Lower) by 2 (Ingroup Identification: Lower vs. Higher) ANOVA yielded a FN discrepancy by ingroup identification interaction effect on reported self-directed negative affect, F(1, 93) = 7.25, p < .008, and no other significant effects. Analyses of simple-main effects revealed that, as predicted, among those participants that perceived themselves as having evaluated the target groups in a way that was highly discrepant with the prescriptions of the intergroup fairness norm, higher identifiers experienced stronger self-directed negative affect than did lower identifiers, F(1, 45) = 3.78, p < .05. By contrast, for lower-FN discrepancy participants, ingroup identification exerted no significant simple-main effect, F < 1.

Table 4. Self-Directed Negative Affect as a Function of Ingroup Identification for Higher Norm Discrepancy Participants

Lower identification

Higher identification

Self-directed negative affect

2.46
(1.01)

3.90
(0.90)

Standard deviations are shown in parenthesis

We then performed analysis of covariance to test the prediction that outgroup derogation, but not ingroup favoritism, would mediate the aforementioned affective impact of ingroup identification on highly FN-discrepant participants. After including both outgroup derogation and ingroup favoritism as covariates, on the self-directed negative affect measure the outgroup derogation covariate was individually and in combination significant, combined F(1, 45) = 4.93, p < .05, and the ingroup favoritism covariate was individually and in combination nonsignificant, Fs < 1. Most importantly, as predicted, inclusion of outgroup derogation as a covariate both individually and in combination with ingroup favoritism reduced to nonsignificance the significant simple-main effect of ingroup identification on self-directed negative affect found for higher FN-discrepant participants in the original ANOVA, F(1, 45) = 3.02, p = .11.

Finally, as predicted, correlation analysis showed that the participants self-reported the intergroup evaluation they had provided to be discrepant with the norm of intergroup fairness to the extent that they had perceived the latter norm as being salient while they were evaluating the target groups, r(94) = -0.44, p < .001.

DISCUSSION

In Study 2 we replicated the findings of Study 1 while addressing some methodological ambiguities and limitations of that study. First, the results of this study provide support for the proposed underlying process for the increase in self-directed negative affect observed in the Study 1: Indeed, we predicted and found the same effect, namely, an increase in self-directed negative affect only among those participants that perceived their pro-ingroup biased intergroup ratings as being highly discrepant with the prescriptions of principles towards intergroup fairness.

[22]
---------------
[23]

Second, we also addressed the research problem of which processes are driving the negative affective consequences of ingroup identification found in the first study. To this issue, in line with evidence from recent research( e.g., Costarelli & Colloca, 2004), Study 2 provided empirical support for the notion proposed in that study that what drives the affective effects of ingroup identification is the least socially justifiable form of pro-ingroup bias (i.e., outgroup derogation). Most importantly, Study 2 showed that it is this very factor that makes higher identifiers experience the higher degree of FN discrepancy and, hence, the higher affective discomfort. Finally, in Study 1 we also assumed that people perceive a discrepancy between prescriptions of intergroup fairness and their pro-ingroup biased intergroup evaluation as a violation of the fairness norm. In line with this notion, in Study 2 we predicted and found that the participants self-reported their intergroup evaluation to be discrepant with the norm of intergroup non-discrimination to the extent that they had perceived this latter norm as being salient while they were evaluating the target groups. Importantly, this finding rules out the possibility that those group members who are more prone to express FN discrepant intergroup evaluations, namely, higher identifiers, do so because they perceive discrimination against the outgroup as the ingroup local norm.

GENERAL DISCUSSION

As predicted, among Italian South-Tyroleans, the social identity-based motive of establishing intergroup distinctiveness affected ingroup-outgroup comparisons. However, in line with evidence from previous relevant research, the establishment of intergroup distinctiveness via outgroup derogation mediated the subsequent elicitation of a discrepancy-based negative affect arousal as a function of ingroup identification.

We found these effects on the very same dependent variable that prior researchers typically used to investigate the research problem that we focused on in the present study. This shows that our findings can complement and extend the relevant literature. Specifically, at the conceptual level, they highlight two interrelated motivational processes that are simultaneously at work. The first process is consistent with social identity theory (Tajfel & Turner, 1979). This theory proposes that individuals derive part of their self-concept, namely, their social identity, through their belonging to social groups. Hence, in line with a motivation to evaluate oneself positively, individuals try to achieve or maintain a positive social identity by establishing ingroup positive distinctiveness, relative to relevant outgroups, through evaluative or behavioral intergroup differentiation (cf. Brewer, 1979),. Our data indicate that the participants engaged in this basic motivational process while completing the intergroup evaluation task, albeit to different degrees.

However, being relatively temperate in the public expression of own pro-ingroup-biased intergroup evaluation and behavior has become the content of important normative prescriptions towards intergroup fairness. Consistent with this notion, van den Bos, Wilke, and Lind (1998) have found that fairness is a particularly salient issue precisely when people are concerned about potential problems that are associated with socially based identity processes. Furthermore, self-categorization theory (Turner, et al., 1987) suggests that individuals are motivated to be sensitive to the content of social norms because acting in accordance with social norms expresses identity at the group-level of inclusiveness, or social identity.

Yet our data also highlight the operation of a second process. Specifically, to the extent that group-based motives give rise to the aforementioned expression of intergroup differentiation,

[23]
---------------
[24]

nonconformity to social norms of nondifferentiation (fairness) may elicit negative affective consequences for the self. In so doing, first, our findings qualify the notion that people try to accommodate intergroup differentiation and fairness together (Singh et al., 1998; Tajfel et al., 1971; Turner, 1983). Additionally, our novel finding that the degree of post-intergroup evaluation negative affect was a function of ingroup identification lends some empirical, however preliminary, support to the novel social identity-based perspective that Ho and Driscoll (1998) have suggested for studies on post-intergroup evaluation affect (cf. Costarelli & Palmonari, 2003; Mackie & Smith, 1999).

In line with evidence from recent research (e.g., Costarelli & Colloca, 2004), our second main finding that outgroup derogation but not ingroup favoritism seems to mediate the self-directed negative affect arousal effect of ingroup identification provides convergent empirical support to the positive-negative asymmetry effect in intergroup evaluation (Mummendey & Otten, 1998) under a conceptually novel perspective, namely, the affective one. We have done so by highlighting the affective costs that contravening to the psychological bases of this asymmetry effect has for group members, and all the more so for those among them that identify the most with the ingroup. Were intergroup evaluations in the positive and negative domain symmetric rather than asymmetric, our experience of self-directed negative affect should have been mediated by their expression of both outgroup derogation and ingroup favoritism. The fact that this was not the case provides convergent evidence in support of the positive-negative asymmetry effect in intergroup evaluation (Mummendey & Otten, 1998).

A notable limitation of our research was that we split our participants on the median of the ingroup identification score to identify relatively low and high ingroup identifiers in spite of the fact that both in Study 1 and 2 the ingroup identification score distribution was negatively skewed (i.e. lower and higher identifiers were in fact rather highly identified with the ingroup, a finding indeed quite plausible given the motivationally strong drive to identify with one’s own ethnic group as a default). Given these non-normal score distributions, modal splits and subsequent use of non-parametric statistics would have been more appropriate (although our finding statistically significant differences by using the more conservative, parametric tests actually argues for, rather than against, the tenability of our findings). We therefore acknowledge that some inferences may have resulted from the methodology used rather than the underlying psychological processes. The current findings should thus be taken with caution and further research should ascertain their empirical tenability.

To conclude, by combining the logic of two distinct literatures, the advances of this exploratory study may contribute to reconcile intraindividual (e.g., Devine et al., 1991; Monteith, 1993, 1996; Monteith et al., 1993) and socially-based perspectives (e.g., Ho and Driscoll, 1998; Mackie & Smith, 1999) accounting for the affective processes that are set in motion by intergroup evaluation.

REFERENCES

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

Branscombe, N. R., & Wann D. L. (1992). Physiological arousal and reactions to outgroup members during competitions that implicate an important social identity. Aggressive Behavior, 18, 85-93.

[24]
---------------
[25]

Branscombe, N. R., & Wann, D. L. (1994). Collective self-esteem consequences of outgroup derogation when a valued social identity is on trial. European Journal of Social Psychology, 24, 641-657.

Breckler, S., & Greenwald, A. (1986). Motivational facets of the self. In Sorrentino R, Higgins T (eds.), Handbook of motivation and cognition. New York: Guilford, 145-164.

Brewer, M.B. (1979). Ingroup favoritism in the minimal intergroup situation: A cognitive motivational analysis. Psychological Bulletin, 86, 307-324.

Brown, R. (1995). Prejudice: Its social psychology. Cambridge, MA: Blackwell.

Brown, R. (2000). Social Identity Theory: Past achievements, current problems, and future challenges. European Journal of Social Psychology, 30, 745-778.

Capozza D., & Manganelli Rattazzi, A. M. (1999). Stereotypes and attribution processes in a multi-ethnic Italian province. Journal of Community and Applied Social Psychology, 9, 231-236.

Costarelli, S., & Colloca, P. (2004). Intergroup conflict, outgroup derogation, and self-directed negative affect among Italian South-Tyroleans. Journal of Social Psychology, 144, 181-189.

Costarelli, S., & Palmonari, A. (2003). Ingroup Ambivalence and Experienced Affect: The Moderating Role of Social Identification. European Journal of Social Psychology, 33, 813-821.

Devine, P. G, Monteith. M. J, Zuwerink, J. R., & Elliott, A. J. (1991). Prejudice with and without compunction. Journal of Personality and Social Psychology, 60, 817-830.

Doosje, B., Ellemers, N., & Spears, R. (1995). Perceived intragroup variability as a function of group status and identification. Journal of Experimental Social Psychology, 31, 410-436.

Hewstone, M., Fincham, F., & Jaspars, J. (1981). Social categorization and similarity in intergroup behaviour: A replication with penalties. European Journal of Social Psychology, 11, 101-107.

Hinkle, S., & Brown, R. (1990). Intergroup comparisons and social identity: Some links and lacunae. In D. Abrahams and M. Hogg (Eds.), Social Identity Theory: Constructive and critical advances. New York: Springer Verlag.

Ho, C. P., & Driscoll, D. (1998). Prejudiced responses and affect. British Journal of Social Psychology, 37, 275-285.

Hogg, M. A, & Abrams, D. (1988). Social identifications: A social psychology of intergroup relations and group processes. London: Routledge.

Jaccard, J. (1998). Interaction effects in factorial analysis of variance. London: Sage, 76-77.

[25]
---------------
[26]

Judd, C. M., & Kenny, D. A. (1981). Process analysis: Estimating mediation in treatment evaluations. Evaluation Review, 5, 602-619.

Kaplan, K. J. (1972). On the ambivalence-indifference problem in attitude theory and measurement: A suggested modification of the semantic differential technique. Psychological Bulletin, 77, 361-372.

Kirchler, E, & Zani, B. (1989). Strategie di differenziazione tra gruppi etnici: Italiani e Tedeschi in Alto Adige. Giornale Italiano di Psicologia, XVI, 1, 61-79.

MacDonald, T. K., & Zanna, M. P. (1998). Cross-dimension ambivalence toward social groups: Can ambivalence affect intentions to hire feminists? Personality and Social Psychology Bulletin, 24, 427-441.

Mackie, D. M., & Smith, E. R.. (1998). Intergroup relations: Insights from a theoretically integrative approach. Psychological Review, 105, 499-529.

Monteith, M. J. (1993). Self-regulation of prejudiced responses. Implications for progress in prejudice reduction efforts. Journal of Personality and Social Psychology, 65, 469-485.

Monteith, M. J. (1996). Contemporary forms of prejudice-related conflict: In search of a nutshell. Personality and Social Psychology Bulletin, 22, 461-473.

Monteith, M. J., Devine, P. G., & Zuwerink, J.R. (1993). Self-directed versus other-directed affect as a consequence of prejudice-related discrepancies. Journal of Personality and Social Psychology, 68, 198-210.

Mummendey, A., & Otten, S. (1998). Positive-negative asymmetry in social discrimination. European Review of Social Psychology, 9, 107-143.

Mummendey, A., & Klink, A., & Brown, R. (2001). Nationalism and patriotism: National identification and outgroup rejection.. British Journal of Social Psychology, 40, 159-172.

Tajfel, H. (1981). Human groups and social categories: Studies in Social Psychology. Cambridge: University Press.

Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1, 149-178.

Tajfel, H., & Turner, J. C. (1986). The social identity theory of intergroup behavior. In S. Austin and W. G. Austin (Eds.), Psychology of intergroup relations. Chicago: Nelson-Hall. 7-24.

Van den Bos, K., Wilke, H. A., & Lind, E. A. (1998). When do we need procedural fairness? The role of trust in authority. Journal of Personality and Social Psychology, 75, 1449-1458.

Wann, D. L., & Branscombe, N. R. (1995). Influence of level of identification with a group and physiological arousal on perceived intergroup complexity. British Journal of Social Psychology, 34 , 223-235.

[26]
---------------
[27]

APPENDIX

Ingroup Identification Scale (Doosje et al., 1995)

1. I see myself as a [member of Group X]

2. I am pleased to be a [member of Group X]

3. I feel strong ties with [members of Group X]

4. I identify with other [members of Group X].

AUTHOR BIOGRAPHIES

Sandro Costarelli is an Assistant Professor at the School of Applied Cognitive Psychology, University of Trento, Italy. His research focuses on social identity, intergroup ambivalence, and prejudice-related affect. E-mail: sandro.costarelli@unitn.it.

Rose Marie Callà received her BA in Sociology at the School of Sociology, University of Trento, Italy. Her research focuses on social identity and intergroup relations. E-mail: rosemarie.callà@unitn.it.

[27]
---------------
[28]