Back to Journal Cover Page


CURRENT RESEARCH IN SOCIAL PSYCHOLOGY


Volume 4, Number 9
Submitted: October 18, 1999
Resubmitted: November 25, 1999
Accepted: November 26, 1999
Publication date: November 27, 1999

ATTRIBUTIONAL BIAS AND IDENTITY IN A CONFLICT REGION: THE MODERATING EFFECTS OF STATUS

John A. Hunter
University of Otago

Maurice Stringer
University of Ulster

[160]
---------------
[161]

ABSTRACT

The relationship between evaluation of group identity and group serving attributional biases was examined among Protestant and Catholic school children in Northern Ireland. Participants were presented with a series of vignettes. These depicted ingroup and outgroup targets who engaged in positive and negative behavior. After attributing cause to the target's behavior, the extent to which participants identified with their respective groups was then assessed. Members of the Catholic group displayed enhanced levels of group serving attributional biases. They were also shown to have less positive group identification scores. These patterns are consistent with the lower status position of this category. Contrary to expectations, the only significant correlations between attributional differentiation and strength of group identity were found among members of the Protestant category. Members of this group emitted more ambivalent attributional patterns. These findings are interpreted as suggesting that, in this particular context, high status group members evaluate their group identity positively by displaying more subtle forms of bias. One limitation of the present study is its reliance on the on the distinction between internal and external attributional dimensions. It is suggested that future research should move beyond this dichotomy and incorporate a multidimensional approach to the study of intergroup causal dimensions.

INTRODUCTION

The popularity and wide ranging nature of attribution research in social psychology has often been acknowledged (Farr 1991; Hunter, Stringer, Millar & Watson 1994). Despite the breadth of interest shown in this area, relatively little attention has been paid to attributional explanations at the intergroup level (e.g., Kelley & Michela 1980; Harvey & Weary 1984). The limited literature that exists, however, indicates that respondents drawn from distinct social categories attribute cause to ingroup and outgroup members differently (e.g., Hunter, Stringer & Watson 1991; Islam & Hewstone 1993; Taylor & Jaggi 1974).

[161]
---------------
[162]

Much of the research carried out on intergroup causal attribution has recently been reviewed by Hewstone (1990). Here it is shown that ingroup favoring/outgroup derogating attributions are not the universal phenomena which had previously been predicted (cf. Pettigrew 1979). Although, group serving attributional biases tend to occur on some dimensions and not others in a particular study, as Hewstone demonstrates, its effects are clearly evident in three spheres: positive and negative outcomes, successes and failures, and group differences. One of the most important theoretical contributions offered in Hewstone's review is the proposal that intergroup attribution and social identity are related. It is argued that group based attributions can be linked to social identity in two main ways. In the first instance, it is postulated that by attributing positive ingroup or negative outgroup behavior to internal, stable or controllable causes a positive social identity can be achieved or enhanced. In the second instance, it is postulated that by attributing negative ingroup or positive outgroup behavior to external, unstable or uncontrollable causes a positive social identity may be maintained or protected.

According to social identity theory groups provide their members with a social identity which they are motivated to evaluate positively. This is done by making comparisons between relevant ingroups and outgroups. Positively discrepant comparisons, or those which achieve positively valued distinctiveness, lead to the attainment of a positive social identity (Tajfel & Turner 1979). As Hinkle & Brown (1990 p.62) point out, this postulate can be translated into a testable hypothesis. That is, that there should be a positive correlation between positive intergroup differentiation and strength of identity.

From this perspective it is possible to develop a framework for the examination of the relationship between attributional bias and strength of identity. In terms of the three spheres of group serving attributional biases outlined by Hewstone (1990), however, only that pertaining to positive and negative outcomes will be used in the present study. Research carried out in this area has revealed two types of attributional activity. These have been defined by Hewstone as categorization and outcome effects. The more pervasive type of finding is known as the categorization effect. This type of attributional pattern refers to the tendency for group members to make; (a) more internal attributions for positive ingroup than positive outgroup behavior; and (b) less internal attributions for negative ingroup than negative outgroup behavior. The less pervasive type of finding is known as the outcome effect. This type of attributional activity refers to the tendency for group members to make; (a) more internal attributions for positive than negative ingroup behavior; and (b) less internal attributions for positive than negative outgroup behavior.

[162]
---------------
[163]

Although it can be hypothesized that each of these four types of attributional activity may be related to strength of identity, it is important to note that, the results of other research carried out under the general rubric of intergroup relations have so far failed to find a consistent pattern of relationships. Indeed, studies implemented to assess the link between ingroup bias and strength of identity reveal positive correlations, negative correlations and correlations which are effectively zero (see Hinkle & Brown 1990 for a review). One attempt to bring some understanding to the confusion inherent in this general area has been proposed by Hinkle & Brown (1990). These authors have recently begun to suggest that not all social categories will engage in positively discrepant intergroup comparisons in order to achieve a positive social identity (see also, Tajfel & Turner 1979). Drawing upon the orthogonal constructs of collectivism-individualism and comparative and non-comparative ideologies Hinkle and Brown, in a series of studies, have recently found support for this view. Recent research indicates that the strongest relationships between intergroup bias and evaluation of identity occur among groups which are both collective and comparative in nature (Brown, Hinkle, Ely, Fox-Cardamone, Maras, Taylor 1992; Hinkle, Brown & Ely in press).

Predictions derived from research and theory in both the present context and elsewhere indicate that such conditions are most likely to be met among category members in conflicted situations. In such circumstances it is generally accepted that group membership is both meaningful and will contain evaluative connotations (Crocker & Luhtanen 1990; Hunter, Platow, Howard & Stringer 1996; Turner 1999). On the basis of such reasoning we may therefore predict that strong positive associations between attributional biases and strength of identity will exist among Catholic and Protestant category members in the context of the Northern Ireland’s continued enmity.

SAMPLE

One hundred and fourteen Protestant and Catholic adolescent males aged between fourteen and fifteen took part in the study. The participants were chosen from religiously segregated secondary schools in Northern Ireland (Murray 1985). Fifty-five of the respondents were Protestant and fifty-nine were Catholic.

[163]
---------------
[164]

STIMULUS MATERIALS AND PROCEDURE

Participants were presented with a series of episodes describing the behavior of an actor in the town where they lived. In each instance respondents were asked to imagine that the actor was behaving towards them. The descriptions were one paragraph long and comprised an event involving either the positive or negative behavior of an ingroup or outgroup member. There were sixteen paragraphs in all. This allowed the perceiver to attribute cause to four examples of each positive and negative ingroup and outgroup behavior. The episodes were presented in a random fashion.

The content of each episode reflect that described and reported elsewhere on the literature on intergroup causal attributions (i.e., Hunter, Stringer & Watson 1993; Islam & Hewstone 1993; Taylor & Jaggi 1974). In an attempt to circumvent the sensitive issue of religion (or more aptly the terms 'Protestant' and 'Catholic') in Northern Irish schools, stereotyped cues were utilized to denote denominational membership. The cues used here were those that referred to actors ‘names’ and ‘schools.’ Prior research has shown that these particular cues are reliable predictors of category membership (Millar & Stringer 1991).

Respondents were asked to imagine that, in the town where they lived, actors with a Protestant name (i.e., David Cavendish) and affiliated with a Protestant school (i.e., Ballymena Grammar school) or a Catholic name (i.e., Patrick O’Hagan) and affiliated with a Catholic school (i.e., Saint Joseph’s Secondary school) behaved towards them in a positive or negative fashion. The situations involved the actor either; (a) giving the respondent a lift in his car on a rainy day or driving through a puddle and splashing him with muddy water; (b) helping or walking on after the respondent falls off a bicycle; (c) giving an extra helping of fish and chips or short changing the respondent; (d) returning or running off with a parent’s lost wage packet. Following each paragraph respondents were presented with four possible explanations as to why the actor had behaved as he had. Two of these represented internal attributions (e.g. ‘that’s just the way he is’) and two represented external attributions (e.g., ‘every one who comes into that shop gets a large portion’). Participants were instructed to choose the one explanation which they thought was the most likely cause for the behavior in question.

[164]
---------------
[165]

To measure social identity, the group identification scale developed by Brown, Condor, Mathews, Wade & Williams (1986) was used. This scale, specifically devised to measure the three core facets of awareness, evaluation and affect, implicit in Tajfel's classic definition of social identity (Tajfel 1978, p.63) was comprised of ten items. Five items affirmed identity and five denied identity. The measure has been repeatedly found to have psychometrically acceptable levels of reliability and validity in both the Northern Irish context and elsewhere (Brown, et al. 1986; Hunter & Stringer 1992). In line with such findings, Cronbach's alpha in the present study revealed an alpha of .75 (n=114).

All questionnaires were completed in a supervised classroom in the participants’ own school. The attribution items were answered first. The identification scale was answered second [although these measures are usually presented in the reverse order, Tajfel & Turner's (1979) predictions, on which this research is based also facilitates the present methodology]. No consultation or discussion among classmates was permitted during the study

RESULTS

After other research in this area (e.g., Hewstone & Ward 1985; Hunter, Stringer & Coleman 1993) the proportion of internal attributions for each of the four main types of episode were summed to produce a single score. Untransformed proportional scores can be seen in Table 1. To overcome heterogeneity of variance, a problem arising from the use of proportions, the arcsine transformation was applied to each of the cell means (Hewstone & Ward 1985; Howell 1987).

A 2 x 2 x 2 mixed model analysis of variance (ANOVA), with one between subjects factors - religion (Protestant x Catholic) and two within subjects factors - target group of evaluation (ingroup x outgroup) x outcome behaviors (positive x negative) was used to analyze the data. A main effect was found for outcome behaviors (F(110,1)= 29.96, p<.001). More internal attributions were made for positive (M=.60) than negative (M=.48) behaviors. This effect was however mediated by the three way interaction found between religion x target actor’s group x outcome behaviors (F(110,1)=7.93, p<.007). To assess the implications of this interaction effect further, a priori comparisons using t tests were carried out. These tests compared mean attributions made for; (a) positive ingroup and outgroup behavior; (b) negative ingroup and outgroup behavior; (c) positive and negative ingroup behavior and; (d) positive and negative outgroup behavior. Analyses were carried out separately for the members of each category.

[165]
---------------
[166]

Table 1. Mean Score of Attributions Made for Actor’s Behavior

 

Positive
Protestant

Positive
Catholic

Negative
Protestant

Negative
Catholic

Protestant
Perceiver
0.64 0.62 0.50 0.50
Catholic
Perceiver
0.52 0.60 0.48 0.42

For members of the Protestant category, the only significant findings were discerned on the attributional dimensions which differentiated between positive and negative ingroup and outgroup behavior. For both ingroup behavior (t(54)= 4.42, p<.001) and outgroup behavior (t(54)=3.44, p<.001) more internal attributions were made for positive than negative actors behavior. Catholic group members were shown to differ significantly in the way they attributed positive ingroup and outgroup behavior (t(58)=2.64, p<.01). More internal attributions were made for positive ingroup behavior. Significant differences were also found in the way Catholics attributed negative ingroup and outgroup behavior (t(55)=2.60, p<.01). More internal attributions were made for negative outgroup behavior. A further effect showed significant differences in the way cause was attributed to positive and negative ingroup behavior (t(58)= 4.08, p<.001). More internal attributions were made for positive than negative behaviors. To ensure that the family-wise error rate did not increase to an unacceptable level a check using Dunn’s test (Bonferroni t) was also carried out. Using this highly conservative procedure (Howell 1987), tests carried out across all analyses revealed that each t significant at the p<.001 level was also significant at the p<.01 level (Dunn’s critical alpha value, 3.18, p<.01). All t’s significant at the p<.01 level were significant at the p<.05 level (Dunn’s critical alpha value, 2.59, p<.05).

A further analysis of variance (ANOVA) was carried out to examine the effects of social category membership (Protestant x Catholic) on respondents identity scores. A main effect was found (F(1, 106)=19.07, p< .001). This revealed that respondents from the Protestant category (M=38.16) had significantly higher group identification scores than their Catholic (M=33.20) counterparts.

[166]
---------------
[167]

ATTRIBUTIONAL DIFFERENTIATION

For each respondent four indices of attributional differentiation were compiled. Following the methodology outlined by Brown, et al. (1986), the first measure was achieved by subtracting the proportion of internal attributions made for positive outgroup behavior from those made for positive ingroup behavior. The second was obtained by subtracting the proportion of internal attributions made for negative ingroup behavior from those made for negative outgroup behavior. A third was obtained by subtracting the proportion of internal attributions made for negative ingroup behavior from those made for positive ingroup behavior. The fourth was obtained by subtracting the proportion of internal attributions made for negative outgroup behavior from those made for positive outgroup behavior. Each of the four indices of indices were correlated with the identity scale.

Table 2. Correlations between Identity and Intergroup Attributional Differentiation

  Positive
Ingroup and Outgroup
Behavior
Negative
Ingroup and Outgroup
Behavior
Positive/Negative
Ingroup
Behavior
Positive/Negative
Outgroup
Behavior
Protestant
(N=55)
0.03 0.18 0.27* 0.37**
Catholic
(N=59)
0.06 -0.02 0.06 0.05

* P<. 05 **p<.01

As can be seen from Table 2, the only correlations between attributional differentiation and evaluation of group identity to reach significance were manifested by members of the Protestant category. Only moderately positive (r's of .27 and .37) these correlations were displayed on those dimensions which differentiated between positive and negative, ingroup and outgroup behavior. These two dimensions were the only ones on which members of this category showed significant levels of attributional differentiation.

[167]
---------------
[168]

DISCUSSION

Strong consistent correlations between group serving attributional biases and strength of group identity failed to emerge. Members of the Catholic category were shown to have less positive identity scores. Respondents belonging to this category were also found to manifest more pronounced levels of group serving attributional biases. Contrary to expectations, the only significant correlations between attributional differentiation and evaluation of group identity were discerned among members of the Protestant category. While only moderately positive (r's of .27 and .37), these associations were found on those dimensions which differentiated between positive and negative ingroup and outgroup behavior. These two dimensions were the only ones on which the members of this category showed significant levels of attributional differentiation.

Overall, both categorization and outcome group serving attributional effects were found. With regard to the former only members of the Catholic group displayed these patterns of bias. Members of this category made more internal attributions for positive ingroup than positive outgroup behavior and less internal attributions for negative ingroup than negative outgroup behavior. Significant outcome effects were evident in that both Protestants and Catholics made more internal attributions for positive than negative ingroup behaviors. When cause was attributed to outgroup behavior only one significant effect emerged. Here members of the Protestant group made more internal attributions for positive than negative outgroup behaviors. This particular effect is normally taken to indicate a trend towards outgroup favoritism (see Hewstone 1990).

The tendency for members of the Catholic group to manifest more pronounced levels of attributional bias runs counter to Hinkle and Brown's (1990) predictions. Such patterns are however consistent with what might be expected among a low status social group which is beginning to challenge a society's accepted pattern of differentiation (Abrams & Hogg 1988; Taylor & McKirnan 1984). Thus, in Northern Ireland, where Catholics are more likely than Protestant to be unemployed, dependent on social welfare and if working to have lower incomes (Bell 1976; Osborne & Cormack 1989) it is (a) largely accepted that Catholics are of lower status than are Protestants and (b) that they perceive this situation to be both illegitimate and unstable [see Whyte (1990) for an excellent discussion].

[168]
---------------
[169]

An alternative (yet not unrelated) explanation for our findings may, be offered in terms of the relative size of the Catholic group. Within the demographic confines of Northern Ireland, members of the Catholic category constitute a numerical minority. In so far as several studies have revealed that the members of numerical minority groups’ often engage in relatively high levels of intergroup discrimination (Brown & Smith 1989; Sachdev & Bourhis 1984), it could thus, be argued that the enhanced levels of group serving attributional biases shown by the Catholics in the current study are more a function of group size than status. Such an argument is, however, difficult to sustain. Essentially, this is because, although Catholics form a numerical minority when considered within the confines of Northern Ireland, they form a large numerical majority when one considers Ireland as a whole. Moreover, given that prior research has revealed that Catholics (who reside within the boundaries of Northern Ireland) identify with and perceive themselves to be living in Ireland, while Protestants (who live in the same geographical region) identify and perceive themselves to living in Northern Ireland, it is entirely possible that the members of both categories consider themselves to be in the numerical majority (Cairns 1987). Future (laboratory based) research can of course, by controlling for status and relative group size, help clarify such issues.

Nevertheless, the lack of any meaningful relationship between attributional bias and evaluation of group identity indicates that, in this real world context, the strategy of enhanced discrimination exhibited by the members of the lower status Catholic category is an ineffective way of achieving a positive group identity. The significant correlations found, between attributional differentiation and evaluation of group identity, among members of the Protestant category reveal a very different picture. The more selective patterns of differentiation displayed by the members of this group appear to be a much more effective way of achieving a positive evaluation of group identity. In this respect, what is surprising is that on one of the two dimensions on which Protestants emitted attributional differentiation more internal causation was ascribed to positive than negative outgroup behavior. This type of responding, which is generally taken to indicate a degree of outgroup favoritism, has been replicated among the members of other types of higher status groups (e.g., Sachdev & Bourhis 1991). These patterns are generally thought to emerge when category members believe their groups status positions to be secure (Brown 1988) and as such they might also be expected to change when category members feel threatened (see for example Hunter et al. 1991, 1994).

[169]
---------------
[170]

The results of the present study indicate that this strategy also facilitates high status category members in their attempts to achieve a positive group identity. One possible way in which this process may be understood is by drawing upon a growing body of evidence which indicates that high status group members, rather than display blatant biases, will sometimes resort to more subtle forms of ingroup favoritism (e.g., Crocker & Luhtanen 1990; Mullen, Brown & Smith 1992). This bias as illustrated by Mummendey and her colleagues can sometimes give the appearance of outgroup favoritism (e.g., Mummendey & Schreiber 1984; Mummendey & Simon 1989). In this sense therefore, the Protestant tactic of making more internal attributions for positive than negative outgroup behavior can be interpreted as a means by which a few ‘special’ Catholics who behave positively are differentiated from the majority who are judged to have behaved negatively. Such a strategy would, as Pettigrew (1979) points out, serve to exclude these ‘special case’ individuals from being proper representatives of the group. Consequently, the few ‘good’ Catholics can then be contrasted with the other members of the group in a way that highlights the group’s shortcomings (e.g., ‘he’s really different, not like other Catholics’) and thus works to enhance Protestant identity.

While admittedly post hoc, such an interpretation does allow for the possibility that there is a discernible link between group serving attributional biases and evaluation of group identity. We would acknowledge, however, that a weakness of the current investigation has been our focus on the internal/external attributional dichotomy. Indeed, we would accept the criticism that, in using these dimensions exclusively, divergent data patterns may (in principal) have been predicted. Research, currently in progress (e.g., Hunter, Reid & Stokell 1999) will attempt to overcome these difficulties by utilizing a multidimensional approach to investigate the attributional dimensions of locus, stability, controllability and globality (see Hewstone 1989 for a review). This work by further adopting the stratagem advanced by Russell and his colleagues (McAuley, Duncan & Russell 1992; Russell 1982), whereby respondents initially write down and then rate their own preferred causal statements, functions to overcome a further shortcoming apparent in previous research carried out in this area (see Hewstone 1990, for a review). Namely, the assumption that researchers can, of their own accord, accurately translate the subjective meaning intended in a person’s causal statements into scientifically objective causal dimensions.

[170]
---------------
[171]

REFERENCES

Abrams, D. Hogg, M. A. (1988). "Comments on the motivational status of self-esteem in social identity and intergroup discrimination." European Journal of Social Psychology, 18:317-334.

Bardon, J. (1992). A History of Ulster. Belfast, Blackstaff Press.

Bell, G. (1976). The Protestants of Ulster. London, Pluto Press.

Brown, R. (1988). Group Processes: Dynamics Within and Between Groups. Oxford, Blackwell.

Brown, R. J., Condor, S., Mathews, A., Wade, G., & Williams, J. A. (1986). "Explaining intergroup differentiation in an industrial organisation." Journal of Occupational Psychology, 59:273-286.

Brown, R., Hinkle, S., Ely, P. G., Fox-Cardamone, L., Maras, P., Taylor, L. A. (1992). "Recognising group diversity: Individualist-collectivist and autonomous-relational social orientations and their implications for intergroup processes." British Journal of Social Psychology, 31:327-342.

Cairns, E. (1987). Caught in Crossfire: Children and the Northern Ireland Conflict. Belfast, Appletree Press.

Crocker, J., Luhtanen, R. (1990). "Collective self-esteem and ingroup bias." Journal of Personality and Social Psychology, 1:60-67.

Farr, R. M. (1991). "Essay review: The continuing saga of attribution theory." British Journal of Social Psychology, 30:89-93.

Harvey, J. H., Weary, G. (1984). "Current issues in attribution research and theory." Annual Review of Psychology, 35:427-459.

[171]
---------------
[172]

Hewstone, M. (1989). Causal Attribution: From Cognitive Processes to Collective Beliefs. Oxford, Blackwell.

Hewstone, M. (1990). "The 'ultimate attribution error'? A review of the literature on intergroup causal attribution." European Journal of Social Psychology, 20:311-335.

Hewstone, M., Ward, C. (1985). "Ethnocentrism and causal attribution in South East Asia." Journal of Personality and Social Psychology, 48:614-623.

Hinkle, S., Brown, R. (1990). "Intergroup comparisons and social identity: Some links and lacunae." In D. Abrams & M. A. Hogg (Eds), Social identity theory: Constructive and critical advances. New York, Harvester/Wheatsheaf.

Hinkle, S., Brown, R., Ely, P. G. (in press). Social identity theory processes: Some limitations and limiting conditions. Revisita de Psicologia Social.

Howell, D. C. (1987). Statistical methods for Psychology. Boston, Duxbury.

Hunter, J. A., Platow, M. J., Howard, M. L. & Stringer, M. (1996). "Social identity and intergroup evaluative bias: Realistic categories and domain specific self-esteem in a conflict setting." European Journal of Social Psychology, 26:631-647.

Hunter, J. A., Reid, J. M. & Stokell, N. M. (1999). "Social attribution, self-esteem and social identity." (manuscript under submission).

Hunter, J. A., Stringer, M. (1992). "The reliability of the group identification scale in Northern Ireland." Perceptual and Motor Skills, 74:114.

Hunter, J. A., Stringer, M., Coleman, J. T. (1993). "Social Explanations and Self-Esteem in Northern Ireland." Journal of Social Psychology, 133:643-671.

[172]
---------------
[173]

Hunter, J. A., Stringer, M., Watson, R. P. (1991). "Intergroup violence and intergroup attributions." British Journal of Social Psychology, 30:261-266.

Hunter, J. A., Stringer, M., Millar, D., Watson, R. P. (1994). "Perceptions of violent events." In A. Guelke (Ed), New Perspectives on the Northern Ireland Conflict. Aldershot, Avebury/Gower.

Islam, M., Hewstone, M. (1993). "Intergroup attributions and affective consequences in majority and minority groups." Journal of Personality and Social Psychology, 64:936-950.,

Kelley, H. H., Michela, J. L. (1980). "Attribution theory and research." Annual Review of Psychology, 31:457-503.

Messick, D. M., Mackie, D. (1989). "Intergroup relations." Annual Review of Social Psychology, 40:45-81.

Millar, D., Stringer, M. (1991). "Group affiliation and confidence in denominational categorization in Northern Ireland." Journal of Community and Applied Social Psychology, 1:223-226.

McAuley, E., Duncan, T. E., Russell, D. W. (1992). "Measuring causal attributions: The revised causal dimension scale (CDSII)." Personality and Social Psychological Bulletin, 18:566-573.

Mullen, B., Brown, R., Smith, C. (1992). "Ingroup bias as a function of salience, relevance and status: An integration." European Journal of Social Psychology, 22:103-122.

Mummendey, A., Schreiber, H. J. (1984). "'Different' just means 'better': Some obvious and hidden pathways to ingroup favoritism." British Journal of Social Psychology, 23:363-367.

Mummendey, A., Simon, B. (1989). "Better or just different III: The impact of comparison dimension and relative group size upon intergroup discrimination." British Journal of Social Psychology, 28:1-16.

[173]
---------------
[174]

Osborne, R. D., Cormack, R. J. (1986). "Religion and unemployment in Northern Ireland." Economic and Social Review, 17:215-225.

Pettigrew, T. F. (1979). "The ultimate attribution error: Extending Allport's cognitive analysis of prejudice." Personality and Social Psychology Bulletin, 5:461-476.

Russell, D. W. (1982). "The causal dimension scale: A measure of how individuals perceive causes." Journal of Personality and Social Psychology, 42:1,137-1,145.

Sachdev, I., Bourhis, R. Y. (1991). "Power and status differentials in majority and minority group relations." European Journal of Social Psychology, 21:1-24.

Tajfel, H. (1978). Differentiation between social groups. Cambridge, Cambridge University Press.

Tajfel, H., Turner, J. C. (1979). "An integrative theory of intergroup conflict." In S. Worchel, W. G. Austin (Eds), The social psychology of intergroup relations. Monterey, Brooks/Cole.

Taylor, D. M., Jaggi, V. (1974). "Ethnocentrism and causal attribution in a South India context." Journal of cross-cultural psychology, 5:162-171.

Taylor, D. M., McKirnan, D. J. (1984). "A five-stage model of intergroup relations." British Journal of Social Psychology, 23:291-300.

Thompson, J. B. (1984). Studies in the Theory of Ideology. Cambridge, Polity.

Turner, J. C. (1999). "Some current issues in research on social identity and self-categorization theories." In N. Ellemers, R. Spears & B. Doojse (Eds.), Social Identity. Blackwell: Oxford.

[174]
---------------
[175]

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., Wetherell, M. A. (1987). Rediscovering the Social Group: A Self-Categorization Theory. Oxford, Blackwell.

Whyte, J. (1990). Interpreting Northern Ireland. Oxford, Clarendon Press.

AUTHOR BIOGRAPHIES

John A. Hunter is a lecturer in social psychology at the University of Otago, Dunedin, in New Zealand. E-mail: jhunter@psy.otago.ac.nz.

Maurice Stringer is senior lecturer and coordinator of research in psychology at the University of Ulster, Coleraine, in Northern Ireland. E-mail: m.stringer@ulst.ac.uk.

[175]
---------------

Back to Journal Cover Page