• Aucun résultat trouvé

Manipulating social identity salience

5.2.1. The salience of meaningful social identities

In the three studies preference was given to manipulating the salience of meaningful (real and artificial) social identities as opposed to minimal group identities. I assumed that if the social identities that were made salient were socially meaningful, it would enhance the general implication of participants in the experiments, in turn facilitating the induction of stronger emotions and of the corresponding appraisals. Ironically, a major drawback of using meaningful social identities is that they are by definition subjective. And, the subjective meaning that participants attribute to social identities in experimental settings may not correspond to the meaning attributed by the researchers (for a critical discussion of this issue see Antaki, Condor and Levine, 1996). More important, Brown (2000) has recently argued that social identity theory has not sufficiently accounted for the fact that different social groups may serve different identity functions, and that the social identification with different groups may have quite different meanings for different people in different contexts (also see Brown & Williams, 1984; Hogg & Williams, 2000). To cite Brown (2000), “It seems to me important to incorporate [these] central dimensions of group diversity and no longer assume that a group is a group is a group as far as key psychological mechanisms are concerned”

(p.761). For example, research by Deaux, Reid, Mizrahi, and Ethier (1995) has shown that when asked to sort a whole range of social identities1 in terms of perceived similarity, five social identity “classes” emerged and people clearly associated different traits to the different classes. These social identity classes included: Personal relationships (e.g. Wife, Mother, Friend), professional (e.g., Psychologist, Student), political affiliations (e.g. Socialist,

Feminist), ethnic/religious (e.g., Hispanic, Catholic), and stigmatised groups (e.g., Alcoholic, Fat Person). In the experiments presented in this thesis, if discrepancies occurred between the meaning that myself (as an experimenter and a psychology student) attributed to the social identities that were made salient, and the actual meaning that may have been attributed by participants (all first year psychology students), then it is possible that the stimuli developed to induce appraisals and emotions did not work as intended.

1 These social identities were produced via spontaneous self-reports.

In Study 1, the intergroup dichotomy that was made salient was real for participants, namely:

Between the University of Geneva vs. University of Lausanne in the context of the hypothetical merger between the two Universities. Despite the fact that participants in the high social identity condition had been pre-selected based on their actual identification with the University of Geneva (based on their higher scores on the private collective self-esteem subscale), and even though in the manipulation check questions they had reported

significantly stronger identification with the University of Geneva (and with other

memberships related to the University), their appraisals and emotion responses only partially corresponded to the type that were hypothesized. The possibility was evoked in the discussion of study 1, that results were influenced by the fact that the sample consisted of first year students. In contrast to second or third year students, first year students may not have had the opportunity to develop their identification with the University of Geneva. In other words, the meaning first year students attributed to the University of Geneva identity may have not corresponded to the meaning that older students, or myself as a researcher and as a doctoral student, attributed to this identity in the context of the potential merger. In general, it may be that the salience of larger scale memberships such as “Gender” or “Nationality” can have a much stronger impact on appraisals and emotion than that the salience of “University

student”. In fact, in study 1 there was some evidence that some of the other memberships that were not at all associated to the university (e.g., including age, nationality, and sex) were also rated as being highly important and positive in the context of the study. In addition, it has been empirically shown that that University identification can vary as a function of the status of different demographic sub-groups of students, so that it is perceived as being less central by minority status students (i.e., South Asian, Asian, Native Canadian) than by majority status students (i.e., White) (Cameron, 1999). Students at the University of Geneva (and/or their parents) come to different countries, but they also come from different Cantons in Switzerland which may differ in terms of size, language, political status, etc (e.g., see Kreis, 1994).

Despite the fact that no significant differences were found between social identity conditions in terms of the ratio of male/female students and Swiss/non-Swiss nationals, and in terms of mean age and mean number of years residing in Geneva, it is not impossible that other aspects of their diversity could have influenced the effectiveness of the social identity salience

manipulations.

In study 2, the intergroup dichotomy that was made salient was artificial: That of the R&D department vs. the PRODUCTION department. These were assigned roles in the context of

the simulation of a virtual business negotiation. In this way, it was hoped to exert greater control over the content and hopefully the meaning of the social categories in the context of experimental task. However, results of the social identity salience manipulation checks revealed that the outgroup’s behaviour may have had an impact (e.g., it was when the

outgroup’s behaviour was competitive that there were no differences in terms of identification with the R&D). In addition, participants’ own behaviour also may have had an impact (e.g., it was when the outgroup’s behaviour was competitive and participants’ own behaviour was competitive, that there were no differences in terms of appraisal of being caused by other’s PRODUCTION outgroup membership).

In study 3, in was hoped that the effects of social identity salience on appraisals and emotions would become clearer if the magnitude of the intergroup differences were enhanced. This was attempted by crossing an additional and real intergroup dichotomy for participants, namely:

R&D/Psychology students vs. PRODUCTION/Business students. Manipulation check items showed no significant differences between social identity salience conditions. What is more, for participants in both social identity condition identification with the R&D department was clearly dominant in comparison to identification with the Psychology student section. Results on appraisals and emotion showed that it was participants in the low social identity condition that were responding in the way that participants in the high social identity condition were expected to respond. In fact, results were suggesting that the social identity salience manipulations could have had the opposite effect (i.e., reducing perceived intergroup differences for participants in the high social identity condition, and enhancing perceived intergroup differences for participants in the low social identity salience condition). As in study 2, additional analyses revealed that participants’ own behaviour and the interaction with the outgroup’s behaviour may have affected identification with the ingroup. In the discussion of study 3, the content of the manipulations and its administration were examined closely and several factors were identified that could have potentially lead to these unexpected results.

5.2.2. The impact of context, time and ongoing interaction

In studies 2 and 3, it was clear that time had had an effect on social identity salience, just as did participants’ own choice of behaviour and the actual interaction with the pre-programmed behaviours of the outgroup. In the discussion of study 2, it was stated that it could not be assumed that the manipulations of social identity salience administered before the negotiation

task remained effective and consistent throughout the 3 negotiation Events. In criticizing standard methods used in social identity research, Antaki, Condor and Levine (1996) have appropriately pointed out that “ the research setting, and the various actions which take place within it, is usually described as if it were a single definable “context”, a “moment” within which the time can be assume to have stood still […]. As a consequence, analysts assume that any aspect of social identity will be (or in hindsight, has been) uniformly “salient” or “not salient” throughout the duration of the experiment.” (p.477). One way to overcome this methodological problem would be to include measures of indicators that have been found to accurately reflect the effect of social identity salience (e.g., ingroup bias) at selected points of the interaction. However, any attempts to capture changes in salience that may occur “online”

will share the same methodological problems encountered by emotion psychologist when trying to capture appraisal and emotion changes “online”. Another interesting possibility to study the impact of context and time is longitudinal research as conducted by Ethier and Deaux (1994). In their study they examined changes in the Hispanic2 social identity of

students at two Ivy League Universities over their first year, and they found that identification and response to threats of the ingroup depended on the initial strength of participants’

identification to the Hispanic social identity.

5.2.3. The impact of individual differences

Another major issue that emerged from the results of study 1, but also studies 2 and 3, concerns the individual differences that may underlie social identity salience processes.

Though individual differences were not addressed in this thesis, additional analyses conducted in study 1 revealed the existence of two subgroups of participants within the high social identity condition showing different appraisal and emotion response patterns. In studies 2 and 3 there was some indirect evidence that participants may have differed in terms of their tendency to “compete”. In study 2, it was shown that the effects of social identity salience condition on appraisals were different when analysing the sub-sample of participants

systematically opting for competition. Several authors have examined the role of dispositional factors in intergroup processes, including trait self-esteem (as shown by Mussweiler, Gabriel,

& Bodenhausen, 2000; and also by Seta & Seta, 1996), situated goals and motives, and needs for security and self-enhancement (e.g., as shown by research on basking in reflected glory by

2 Students who claimed the Hispanic identity also used terms like Latino, Chicana, Puerto Rican, Mexican-American to describe themselves.

Cialdini, Borden, Thorne, Walker, Freeman, & Sloan, 1976; or by research by Simon,

Pantaleo, & Mummendey, 1995). In his recent critical review of social identity theory, Brown (2000) has reminded us of the importance in recognizing that “there may be both group and individual differences in the propensity to engage in intergroup comparisons” (p.762). In fact, in future research it would be important to assess people’s general orientation to interpersonal and intergroup comparisons by using instruments such as the one developed by Brown and his associates (Brown, Hinkle, Ely, Fox-Cardamone, Maras & Taylor, 1992) assessing

individualist–collectivist and autonomous–relational orientations, or the one developed by Gibbons and Buunk (1999) assessing social comparison orientation.