Investigating the Effect of Group Status on In-Group Identification

The main goal of the present research was to investigate the effect that group status had on different types of identification with social groups. In addition, the study aimed to provide further support for the distinction between centrality, social, communal, and interdependent types of in-group identity. It experimentally manipulated the status of laboratory-based groups in order to examine whether membership in a low status group would be associated with an increase only in social identification and whether this effect would be moderated by culture. Consistent with predictions, the results from a series of 2x2 between-subjects ANOVAs (N=108) revealed that a significant main effect of study condition (group status) occurred in relation to social identification but not in relation to centrality, communal, and interdependent identification. Participants in the moderately positive status group scored significantly higher on social identification than participants in the extremely positive status group. As expected, this main effect was qualified by culture with only collectivistic individuals’ ingroup identification differing significantly between the two conditions. The present work adds to the findings of previous research that has examined the link between group status and in-group identification and could be used to address new issues in the group identification research.


Theoretical Background
There is a general agreement in the group identity research that group identification is a multidimensional construct (Ellemers et al., 1999;Jackson & Smith, 1999;Hinkle et al.,1989;Leach et al., 2008). In a review of the relevant literature in the area (Milanov et al., 2014a) examined four core types of identification with social groups and commented on some specific characteristics and processes that separated these types of identification from other previously investigated theoretical structures of identity. Social identification (Tajfel & Turner, 1979;Turner et al., 1987) is based on the processes of self-categorization and depersonalization, where individuals lose their sense of individuality and perceive themselves as interchangeable members of their group. Centrality (Ashmore et al., 2004;Cameron, 2004;Leach et al., 2008) combines the importance of the group for the selfconcept and the salience of the group and the group membership. Communal and interdependent identification are related to the specific interpersonal interactions through which group members identify with other group members without losing their sense of individuality (Milanov et al., 2014a). The main differences between these two types of ingroup identification are the particular type of relationships (i.e., communal or exchange relationships; Clark & Mills, 1979;Mills & Clark, 1994) between the members of the group.
These relationships shape individuals' expectations towards the group membership and constitute the nature of the interpersonal contacts within the group.
There is some evidence that members of minority (low status) groups tend to enhance their positive group identification by increasing the perceived similarity between in-group members (Simon, 1992;Simon & Brown, 1987). This increase in perceived in-group similarity is intended to secure high self-esteem because homogeneous in-groups are seen to provide more solidarity and social support to their members than are heterogeneous in-groups.
Researchers (Ashmore et al., 2004;Ellemers et al. 1999;Jackson & Smith, 1999;Turner et al., 1987) have operationalized social identification in terms of the perceived similarity between the self and other group members (e.g., I am quite similar to the other people in my group.). In contrast, none of the other three types of identification that are investigated here (i.e. centrality, communal, and interdependent) consider the perception of in-group similarity in their conceptualization. It could be proposed then that only social identification would be affected by changes in in-group status. In the context of different status groups, people Psychological Thought 2020, Vol. 13(1), 240-253 South-West University "Neofit Rilski" https://doi.org/10.37708/psyct.v13i1.438 should react to the threat of membership in a lower status group by enhancing their perceived in-group similarity and, consequently, their social identification.
In addition, it could be expected that people from collectivistic cultures would show significantly higher levels of social identification than people from individualistic cultures because collectivists are less concerned about retaining their sense of individuality in the group (Gardner et al., 1999;Oyserman et al., 2002). Therefore, collectivists are more likely to engage in a type of identification that is associated with depersonalization and perception of similarity with others in the group (i.e., social identification). In terms of the investigated relationship between group status and social identification then, it could be expected that the effects of group status on social identification may be moderated by culture with people from collectivistic cultures being more predisposed to enhance perceived in-group similarity in response to relatively low group status.
Previous research (Branscombe et al.,1999;Jetten et al., 2001) has found that identification increases following low group status under certain conditions. However, these studies confounded different types of identification in their measures. For example, Jetten et al. (2001) investigated the group identification of people with body piercing, and Branscombe et al. (1999) study assessed the group identification of African Americans. In such cases, identification with the group is primarily based on perceived similarities between group members and depersonalization. Hence, it could be assumed that these studies measured only participants' social identification without considering any other types of in-group identity.
In addition, previous research has tended to operationalise low status groups as groups that have failed at a certain task or performed below a certain norm (e.g., Ellemers et al., 1999;Turner et al., 1984). Although failure and low status often go hand-in-hand, it is not necessary to fail at a task in order to be a low-status group. Only doing less well than other groups is sufficient to accrue relatively low status. Furthermore, it is not necessary to be a member of a low-status group in order to fail at a task: High-status groups may fail at tasks without doing any serious damage to their status. Theoretically, failure may have qualitatively distinct effects on identification compared to low status per se because it provides less possibility of in-group improvement (Ouwerkerk et al., 2000). The present study unconfounds status and task failure by making sure that the low status in-group comes second in an intergroup competition, rather than last. This experimental research design would allow the comparison of high and low-status groups independent of failure.

Aims and Hypotheses
The in-group status hypothesis here aimed to provide an important test of the distinction between different types of identification with social groups by revealing the effect of group status on group identity. It only predicted an increase in social identification, not centrality, communal or interdependent identification. More specifically, it was expected that: (1) Participants in the moderately positive in-group condition will show higher social identification than participants in the extremely positive in-group condition and, (2) This effect will be more pronounced for people from collectivistic cultures than for those from individualistic cultures.

Method
The research used a computer-based questionnaire. The status of laboratory-based social groups was experimentally manipulated by producing a moderately positive in-group and an extremely positive in-group.

Design and Participants
The study consisted of a 2 (in-group status: extremely positive/moderately positive) x 2 (cultural background: individualistic/collectivistic) between-subjects design. In order to obtain an approximately equal number of people from individualistic and collectivistic cultural backgrounds (see Oyserman et al., 2002 for individualism/collectivism classification criteria) in each experimental condition, the researcher asked potential participants about their country of origin before making an appointment for their participation. Some potential participants were excluded from the research on the basis of this information.
During a two-month period 122 participants aged 18 years and over were recruited. The vast majority of participants were university students. A manipulation check item showed that 12 participants did not remember the ranking of their group. These participants were excluded from the data analysis. Two more participants were also excluded from the analyses as there was not enough data to decide whether they should be coded as individualists or collectivists.

Procedure
The questionnaire was presented on a computer. The participants were required to attend one research session during which they (1) completed a group generation task, (2) answered several group profile questions, and then (3) responded to a series of dependent measures.
Each research session took approximately 35 minutes to complete.
In the group generation task, the researcher asked participants to generate a group of five people that included themselves and four other family members and/or friends. This group was called the RED group. Then, participants responded to four group profile questions that asked how many people in their group, including themselves, possessed certain characteristics (e.g. "How many members of your group are male?", "How many members of your group have dark hair?", etc.).
In order to experimentally manipulate in-group status, participants were informed that the computer would compare their responses with the responses obtained from four other people who had recently taken part in the same study and responded to the same questions. The groups of the other four people were presented as the BLUE, GREEN, ORANGE, and YELLOW groups. The computer then provided a ranking of the five groups. This ranking was supposedly based on each group's match to a wining group profile that had been randomly generated by the researcher before the study commenced. Participants read that the group that was ranked in first place was the group that most closely matched the winning group profile, the group that was ranked in second place was the next closest matching group, and so on. A percentage score appeared next to each group in the list to indicate how closely each group matched the winning group profile. A highlighted message followed the ranking, advising participants to spend few moments memorizing their group position because this information would need to be recalled later on.
In fact, there was no winning group profile, and the four other participants and their groups that were supposedly involved in the competition were fictitious. This bogus situation allowed the computer to randomly assign participants' groups to either first place (extremely positive in-group status) or second place (moderately positive in-group status) without arousing the suspicions of the participants. Participants were then asked to complete a questionnaire while thinking only about the group that they generated in the group building task. The main part of the questionnaire was a 20item scale that measured four different types of identification with social groups (Milanov et al., 2014b). Example items from this scale include: "My group is important to my sense of who I am" (centrality), "I am quite similar to the other people in my group" (social identification), "I empathize with the other people in my group" (communal identification), and "When I give something to another person in my groups, I generally expect something in return" (interdependent identification). Participants responded to each statement using a fivepoint Likert-type scale (1= strongly disagree, 5=strongly agree).
People then provided their age and gender and responded to a measure of culture. This measure of culture incorporated four items which taped both objective and subjective information at the level of specific countries, cultural backgrounds and languages. These items gave combined information about broader individualistic and collectivistic cultural variations. The questionnaire also included three manipulation check items and group feeling items (e.g., "What position was your group in the ranking?", "How good do you feel about your group?"). This gave an opportunity to investigate different aspects of the in-group status manipulation effect.

Factor Structure of the Types of Identification Scale (TIS)
The combined results of a principal axis factor analysis with promax rotation (item loadings ≥ .30 as cut-of criteria), a scree-plot test (Cattell, 1966), and a parallel analysis (Horn, 1965;Watkins, 2000) confirmed the four-factor structure of the identification measure. The four factors accounted for 51.20% of the total variance and were structurally consistent with prior theory in relation to centrality, social, communal and interdependent identification.
Cronbach's alphas for all of the subscales ranged between 0.6 and 0. 8 and were in the diapason recommended by Clark & Watson (1995) for good or adequate reliability.

Variations in Types of Identification as a Function of Ingroup Status
To examine the effect of in-group status and culture on each of the four types of in-group identification, a series of 2 (in-group status: extremely positive/moderately positive) x 2 (culture: individualistic/collectivistic) between-subject ANOVAs on each of the TIS' subscales Psychological Thought 2020, Vol. 13(1), 240-253 South-West University "Neofit Rilski" https://doi.org/10.37708/psyct.v13i1.438 were performed. In-group status was represented by the experimental conditions (in-group ranked first, in-group ranked second). Culture was represented by each of the four culture items (nationality, country, language, and cultural background).

Results based on nationality
When culture was based on nationality, there was a significant main effect of ingroup status on social identification , F(1, 104)

Results based on country of origin
When culture was based on country of origin, the pattern of results was similar to that reported above. There was a significant main effect of ingroup status on social identification, Their effects of group status on centrality, communal, and interdependent identification were nonsignificant (ps > .17).

Results based on cultural background.
When culture was based on cultural background, the results were partly different to those above. Consistent with the previous results, there was a significant main effect of ingroup status on social identification, F(1, 72) = 5.93, p = .02, η p ² = .08. Again, participants in a moderately positive status condition scored higher on social identification (M = 3.11) than participants in the extremely positive status condition (M = 2.74). As before, this main effect was qualified by a two-way interaction between status condition and cultural background,

Results based on language.
When culture was based on language, the results were relatively similar to those based on nationality and country of origin. As before, there was a significant main effect of study condition on social identification, F(1, 104) = 6.70, p = .01, η p ² = .06. Participants in the moderately positive status condition scored higher on social identification (M = 3.13) than Psychological Thought 2020, Vol. 13(1), 240-253 South-West University "Neofit Rilski" https://doi.org/10.37708/psyct.v13i1.438 participants in the extremely positive status condition (M = 2.77). Unlike previous results, however, this main effect was not qualified by a significant interaction between status condition and language (p = .20). Their effects of group status on centrality, communal, and interdependent identification were nonsignificant (ps > .18).

Results based on a combined index of culture
Although the previous analyses revealed a fairly consistent pattern of results using different measures of culture, there were some discrepancies. In order to obtain a more reliable analysis of the interaction between group status and culture, a Combined Index of Culture was created and used as a categorical variable. More specifically, the scores from the nationality, country, language, and cultural background responses were summed to form a

Discussion
Based on previous research (Simon, 1992;Simon & Brown, 1987;Turner et al., 1984), it was proposed that people would react to low in-group status by increasing their social identification with the group. Therefore, in the analyses, it was of particular interest to reveal the effect of in-group status on each of the four subscales of the identification measure.
Consistent with predictions, the results showed a significant main effect of study condition (in-group status) on social identification. Participants in the moderately positive status condition scored significantly higher on social identification than participants in the extremely positive status condition. As expected, the main effect of in-group status tended to be nonsignificant for centrality, communal, and interdependent identification.
There was a slight discrepancy in the results based on one of the different operationalisations of culture: When culture was based on cultural background, participants in the extremely positive status condition showed significantly higher interdependent identification than participants in the moderately positive status condition. In addition, participants in the extremely positive status condition showed significantly lower communal identification than participants in the moderately positive status condition. However, these main effects were not significant when culture was based on nationality, country of origin, language, or the Combined index of culture. The above difference may be explained in terms of the coding of the data obtained from the cultural background item. Based on this item, 33 participants indicated having "mixed" (individualistic/collectivistic) or "other" cultural background. Because these responses were coded as missing, this is likely to have affected the reliability of the results based on this measure of culture.
Further analyses of the effect of in-group status on social identification revealed that this effect was moderated by culture. Collectivistic participants in the moderately positive status condition had significantly higher levels of social identification than collectivistic participants in the extremely positive status condition. In contrast, there was no significant difference in individualistic participants' social identification in both status conditions. It is possible to explain this result in terms of cross-cultural differences in reactions to low in-group status.
Previous research (Heine et al., 2001;Heine et al., 1999;Heine & Renshaw, 2002) suggests that people from individualistic cultures are more prone to individual self-enhancement because of its perceived beneficiality and functionality in society, whereas people from collectivistic cultures "focus more on maintaining positive evaluations of their groups" (Heine Psychological , 1999, p. 783) and are "more likely to view their self and their performance as potentially improvable" (Heine et al., 2001, p. 606). Consistent with this interpretation, it is possible that collectivistic participants in the present study were more prone to enhance their social identification with their group because they were more concerned about facilitating their group's performance.

Study Limitations
Couple of limitations of the present study should be taken into consideration. First, the group generating task in the group status manipulation did not involve different types of groups (see Lickel et al., 2000). Instead, participants were asked to think only about a group that consisted of close friends and family. It is possible then that the use of this specific group as the sole target in the study may have affected participants' responses on the identification measure by enhancing one particular type of identification at the expense of other types of identity.
Second, the present research uses only one set of data from mainly student participants to tests its main hypothesis. Given the revealed interactions between group status and culture, larger socioeconomically and culturally diverse samples are needed in order to provide more detailed evidence for the effect of group status on different types of identification with social groups.

Conclusion
Overall, the pattern of variations in the analyses confirmed the prediction that an increase in identification as a response to low in-group status occurred only in relation to social identification and not in relation to centrality, communal or interdependent identification. The result showed that membership in a low status group enhances the perceived similarities between group members, and consequently, their social identification with that group.
However, types of in-group identification that do not involve the perception of similarity between members (i.e. centrality, communal, and interdependent identification) remain unaffected by differences in the status of the salient group. In addition, the occurring changes in identification as a result of different group status appeared to be affected by aspects of culture. Individualistic participants' social identification did not differ significantly between the moderately positive and the extremely positive status conditions. In contrast, collectivistic participant increased their social identification with their group in the lower status condition when the need for improvement or better performance was made salient. It can be concluded Psychological Thought 2020, Vol. 13(1), 240-253 South-West University "Neofit Rilski" https://doi.org/10.37708/psyct.v13i1.438 then, that the relationship between group status and in-group identification depends on (a) the specific type of in-group identification that is being investigated and (b) the particular culture of the identifying group members. Future research in this area should continue to consider each of these variables more carefully because studies in this particular direction have the potential to address new issues in group identification research.

Funding/Financial Support
The author has no funding to report.

Other Support/Acknowledgement
The author has no support to report.