Psychology of ethnic discrimination

Размещено на http://www.allbest.ru/

Размещено на http://www.allbest.ru/

School of Social Behavioral Science

Psychology of ethnic discrimination

Work and organizational psychology

Student: Aleksandra Kotsuba

Supervisor: Dr. Byron G. Adams

Abstract

It is crucial for international companies to support the work attachment of expats in order to enhance their performance. The current study attempts to shed light on this problem by focusing on how induced thoughts about discrimination and exclusion could affect participants' level of ethnic and organizational identity. The participants were expats of different nationalities who are currently working in Russia. It was found that when expats have a reason to think that they are being discriminated on the basis of their ethnicity, they can undergo the following changes after the manipulation: an increased ethnic identity and a decreased organizational identity. However, this effect was not observed in the participants who were a subject of exclusion treatment. Within this manipulation, the participants were indirectly made to feel excluded, with the cause of the exclusion being non-specified. It was revealed that the level of both identities did not change significantly. These results confirm the need of companies to prevent discrimination and deal with its potential causes since they can lead to negative consequences both for the employee and for the organization.

Keywords: discrimination, exclusion, treatment, organizational identity, ethnic identity, experiment, expats, Russian working environment

The effect of exclusion and discrimination on ethnic and organizational identity in the workplace

With the increase in internationalization, expatriation and mobility, the possibility of people experiencing exclusion or discrimination from the majority and host members is becoming more and more probable (Wodak, 2008). Experiencing discrimination and exclusion is believed to negatively affect the way people identify with their organizations as it prevents them from perceiving themselves as a part of it (Gioia, Schultz & Corley, 2000; Lindgren & Wеhlin, 2001). This results in them often seeking support from their own group (i.e., ethnic, cultural, and national groups) as a means to counter the effect of their discriminatory or excluded experiences which, in turn, leads to an increase in their in-group identity. This phenomenon is explained by the "Rejection-Identification Model (RIM)" (Branscombe, Schmitt, & Harvey, 1999). Since only a few experimental studies have been carried out to confirm this effect, the following research question is formulated: How can exclusion and discrimination treatment affect expats' in-group (ethnic) and out-group (organizational) identities? The purpose of this research is to examine the effects of exclusion and discrimination treatment on ethnic and organizational identity of employees of different nationalities in the context of the Russian labor market. Moreover, due to the fact that no one yet has examined to what extent the effect of discrimination on identities differs from the effect of exclusion on it, since they are somewhat different concepts (Leets & Sunwolf, 2005; Jackson, Brown, & Kirby, 1998), I aimed to determine this as well.

The literature review presents the main constructs which have been studied within the research and the relation between them. In particular, it covers the concept of social identity, which includes ethnic and organizational identities. Furthermore, it presents the concepts of exclusion and discrimination, how they are connected with one another and what impact they have on ethnic and organizational identities. The final part of literature review takes a look at the hypotheses which seek to address the gaps in current knowledge in this field.

Social identity in the organization

Social identity is a concept used to explain how people define themselves based on their membership within different groups (Tajfel, 1974) by focusing on their similarities with the in-group and differences with the out-group. It is important to say that it is natural for every human to attribute themselves to a certain social group to which they belong and with which they share similarities (Baumeister & Leary, 1995). However, with some groups it can be done easier than with others (Tajfel, 1974). If people are not able to identify themselves with certain social groups due to some reason, they can compensate it through stronger identification with another social group (Ashforth, 1985). In this research discrimination and exclusion are regarded as reasons which can affect social identity. In particular, two forms of social identity (ethnic and organizational) are considered.

Ethnic identity is defined by how people identify themselves in terms of ethnicity (Phinney et al., 2001). Ethnic identity becomes stronger when people feel as a part of their ethnic group. Furthermore, strong ethnic identity positively correlates with a high level of life satisfaction and a high level of self-esteem (Mшllersen & Holte, 2008; Smith & Silva, 2011). Therefore, it can be concluded that it is essential to develop a strong ethnic identity in order to enhance people's well-being,

Organizational identity is defined by how a person is affiliated and sees himself/herself in the organization (Ashforth & Mael, 1989). It allows employees to figure out who they are and for what they stand as an organization. It is based on organizational values and principles which relate to interpretation of problems that the company faces, the way of communication, resolving conflicts, etc. (Voss & Cable, 2006).

In order to form an organizational identity, a company must promote its history, be able to differentiate itself from similar organizations and rely on deeply rooted company values (Whetten, 2006). According to findings of Gioia, Schultz and Corley (2000), organizational identity becomes stronger when employees feel as a part of their organization. It can be seen when an organization's values or principles match an employee's self-concept (beliefs about themselves). Organizational identity is very important for companies since it reflects employees' commitment, engagement and job satisfaction (Dutton et al., 1994).

Exclusion and discrimination in the organization

Mere exclusion is the opposite concept of inclusion and it is characterized by a lack of sense of belonging to the community and lack of participation in formal and informal social networks (Kurzban & Leary, 200; Brewer, 1991). Robinson and his colleagues (2013) showed that a person's self-perception of exclusion most often represents the real situation where people are actually excluded by their group members (e.g. they do not talk to them). In the context of a workplace, excluded employees are usually not invited to any corporate events organized by their colleagues. In addition, they can also be excluded from job related tasks. For example, co-workers can ignore their ideas and suggestions during discussions of work projects (Robinson et al., 2013). Some possible reasons for exclusion can be their appearance, behavior, personal characteristics and etc. Since these aspects are changeable, people are able to adjust them and, thus overcome exclusion and become included again (Brewer, 1991).

Exclusion can be also viewed as a form of discrimination (Wodak, 2008) in which people are excluded from social contacts because of their gender, race, religion, age and etc. However, discrimination, in contrast to mere exclusion, leads to people being excluded only because of their group-membership, which is nearly impossible to change (Jackson, Brown, & Kirby, 1998). Another definition of discrimination is unfair and unequal treatment of people because of their group belonging (Simpson & Yinger, 2013).

It is important to examine exclusion and discrimination since they can have negative effects in many aspects of social life (e.g. work, university, school and etc.). For example, they both can affect people's psychological health. It is proven that excluded and discriminated people tend to experience low self-esteem, helplessness, and lower affective commitment (Eisenberger et al., 2003; Triana et al., 2015). In a working setting excluded and discriminated employees have low job-satisfaction, and turnover intentions, as a result they tend to manifest antisocial reactions, dishonesty, social loafing and job withdrawal (Kouchaki & Wareham, 2015; Ferris, Brown, & Heller, 2009; O'Reilly et al., 2015). Another unfavorable effect of both concepts is that they can lead to an increase in conflicts, decrease in confidence and difficulties in communication with other people (Tajfel & Turner, 1979).

However, the negative effect of discrimination may be stronger than the effect of mere exclusion. In particular, people tend to discriminate others because of negative stereotypical expectations about the discriminated people's behavior, abilities and social status associated with their group membership (Buengeler & Den Hartog, 2015). These negative stereotypes complicate the process of dealing with discrimination. For example, people of certain ethnicity can be treated unfairly by out-group members because of the negative stereotype about their ethnic belongingness (Robson & Breems, 1985; Wortley, 1996). Since discriminated people cannot change their ethnicity, they are struggling to deal with the negative attitude of the out-group members toward them. Moreover, this frustrates them and negatively affects their well-being (Triana et al., 2015). However, in the case of exclusion, people can easily cope with it since they might know why they are excluded and how to change the current situation in order to get back into the group (Brewer, 1991). Moreover, people who are excluded from the group are just denied social contact by their group members and are not necessarily exposed to negative treatment (Kurzban & Leary, 2001). This is also the reason why the effect of discrimination can be stronger than that of exclusion. Overall, exclusion and discrimination are common concepts that can be seen in everyday life and they can lead to negative outcomes to different extents.

Exclusion, discrimination, ethnic and organizational identity

According to the RIM model, exclusion/discrimination by members of out-group is perceived less negatively by minorities who identify themselves stronger with their own in-group (Branscombe et al., 1999). Therefore, people tend to identify with their in-group much stronger and with the out-group much weaker when they feel excluded/discriminated. Within this research, the organization is considered as the out-group, while the ethnic group is considered as the in-group.

There is substantial evidence that people's social identities are not constant and can be strengthened or weakened by certain situations in which they feel discriminated in relation to their ethnicity (Breakwell, 2015). There can be two opposite effects - in the first case it can have a weakening effect on the ethnic identity when people perceive discrimination. In the second case the effect might be strengthening, which goes in line with the RIM model. In the first case, minorities which experience discrimination at work concerning their ethnicity can feel ashamed of their affiliation and undergo minimization of the significance of their ethnic group (Phinney, Chavira, & Tate 1993). They go through the process of socialization and try to achieve identification with the out-group (organization identity) which in turn leads to weakening of ethnic identity (Spencer & Markstrom-Adams 1990). However, it should be noted that little research of this effect has been done and it is safe to say that it occurs with a significantly lower probability than the second one.

In the second case some studies demonstrate the moderating effect of ethnic identity on the perception of discrimination (Noh, Beiser, Kaspar, & Hou, 1999; Dion & Pak 1992). That is, those minorities that have a strong ethnic identity (for example, preserve the language, ethnic values and traditions) may perceive racial discrimination from the out-group less negatively. Therefore, they often create communities based on their ethnicity where they are protected from discrimination and can feel free to express themselves (Dion & Pak 1992). As it was mentioned in RIM model, discrimination, as a rule, not only enhances one's own ethnic identity, but it also weakens the out-group identity (Branscombe et al., 1999). For example, migrants experiencing discrimination are less likely to develop an identity with respect to the host country. On the contrary, migrants who do not experience discrimination can strengthen the identity with the out-group and weaken the identity with the in-group (Erikson, 1968).

According to Gioia, Schultz and Corley (2000) the relationship between discrimination and organizational identity is negative since discriminated employees feel undervalued and fail to adapt in the workplace. Moreover, members of minority groups start focusing on their differences from the members of the out-group and the company itself. As a result, their organizational identity weakens even more (Lindgren & Wеhlin, 2001). That is to say, organizational identity is a dynamic and changing feature that can depend on discrimination at work.

Thus, according to the RIM model and the theories mentioned above, a conclusion may be drawn that exclusion and discrimination affect the social identity of employees in an organization in a similar way (strengthen the ethnic identity and weaken the organizational identity). However, since this effect was still not proven in experimental conditions it will be addressed in this study. Moreover, since it was mentioned that exclusion and discrimination are somewhat different concepts, that can lead to negative outcomes to different extents (Leets & Sunwolf, 2005; Jackson, Brown, & Kirby, 1998), it was assumed that their impact on both identities can also be different. In particular, a person discriminated due to his ethnicity can be more affected by that situation since it is more difficult for him to deal with this exclusion (Robson & Breems, 1985; Wortley, 1996). Therefore, compared to mere exclusion, a discriminated person might seek more support from his own ethnic group, than from his organization (Noh, Beiser, Kaspar, & Hou, 1999; Dion & Pak 1992). Taking into consideration the gaps in current knowledge in this field both hypotheses are formulated:

Hypothesis 1: Non-Russian employees after an exclusion and discrimination manipulation will experience lower organizational identity than before manipulation (a) and in discrimination condition its effect will be stronger than in exclusion condition (b).

Hypothesis 2: Non-Russian employees after an exclusion and discrimination manipulation will experience higher ethnic identity than before manipulation (a) and in discrimination condition its effect will be stronger than in exclusion condition (b).

Method Sampling and data collection

According to the power analysis done in G-power (Faul, Erdfelder, Lang, & Buchner, 2007) for Mixed ANOVA (repeated measures, between factors) with effect size of.25, I required a sample size of 120 participants to get a statistically significant power of.80 with an б level of.05. In order to account for complete measure, I aimed for higher significant power. Therefore, data were collected from 231 working expats in Russia. Working expats were found via Intermark which specializes in providing services for companies, who require assistance in relocation and immigration of expats from all around the world to Russia (Intermark, 2008) and via different groups such as “Expats In Russia” in social networks (Facebook and VK). The requirement set for the participants was to be a non-Russian citizen living and working in the Russian Federation. After collecting the data, Missing Value Analysis was performed (Supplementary file 1), after which 80 participants, who did not fill the questionnaire fully, were excluded (49 people followed the link without doing the survey; 31 out of 80 people left the scale non-filled the second time the identities were measured after the treatment, with 3 of them also not answering the control questions about exclusion, and 6 of them not answering the control questions about discrimination). As a result, the total sample was represented by 151 working expats (73% men, Mage= 29.39 years, SD = 7.40). It was divided into 37 participants in the discrimination group (76% men, Mage=29.16 years, SD=6.71), 42 participants in the exclusion group (64% men, Mage=29.48 years, SD=8.35) and 72 participants in the control group (76% men, Mage=29.46). The data concerning nationality, religion, years of experience in a current organization, years of experience as an expat in total and in Russia are provided in the Table 1 below.

Design

This study followed the mixed factorial design, which is a combination of between and within subjects factors. The within-subjects factor is Time (measurement of ethnic and organizational identity before and after the treatment of each condition) and the between-subjects factor are conditions (measurement of differences between discrimination, exclusion and control conditions). Following this design, participants were recruited and asked to click a web link to an online questionnaire at the “Qualtrics” survey tool which assigned them to one of these three conditions. As can be seen from descriptive statistics, the distribution of participants in each group was unequal. It resulted from the fact that some participants followed the link with the questionnaire and, being automatically assigned either discrimination or exclusion condition, closed it without filling anything. This led to a lot of participants being assigned to control condition. It is important to note that refusal to participate was definitely not caused by the assigned condition since participants were not aware of it.

By including the control group, which receives neutral treatment, I allow to examine the effect of independent variables (exclusion or/and discrimination conditions) on dependent variables (Ethnic and Organizational identity) holding all the other factors fixed. By this I prove that only exclusion and discrimination treatments can affect the level of both identities. Also, including the control group enables me to perform manipulation check. Thus, it can allow to check that neither does neutral treatment affect the levels of perceived exclusion in contrast to exclusion treatment, nor does it affect the levels of perceived discrimination in contrast to discrimination treatment.

Procedure

In the first step of the experimental procedure, participants of the 3 groups filled questionnaires which measured the initial level of their ethnic and organizational identity.

In the second step participants from all groups were asked several sociodemographic questions. Furthermore, in order to increase the time between the first and the second measurement of Ethnic and Organizational identity, working expats were asked to describe the reason why they decided to move to Russia.

In the third step, the exclusion and the discrimination groups were asked to read an article published in the newsletter, devoted to the problem of poor performance among employees because of exclusion and/or discrimination of non-Russian people by Russians (Appendix A; Appendix B). They were informed that this article was important to read since it was written by their sister-company (with similar corporate culture, values and company goals), therefore the problem which they faced could also arise in their company. Then, the participants were asked to imagine that they were in the position of the person from the article (excluded or discriminated) and were asked to think about how it might make them feel. Moreover, in order to increase the role immersion, expats wrote a text about their personal experience of discrimination/exclusion in their workplace. The participants from the control group read a content-neutral text about an effective training program on Time Management at their sister-company (Appendix C). They were also told that it was important to read it since such training might be useful for them as well.

In the fourth step, participants from each group rated several items in order to check the efficiency of exclusion or/and discrimination manipulation. The exclusion group answered three questions about exclusion. The discrimination group answered three questions about discrimination. The control group was asked to rate all six questions related to both discrimination and exclusion (see measures for control variables).

In the last step, the level of organizational and ethnic identity among all participants was measured again. The research tools for measuring dependent variables and control variables which were used for manipulation check (as well as the sources of these tools, data on their validity and reliability) are reflected bellow.

Measures

Sociodemographic variables. Questions like age, gender, nationality, religion, work experience in current organization were asked. There were also questions about how long the participants have been working in Russia as an expats and how long they have been working as an expat in any other country.

Dependent variables: Organizational identity was measured with the six items used by Mael and Ashforth (1992) (e.g., “When someone criticizes my firm, it feels like a personal insult”). The evaluation was carried out using a 5-point scale from 1 (strongly disagree) to 5 (strongly agree) (Appendix D). The reliability of the scale among all groups was above 0.62 (see Table 2). Ethnic identity was measured with the 6 items used by Phinney (1992). This measure included two scales: ethnic identity search (e.g. “I have put effort in learning more about my ethnicity”) and commitment to the ethnic group (e.g. “I feel a strong attachment toward my ethnic group”). The evaluation was carried out using a 5-point scale from 1 (strongly disagree) to 5 (strongly agree) (Appendix E). Cronbach's б of Ethnic identity among all groups was above 0.80 (see Table 2).

Control variables: Level of exclusion was measured with three items: To what extent do you feel excluded by the company's members? Do your team members have the desire to include you in common work related projects? Do you feel generally accepted by your team members? The evaluation was carried out using a 7-point scale from 1 (very much) to 7 (not at all). The Cronbach's б of the scale in each group was above 0.64 and KMO value was above 0.53, which levels were tolerable (Griethuijsen et al., 2014; Cerny & Kaiser, 1977). Level of discrimination was measured with three items: To what extent do you feel discriminated by your company's members? Do your team members have the desire to include you in common work-related projects, despite your nationality and all the stereotypes related to it? Do you feel generally accepted by your team members despite your nationality? The evaluation was carried out using a 7-point scale from 1 (very much) to 7 (not at all). The Cronbach's б of the scale in each group was above 0.62 and KMO value was above 0.52, which levels were tolerable (Griethuijsen et al., 2014; Cerny & Kaiser, 1977). The examples of control variables were found in the article of Mok and De Cremer (2015).

Ethical Consideration

Before conducting the experiment, the participants were informed about the procedure and their participation was voluntary. Therefore, the participants had the option to refuse to participate at the beginning or at any time during the experiment. The study was anonymous and required only socio-demographical data from the participants. After finishing the experimental procedure, the participants read the debriefing part devoted to the aim of this study (Appendix F). Furthermore, the participants had an opportunity to talk with the researchers about their feelings after the experiment.

Results

Plan for Primary analysis

The analysis of the gathered data was performed in SPSS 22. Before testing the hypotheses, a primary analysis was conducted in order to see whether the data was suitable for further analysis or not. Firstly, the reliability of dependent variables was checked by Cronbach's б and KMO. Secondly, the means of these variables were measured using MANOVA test in order to see differences between groups. Thirdly, dependent variables were correlated between each other in each group and in total. For the correlation the Pearson's correlation coefficient was used since it detects any significant relations between the variables. Fourth, a manipulation check was conducted in order to ensure that the treatments were working for each group. Finally, several assumptions of Mixed ANOVA test were analyzed before testing hypothesis.

The Factor analyses

The reliability tests were performed in order to verify that dependent variables of each group could be used for further research. It was found that in the exclusion group both Organizational identity and Ethnic Identity (before and after) had KMO value above.70, Bartlett's Test of Sphericity was significant (p <.001), and analyses were explaining above 60 % of total variance. The exclusion group had quite similar measures: KMO was above.70, Bartlett's Test of Sphericity were significant (p <.001) and analyses were explaining above 49 % of total variance. Regarding the discrimination group, the reliability of its variables was also good: KMO was above.70, Bartlett's Test of Sphericity was significant (p <.001) and all variables explained no less than 57.50% of total variance, despite of organizational identity' variable which were conducted before manipulation. This variable had KMO value =.40, which is not desirable (Cerny & Kaiser, 1977) but the Bartlett's Test of Sphericity was significant (p <.001). Furthermore, according to Cronbach б the reliability of this scale was above.62, which can be seen as sufficient (Griethuijsen et al., 2014). Moreover, since I did the repeated measure and could not easily delete some unsuitable items of this specific scale, as it can affect the reliability of other similar scales (before and after organizational identity of all groups), I decided to keep the scale at this level. The results of KMO reliability test can be seen in Supplementary file 2. The reliabilities expressed by Cronbach's б can be seen in Table 2.

One-way MANOVA was performed, and the analysis conducted revealed that there were no significant differences between exclusion, discrimination and control conditions on the levels of Ethnic and Organizational identity.

Correlations

The table below presents the correlations between dependent variables, where no correlation was higher than.94. The only significant relationships were the ones between each identity's before and after levels among all groups and in total. There was not found significant correlations between organizational and ethnic identity in all levels.

Manipulation check

For the manipulation check, I first performed an independent samples t-test in order to compare the mean exclusion levels between exclusion and control conditions. The test showed a significant difference between the mean levels of the exclusion group (M = 3.33, SD = 1.32) and the control group (M = 2.67, SD = 1.12; t(112) = 2.62, p =.010, two-tailed). The magnitude of the differences in the means (M =.63, 95% CI:.15 to 1.11) was moderate (зp 2 =.058). These results confirmed that exclusion was perceived less by participants in the control group as compared to the ones in the exclusion group (Table 4).

In order to compare the mean discrimination levels between the discrimination and the control groups, I ran a second independent samples t-test. The test showed that there was a significant difference between the mean levels of the discrimination group (M = 3.57, SD = 1.12) and the control group (M = 2.70, SD = 1.21; t(107) = 3.56, p =.001, two-tailed). The magnitude of the differences in the means (M =.86, 95% CI:.38 to 1.34) was moderate (зp 2 =.106). These results confirmed that discrimination was perceived less by participants in the control group as compared to the ones in the discrimination group (Table 4).

discrimination ethnic organizational identity

Assumptions for Mixed ANOVA

Before using Mixed ANOVA several underlying assumptions had to be checked. One of these was to check for any outliers which could reduce the accuracy of this study's results by distorting the differences between the conditions. It was found 8 outliers in the dependent variables among the 3 conditions. However, after conducting the analysis with and without outliers, it became evident that the results for the research goal did not change and for this reason it was decided to keep the outliers within the sample. Another assumption required a check of dependent variables (before and after among the 3 conditions) for normal distribution. The analysis conducted showed that normal distribution was violated. However, since the deviation was not significant and Mixed ANOVA is not very sensitive to moderate deviations from normality (Kalla, 2011), it was concluded that the dependent variables were approximately normally distributed. For the last assumption Levene's test was performed in order to check for homogeneity of dependent variables. This assumption was satisfied since Sig. statistics for both variables (before and after) were all higher than.05, which demonstrated homoscedasticity in all the 4 variables.

Hypothesises testing

In order to check Hypothesis 1, stating that Non-Russian employees after exclusion and discrimination manipulation will experience lower organizational identity than before manipulation (a) and that in discrimination condition its effect will be stronger than in exclusion condition (b) Mixed ANOVA test was conducted. The results of this analysis showed that the main effect comparing the 3 conditions in terms of organizational identity showed no significance (F(2, 148) =.11, p =.898, зp 2 =.001). Furthermore, according to the Wilks' Lambda test the main effect of time was significant [Wilks' =.94, F(1, 148) = 8.85, p =.003, зp 2=.056] with the three groups in total experiencing a decrease in the level of organizational identity across the two time periods examined. However, Wilks' Lambda showed that there was no significant interaction effect between time and condition to which participants were assigned [Wilks'=.97, F(2, 148) = 2.29, p =.105, зp 2 =.03] with medium effect size. The Wilks' Lambda for the main effect of time and the interaction effect can be seen in Table 5.

Table 5

Wilks'Lambda test for organizational identity

Note: **indicate significance at the 5 % level

In spite of the fact that there was no significant interaction effect the Pairwise Comparisons table showed that only the discrimination group experienced a significant decrease in the level of organizational identity after the manipulation with (M =.171). Wilks' Lambda showed following effect [Wilks' =.95, F(1, 148) =7.90, p =.006, зp 2=..051]. In contrast, the exclusion and control groups experienced a non-significant decrease in the level of organizational identity with (M =.095), [Wilks' =.98, F(1, 148) = 2.78, p=.098, зp 2=.018] and with (M =.014), [Wilks' =.99, F(1, 148) =.101, p=.751, зp 2 =.001] respectively. These results proved that Hypothesis 1a was partially supported only in relation to the discrimination manipulation but not in relation to the exclusion one. However, Hypothesis 1b was not supported since the interaction term was not significant (p=.105) and it will be incorrect to say that the discrimination treatment affected the levels of organizational identity significantly differently from the exclusion treatment. Possibly, an interaction effect could be found if the discrimination condition group had more participants in which case the effect size would be higher than the current one (зp 2=..051). The results of the Pairwise comparisons and Wilks' Lambda for each condition can be seen in (Table 6, Fig.2) and (Appendix G) respectively.

Table 6

Pairwise Comparisons for organizational identity

Figure 1. The level of organizational identity before and after the treatment in each condition

Mixed ANOVA test was performed in order to assess Hypothesis 2, stating that Non-Russian employees after an exclusion and discrimination manipulation will experience higher ethnic identity than before manipulation (a) and that in discrimination condition its effect will be stronger than in exclusion condition (b). The results show that the main effect comparing the 3 conditions in terms of ethnic identity showed no significance (F(2, 148) =.27, p =.765, зp 2 =.004). However, according to the Wilks' Lambda the main effect of time was significant [Wilks' =.95, F(1, 148) = 8.46, p =.004, зp 2 =.054] with the three groups in total experiencing an increase in the level of ethnic identity across the two time periods examined. Moreover, the interaction effect between time and the condition to which participants were assigned was also significant [Wilks' =.94, F(2, 148) = 4.69, p =.011, зp 2 =.06]. The Wilks' Lambdas for the main effect of time and the interaction effect can be seen in Table 7.

Table 7

Wilks'Lambda test for ethnic identity

Note: **indicate significance at the 5 % level

Taking into account the significance of the interaction term, I can conclude that the treatment affected the levels of ethnic identity between the three groups differently. Moreover, the discrimination treatment was the main cause of the significant interaction effect which was found. It can be seen in The Pairwise Comparisons table which demonstrated that only the discrimination group experienced a significant increase in the level of ethnic identity, with (M =.243) after manipulation. Wilks' Lambda showed that this effect was significantly strong [Wilks' =.91, F(1, 148) = 14.44, p <0.001, зp 2=.089], which could not be said for the other 2 groups. The exclusion group and control group experienced a non-significant increase in the level of ethnic identity, with (M=.036), [Wilks' =.99, F(1, 148) =.35, p=.553, зp 2=.002] and with (M =.009), [Wilks' =.1, F(1, 148) =.041, p=.840, зp 2=.000] respectively. These results supported Hypotheses 2a concerning the discrimination treatment but not the exclusion treatment. Furthermore, since the desired effect was found only in the discrimination condition Hypothesis 2b was supported as well. The results of Pairwise comparisons and Wilks' Lambda for each condition can be seen in (Table 8, Fig.2) and (Appendix H) respectively.

Figure 2. The level of ethnic identity before and after the treatment in each condition

Discussion

This research investigated how and to what extent the levels of ethnic and organizational identity of expats, primed by exclusion or discrimination treatment, could be changed. Furthermore, since the differences in the extent of both discrimination and exclusion effects on identities had not been deeply studied before, this research aimed to analyse this as well. Moreover, to our knowledge these effects were not examined in experimental conditions. Therefore, I sought to close this gap by conducting a study in which participants were randomly assigned to discrimination, exclusion or control condition. In doing so, I wanted to examine how induced feelings of discrimination and exclusion affected both identities. The manipulation check showed that expats who were assigned to the discrimination condition actually felt more discriminated than those who were assigned to the control condition. Furthermore, participants who were assigned to the Exclusion condition felt more excluded compared to the participants from assigned to the control condition. These results confirmed that expats who were induced with the thought of discrimination or exclusion were affected by this manipulation in comparison to the control group which had neutral treatment.

After verifying that the manipulations were effective, I turned to the main research question. The results of this study showed that discrimination treatment influenced both identities by increasing the level of ethnic identity and decreasing the level of organizational identity. This can be explained by two reasons presented below:

In general, discrimination is negatively perceived by people. Firstly, people who are discriminated or, as in case of this study, are induced with this feeling, have difficulties experiencing it because of their inability to change their ethnicity in order to be accepted by out-group (Robson & Breems, 1985; Wortley, 1996). Secondly, discrimination itself negatively affects peoples' well-being resulting in health problems, low self-esteem, helplessness and decrease of self-confidence (Eisenberger et al., 2003; Triana et al., 2015). This leads people to try and diminish the negative effects in different ways. Several studies show that discrimination can affect people less negatively if they identify themselves stronger with their ethnic group than with the out-group (Branscombe et al., 1999; Noh, Beiser, Kaspar, & Hou, 1999; Dion & Pak 1992). Furthermore, stronger identification with their ethnic group occurs when they are seeking support and protection from discrimination (Dion & Pak 1992). Weaker identification with their organization occurs when discriminated employees experience difficulties in feeling a part of their organization (Gioia, Schultz & Corley, 2000). As a result, they are getting further from it (Lindgren & Wеhlin, 2001). The results of this experiment confirmed that expats in discrimination condition adjusted the levels of both identities as a way of coping with discrimination treatment.

The findings related to exclusion are less clear since treatment did not have any effect on both identities. Therefore, contrary to my study's expectation and theoretical support of the RIM model, the levels of organizational and ethnic identity did not change after this treatment: the ethnic identity level did not increase and the organizational identity level did not decrease (Branscombe et al., 1999). This can be explained by several facts. Firstly, participants from the exclusion group could think that exclusion was not based on their ethnicity and therefore, they could change their behavior in order to be included again. Provided they could reestablish their relationship with the out-group, they did not need to seek support from their ethnic group (Brewer, 1991). Secondly, it could be explained by a methodological reason. In particular, due to the short time between the two measures, some participants might have given the same answers twice without relying on the treatment.

Therefore, these results clearly showed that expats who were induced with the thought of discrimination got closer to their ethnic group and further from their organization after this treatment. However, there was no such effect in the exclusion group. Despite these results, I still could not state that discrimination had a stronger effect on organizational identity than exclusion (H1b). One possible reason can be some statistical issues. In particular, the discrimination group consisted less participants than the other 2 groups and this made it difficult to find the desired effect. Another possible reason can be that fact that although exclusion and discrimination are somewhat different concepts (Leets & Sunwolf, 2005; Jackson, Brown, & Kirby, 1998), they have many similarities (Wodak, 2008). Therefore, finding any significant differences between these conditions is complicated since it is difficult to differentiate between the specificity of both treatments. In conclusion, since the effect in the discrimination condition was significant when comparing both identities before and after treatment, it is important to take these results into account.

Implications

This study contributes to the current state of the methodological, theoretical and practical aspects related to discrimination and exclusion.

Regarding methodological contribution, it is the first time when exclusion and discrimination have been investigated on the sample of working expats in Russia. Thus, by using Russian setting I shed some light on contextual aspects of exclusion and discrimination in relation to identities. Moreover, in previous studies the most effective and ethical way to make people feel discriminated or excluded was through the cyberball technique, autobiographical methods, scenarios and etc. (Mok & De Cremer, 2015; Gonsalkorale & Williams, 2007; National Research Council, 2004), but in this study I showed a different way of manipulating both feelings. In other words, this study used a novel treatment which proved to be ethical and effective. Therefore, effectiveness of the treatments was verified for its future use in experiments related to discrimination and exclusion.

Regarding theoretical contribution despite of my forecasts, it was showed that the exclusion conditions did not have significant impact on ethnic and organizational identity. However, it was found that even the mere thought about discrimination in expats' workplace can get them closer to their ethnic group and further away from their organization. According to previous studies, low organizational identity leads to negative consequences (Dutton et al., 1994). Therefore, working expats who think that they are discriminated can get less work-oriented, job-engaged and motivated (Dutton et al., 1994). In addition, thoughts about discrimination can also encourage them to quit or, migrate back to their own country because they would search support from their own ethnic group (Noh, Beiser, Kaspar, & Hou, 1999). This can have a negative impact on an organization especially in case of this research, since a large number of working expats who participated in this study were at high positions in their current organizations and were valued greatly by their employers (Intermark, 2008). Therefore, if organizations lose such valued employees, they will not be able to benefit from their knowledge and experience which, in turn, can affect the whole performance of the company (Bonache & Noethen, 2014).

Due to these reasons, every organization should not only protect its people from experiencing discrimination at their workplace but do their best to keep them away from even thoughts that they can be subject to discrimination. It can be done in different ways. First of all, it is important to care about the organizational culture and values, which preferably should manifest diversity and openness to different mindsets. The policy should also highlight the potential costs of ignoring discrimination in the workplace (Kimberlee, 2019). Secondly, it is worth conducting different trainings which enhance communication between employees and enable them to familiarize themselves with different cultures and teach them how to work effectively as a team (Kimberlee, 2019). This advice can help organizations prevent foreign employees from feeling discriminated. To sum up, this study provides strong empirical support for the need to deal with discrimination itself and reasons for its occurrence as it can lead to negative consequences both for the employee and for the organization.

Limitations

This work had several limitations which should be mentioned. First of all, the current research suffers from unequal distribution of the participants into 3 ...