Third Culture Kids, their diversity beliefs and their intercultural competences

Abstract Positive diversity beliefs are known to increase the effectiveness of diverse groups in organizations and society. Early cross-cultural experiences might facilitate developing these diversity beliefs. This study examined the relationship between being a third-culture kid (TCK) and one’s diversity beliefs, and the possible mediation of this relationship by intercultural competences. Data came from 1454 respondents, 17–19 years of age, who filled out the Intercultural Readiness Check (IRC) between 2011 and 2016 (49.1 % female, 50.9 % male). 65.0 % of the respondents had specified their nationality as Dutch, and 35.0 % had indicated a different nationality. 550 respondents (37.8 %) had spent one or more years abroad and were thus classified as TCKs. Their diversity beliefs were compared to 904 respondents (62.2 %) who had never lived abroad, and who therefore were classified as non-TCKs. A mediated regression analysis showed that TCKs had higher positive diversity beliefs than non-TCKs, and that this relationship was mediated by the degree to which they had developed specific intercultural competences, being intercultural sensitivity and building commitment. These findings show that early cross-cultural life experiences help individuals to develop intercultural competences and positive diversity beliefs. Equipped with these competences and beliefs, TCKs can add value to organizations and society.

A person's experiences with diversity will influence his or her beliefs about diversity . Growing up cross-culturally is a special form of cross-cultural experience and is likely to have an impact on a person's perspectives on cultural differences and beliefs about diversity. This experience relates to perspectives and beliefs regarding culture and ethnicity, and to those about gender, age, and other diversity aspects (Pollock & Van Reken, 1999).
This study investigates the impact of early cross-cultural life experiences on diversity beliefs by focusing on Third Culture Kids (TCKs). TCKs are defined as children who spend a significant part of their first eighteen years of life "accompanying one's parent(s) into a country that is different from at least one parent's passport country(ies) due to a parent's choice of work or advanced training" (Pollock, Van Reken, & Pollock, 2010). Van Reken (1999/2009) explain that a 'significant part' implies one year or more, and for a Third Culture Kid, the period abroad must occur during the developmental years, which range from birth to eighteen years of age (p. 21).
Already at an early age, TCKs might have been able to discover what other cultures have to offer, as they had a chance to get to know and relate to members of other cultures and learn from them (Pollock & Van Reken, 1999). TCKs are among the most-researched types of the broad category of cross-cultural kids (CCKs; Fig. 1). CCKs encompass many subgroups, such as children of refugees and children of immigrants, and collectively refer to people who are "living or have lived inor meaningfully interacted with two or more cultural environments for a significant period of time during childhood" (Pollock & Van Reken, 1999. The characteristic that differentiates TCKs from other CCKs is that TCKs went abroad during their childhood years because of the employment of one or both of their parents. Often the label Third Culture Kids in research refers to children as well as to adults with such cross-cultural childhood experience. TCKs can consequently feel that they belong to all the cultures they have lived in, implying they have developed a so-called third culture of their own (Moore & Barker, 2012;Pollock & Van Reken, 1999). Similarly, Tanu (2015, p. 16) states that studying Third Culture Kids will contribute "to an understanding of the way international mobility impacts upon the individual." To better understand the conditions leading to valuing diversity more, this study investigates whether and how intercultural competences mediate the relationship between early cross-cultural experiences as a Third Culture Kid and diversity beliefs. Growing up as a TCK will bring along a range of cross-cultural experiences, yet at the same time it in itself says little about what the effects of these experiences are on an individual's skills, thoughts and behavior. Given the fact that TCKs do not chose to become TCK, as this is a result their result of their parent(s) moving abroad for their work, their parents' move may result in a variety of cross-cultural experiences, with more or less contact with host cultures. For example, the experiences of military TCKs (Brats; Wertsch, 1991) growing up on a compound will probably imply having significant experiences with people from other cultures, but not necessarily with people from the host culture itself. Regardless of the specific nature of the cross-cultural experiences TCKs may have had, TCKs can be expected to have developed more positive diversity beliefs. The present study investigates to what extent this relationship may be influenced through their developed intercultural competences. Non-TCKs have not lived abroad during their years of development but might have had crosscultural experiences while growing up within their own culture. Yet, in comparison to non-TCKs, during their time abroad TCKs will be more strongly immersed in cross-cultural life. It can therefore be expected that such significant cross-cultural experiences will have more impact on their intercultural skills and through such skills on their diversity beliefs.
Intercultural competences are defined as "the knowledge, motivation, and skills to interact effectively and appropriately with members of different cultures" (Wiseman, 2002, p. 208). Examples of intercultural competences are intercultural sensitivity (Bhawuk & Brislin, 1992;Cleveland, Mangone, & Adams, 1960;Hawes & Kealey, 1981) and tolerance for ambiguity (Brislin, 1981;Deller, 2000;Hammer, Gudykunst, & Wiseman, 1978). Brinkmann and Van Weerdenburg (2014) found that people who had spent more than 2 years abroad, scored higher on the competence of intercultural sensitivity than people with less or no experiences of living abroad, provided they have friends from other cultures. Earlier research indicated that TCKs are interculturally more competent than non-TCKs: TCKs were found, for example, to score higher on measures of intercultural sensitivity, open-mindedness, and respect for and tolerance of others (De Waal & Born, 2020;Hayden, Rancic, & Thompson, 2000;Selmer & Lam, 2003).
We therefore aim to find out whether intercultural competences are critical to understanding the relationship between TCK-statusthat is, being a TCK or notand an individual's diversity beliefs. The next section will discuss the concept of diversity beliefs, and how these may be influenced by early (before the age of 19 years) cross-cultural life experiences. Subsequently, intercultural competences and the role these competences may play in the relationship between TCK-status and diversity beliefs will be elaborated upon (Fig. 2).

Diversity beliefs
Diversity research focuses on a range of differences between individuals, for example on differences in ethnicity, gender, nationality, age, values, educational and professional background, and on how these differences influence how individuals interact and cooperate (Karolidis & Vouzas, 2019;Weber, Sadri, & Gentry, 2018). Empirical studies until now show that the influence of diversity on interacting and cooperating with others critically depends on the extent to which people involved believe that their differences add value to cooperation in a team setting (Homan et al., 2007). Homan et al. (2007), for example, found that team members who believed in the value of diversity for the team were better in exploring and analyzing information critical to performance than were team members who did not see the value of diversity for their team.
These beliefs about diversity have been labeled diversity beliefs, and are defined as "beliefs about the value of diversity to group functioning" (Stegmann, 2011, p. 44). In line with Homan, Greer, Jehn, andKoning (2010), Homan, Buengeler, Eckhoff, Van Ginkel, andVoelpel (2015) and Stegmann (2011), the present study views diversity beliefs as the positive feelings and cognitions that people have about diversity. In the following, we elaborate on several empirical studies which confirm the relevance of diversity beliefs. These studies mainly focus on the functioning of people in groups.
A first study, by Van Dick, Van Knippenberg, Hägele, Guillaume, and Brodbeck (2008), reported that team members who believed in the value of diversity for the team were better in exploring and analyzing information critical to performance than were team members who did not see the value of diversity for their team. Group members with more positive diversity beliefs were also more aware of the diversity within their team and also identified more strongly with their team than group members with less positive diversity beliefs. In a second study showing the relevance of diversity beliefs, Van Oudenhoven-Van der Zee, Paulus, Vos, and Parathasaray (2009) found that respondents with more positive diversity beliefs expected more positive outcomes of their workgroup than did respondents with less positive diversity beliefs. Thirdly, research has shown how positive diversity beliefs also reduce the risk for teams to fall apart into subgroups. To illustrate, Homan et al. (2010) classified 39 teams in terms of the extent to which each faced an objective risk of subgroup formation. This risk was calculated based on the extent to which multiple diversity features were distributed across team members. The results showed that when members of a high-risk team valued their diversity, these members reported seeing no subgroups in their teams. In contrast, when members of high-risk teams did not value their diversity, they did see these subgroups had formed in their team. More recently, Van Knippenberg, Van Ginkel, and Homan (2013) argued that diversity beliefs need to be combined with a diversity-related skills-set to be able to positively influence cooperation within diverse teams. The combination of believing that diversity has value with practical skills to make this belief effective was coined by Van Knippenberg and colleagues as the diversity mindset (Van Knippenberg, Van Ginkel, & Homan, 2013, p. 183). A diversity mindset encompasses a person's diversity beliefs, knowledge about diversity, general person-oriented competences and this person's more specific intercultural competences.

Diversity beliefs and (early) cross-cultural experiences
How do positive diversity beliefs develop, that is, how do people learn to value the diversity between people they encounter? Early cross-cultural experiences seem to play a decisive role. We will review evidence from which the importance of cross-cultural experiences for the formation of positive diversity beliefs may be derived.
In a study involving 196 (adult) TCKs, Melles and Schwartz (2013) found that the degree of exposure to other cultures during their developmental years was related to being less prejudiced towards other people in general. Their results showed that TCKs' level of exposure to and contact with different cultures during their upbringing negatively predicted their levels of reported prejudice. The authors noted that hardly any research had been done on the relationship between being a TCK and level of prejudice: There is "… limited quantitative evidence that the TCK's experience impacts prejudicial attitudes" (Melles & Schwartz, 2013 p. 261). Likewise, to our knowledge no research until now has investigated how TCK-experience impacts positive diversity beliefs.
Specific evidence of the relationship between early cross-cultural experiences and diversity beliefs, although not collected among TCKs, can be found in a study on diversity attitudes of youth in the USA by Weiler, Helfrich, Palermo and Zimmerman (2013). These authors concluded: "Adolescents who were more knowledgeable of other cultures were more likely to understand and accept individuals from other cultures" (Weiler, Helfrich, Palermo, & Zimmerman, 2013, p. 34). Further evidence of the relationship between early cross-cultural experiences and diversity beliefs was found in the constellation of classrooms in the Netherlands. When looking at the influence of ethnic diversity in the classroom on diversity beliefs among adolescents, Van Geel and Vedder (2010) found that one's ethnicity was a significant predictor of multiculturalism attitudes, which includes "the conviction that access to other cultures enriches one's own life" (p. 549). Within a Dutch context, these researchers found that immigrants held stronger multicultural attitudes than did nationals. The immigrant students had parents who were born in Turkey (37), Morocco (36), Surinam (15), and the Netherlands Antilles (15). A total of 60 students had different non-western roots in countries such as Afghanistan, Algeria, and the Dominican Republic. A total of 20 students were western immigrants from Europe and North America. A gender difference was also found: Girls in this study held stronger multicultural attitudes than boys. This study among native and immigrant students demonstrates that positive attitudes towards cultural diversity were stronger among those with a multicultural background.
Although the above studies have not specifically investigated the relationship between the TCK-experience and diversity beliefs, there is some evidence from the specific domain of work that TCKs may contribute specifically to (culturally) diverse work environments, for example through a more international orientation. In a study comparing 63 British TCK adolescents with two non-TCK groups (103 adolescents from Hong Kong and 88 adolescents from the UK), Selmer and Lam (2003) showed that the TCK group was more open than the non-TCK groups to aspects of international mobility, such as pursuing an international career, international traveling, and interest in foreign languages. TCK's international orientation was more strongly related to intercultural competences (e. g., open-mindedness and tolerance) than was the case among the two non-TCK groups. Similarly, Cottrell (2002) found that the length of time US citizens had lived outside the USA as TCKs, the more countries they had lived in, and the more often they had spoken another language than their native tongue at work, the more international their career path had become.
Finally, in a small qualitative study, Sellers (2011) interviewed fifteen adult TCKs about themes which had evolved from their experiences of living outside their passport country. Nine participants thought that, as a result of their time living abroad, they had developed cultural sensitivity and an understanding of diversity (Sellers, 2011, p. 55). Although Sellers (2011) interviewed only a small group of TCKs and did not compare the TCKs self-formulated insights to those of non-TCKs, these findings together with the above studies lead us to expect TCKs cross-cultural experiences to be related to more positive diversity beliefs. Hence, Hypothesis 1 was formulated as follows: Hypothesis 1. TCKs have stronger diversity beliefs than non-TCKs.

Intercultural competences, diversity beliefs and TCKs
There is a rich and on-going discussion about the nature of intercultural competences, and how intercultural competences can best be defined (Engle & Engle, 2004;Miska, Stahl, & Mendenhall, 2014;Ruben & Kealey, 1979;Ruben, 1989;Spitzberg & Changnon, 2009;Thomas & Fitzsimmons, 2008;Van der Poel, 2016). We follow Wiseman (2002), who defines intercultural competences as "the knowledge, motivation, and skills to interact effectively and appropriately with members of different cultures" (p. 208). In relation to the intercultural aspect of interaction among people, Van der Poel (2016) states that "The term intercultural refers to the 'space' where a person from one culture meets a person from a different culture and where both are challenged to navigate the cultural differences by actively seeking (new) meaning" (p. 169).
Intercultural competences differ from diversity beliefs in that they encompass psychological factors supporting intercultural interaction, including knowledge, motivation, behavioral skills and attitudes. Intercultural competences may also include factors like meta-cognition, that is, reflections about ones's intercultural strategies (Egan & Bendick, 2008). Unlike diversity beliefs, intercultural competences refer to an individual's ability to behave effectively and appropriately towards others in an intercultural context (Deardorff, 2011). These behaviors will critically influence whether the interaction with these others overall is experienced as positive, neutral, or negative, and through these experiences they ultimately may affect the degree to which the individual adopts positive diversity beliefs.
One of the well-known models of intercultural competences is the Intercultural Readiness Model of Van Weerdenburg (2003, 2014;Van der Zee & Brinkmann, 2004). This model consists of the following four intercultural competences: Intercultural sensitivity has been defined as the ability to take an active interest in others, their cultural background, needs and perspectives. Intercultural communication is defined as the ability to monitor and adjust one's communicative behaviors when communicating with people from different cultures. Intercultural sensitivity and intercultural communication concern paying extra attention to the needs and perspectives of the other party in intercultural interactions. Building commitment is defined as the ability to influence one's social environment, based on a concern for integrating different perspectives. More specifically, the building commitment competence concerns the skill in intercultural interactions to bring people together around shared goals, by integrating their different perspectives, and even changing one's original goals if necessary. Inclusion of this competence in the Intercultural Readiness Model was inspired by Hampden-Turner and Trompenaars' (2008) work on dilemma reconciliation as a means for resolving tensions between seemingly opposing value orientations. Lastly, managing uncertainty refers to the ability to appreciate uncertainty and complexity of culturally diverse environments as an opportunity for personal development. Egan and Bendick (2008) argued that training individuals in intercultural competence will teach people to "analyze individuals as unique, complex combinations of dimensions of diversity" (p. 391), implying that a higher level of cultural competence leads to more positive diversity beliefs. Confirming their thinking, a survey study by Brinkmann and Van Weerdenburg (2014) indeed showed that intercultural competences have a positive relationship with diversity beliefs, with particularly strong relationships for building commitment and managing uncertainty. The sample was collected in 2011 and 2012 and consisted of 4930 people from all age groups.
While growing up, living and working cross border, one may have developed stronger intercultural competences, and multicultural personality characteristics and other skills. For instance, among a sample of 79 TCK-teenagers Dewaele and Van Oudenhoven (2009) found positive effects of being a TCK and multilingualism on open-mindedness and cultural empathy, but negative effects on emotional stability. TCKs have been found to differ from non-TCKs on many characteristics: TCKs score higher on multi-cultural personality characteristics, speak more languages, have a different sense of identity and belonging, they score higher on well-being than non-TCKs, and are better adjusted to other cultures (Cottrell, 2002, ;Dewaele & Van Oudenhoven, 2009;Hoersting & Jenkins, 2011;Moore & Barker, 2012;Tannenbaum & Tseng, 2015). Moreover, a study by De Waal and Born (2020) found that TCKs have a stronger preference to use a transformational leadership style than non-TCKs, a relationship which was mediated by the multi-cultural personality traits of open-mindedness and flexibility. Related to the topic of adjustment, in an exploratory study comparing 71 self-initiated TCK-expatriates with 196 self-initiated non-TCK expatriates, Selmer and Lauring (2014) found that TCK-expatriates rated themselves higher on cultural adjustment than did non-TCK expatriates. Contrary to the predictions, however, TCK-and non-TCK-expatriates did not differ in self-rated job-and interaction-adjustment. Similarly, the Lyttle, Barker, and Cornwell (2011) compared 74 TCKs to 68 non-TCKs in terms of their ability to assess social situations, and to empathize with others. These authors assessed social and emotional interpersonal sensitivity by both self-reports and performance-based measures. Regardless of the measure with which the competence was measured, TCKs scored higher than non-TCKs on social interpersonal sensitivity, whereas non-TCKs scored higher than TCKs on emotional interpersonal sensitivity.
Integrating the above findings that report that intercultural competences are positively related to diversity beliefs, and findings that TCKs score higher than non-TCKs on intercultural competences, we propose that intercultural competences mediate the relationship between TCK-status and diversity beliefs as follows: Hypothesis 2. Intercultural sensitivity (2a), intercultural communication (2b), building commitment (2c) and managing uncertainty (2d) mediate the relationship between TCK-status and diversity beliefs.

Participants and procedure
The present study drew on data from 1465 respondents who had assessed their intercultural competences using the Intercultural Readiness Check (IRC), the questionnaire developed to assess the above four competences (Brinkmann & Van Weerdenburg, 2014). Respondents were students of higher education, being 17-19 years of age, mostly studying at one of four large Dutch applied universities. They had completed the IRC in the years 2011 to 2016 as part of their course fulfilment at an institute of higher education preparing them for an international part of their studies, for instance for an internship abroad.
Respondents differed in terms of prior experience abroad. Respondents who indicated having never lived abroad were classified as non-TCKs; respondents who indicated having spent one year or more abroad were classified as TCKs. According to the TCK-definition of Van Reken (1999/2009) having lived abroad during a significant part can indicate as little as one year and must occur during one's developmental years, which range from birth to eighteen years of age (p. 21). Eleven participants were excluded from the study because they had selected the same response option for all IRC items, including those that were scored inversely, which made their results invalid. The final sample therefore included N = 1454 participants, n = 550 of which were classified as TCKs, and n = 904 as non-TCKs who served as a comparison group. All participants completed the questionnaire in English. While we do not have information about participants' native language, they chose English over the seven other languages (Chinese, Japanese, Dutch, German, French, Spanish and Brazilian Portuguese) in which the IRC is available. Moreover, a Webfx Flesch Kincaid Grade Level readability test (WebFX Reviews, 2019) indicated that the items can be easily understood by most American 13-year-olds and older. We are therefore confident that participants who filled out the items in English were capable of comprehending these.

Demographics
Participants' gender, nationality, the number of languages spoken (but not fluency) as well as their highest educational level were recorded (see Table 1). The sample was nearly equally divided into men (50.9 %, n = 740) and women (49.1 %, n = 714). A small majority (65.0 %) had the Dutch nationality (n = 945; including those who reported other nationalities besides their Dutch one), and a large majority (97.5 %) spoke two or more languages (n = 1418). A large percentage (69.9 %) indicated having received a degree in higher education (n = 1016).
The distribution of gender was similar across TCKs and non-TCKs, while the percentages of having a Dutch nationality, speaking two or more languages, and highest educational degree attained, differed between TCKs and non-TCKs, as indicated by a significant Chi-square test (p < .05). A larger proportion of non-TCKs compared to TCKs had the Dutch nationality (79.9 % vs. 40.5 %). A somewhat larger proportion of TCKs compared to non-TCKs spoke two or more languages (98.9 % vs. 96.7 %). Slightly more TCKs than non-TCKs had a university degree (19.1 % vs. 10.7 %). More non-TCKs compared to TCKs had a lower educational degree (0.7 % vs. 0.0 %), a degree of middle education (10.2 % vs. 6.2 %) or a degree in higher education (70.0 % vs. 69.6 %). More non-TCKs than TCKs selected other as their highest level of education attained (8.4 % vs. 5.1 %).

TCK-status
A single criterion was used to define the participants' status as a TCK: Participants who had lived abroad for 1 year or longer were classified as TCKs, those who had never lived abroad were classified as non-TCKs.

Diversity beliefs
Diversity beliefs were measured with four statements and scored on a 7-point Likert scale ranging from 1 (Completely disagree) to 7 (Completely agree). The statements (alpha reliability of .89) had been developed by Homan et al. (2010) as a diversity beliefs scale.
Respondents were asked to indicate to what extent they agreed with each of the four following statements: 'Diversity is an asset for Note. a Whether someone spoke two or more languages (≥2 languages) was self-reported, and does not necessarily imply fluency in those languages.
teams'; 'I believe that diversity is good'; 'I enjoy working with different people'; 'I feel enthusiastic about diversity'. The final diversity beliefs score represents the average among the four items. The minimum possible score was 1 (lowest possible diversity beliefs) to 7 (highest possible diversity beliefs). The distribution of scores showed a negative skew caused by a ceiling effect. This indicates that the scale did not differentiate well among people with high scores. The English version of the Intercultural Readiness Check, used in this study, had two short additional scales, the 4-item diversity beliefs scale (Homan et al., 2010) and a 4-item scale assessing the degree to which participants perceived subgroups being present in their team; the latter scale was not used for this study. Respondents were informed, by a separate statement preceding these two scales, that their answers to the scales were intended for research purposes only and would not influence their results on the IRC (Brinkmann & Van Weerdenburg, 2014).

Intercultural competences
Intercultural indicate to what extent they consider each statement to be applicable to them by means of a 5-point Likert Scale ranging from 1 (totally not applicable) to 5 (completely applicable). The score on each intercultural competence is the average response on items of the respective intercultural competence ranging from 1 (lowest possible intercultural competence level) to 5 (highest possible intercultural competence level). The IRC has been in use since 2010 in various settings of intercultural training and coaching (Lyubovnikova, Napiersky, & Vlachopoulos, 2015;Van der Poel, 2016). Further details on this measure, including quality criteria and example items, are provided in Appendix B.

Statistical procedure
Statistical analyses were performed using the Statistical Package for the Social Sciences Version 25 (IBM SPSS 25). We controlled for the variables of gender, Dutch nationality, and level of education. TCKs and non-TCKs differed in the proportion of participants speaking two or more languages. However, we considered it likely that the number of languages spoken is affected by both TCK-status as well as diversity beliefs, rather than the number of languages spoken affecting diversity beliefs. In this situation, the number of languages spoken should not be controlled for, as we would otherwise introduce endogenous selection bias (Elwert & Winship, 2014).
Differences between TCKs and non-TCKs on diversity beliefs were tested using students t-tests and linear regression analysis. We observed a ceiling effect for the diversity beliefs variable, which violated the assumption of homoscedastic errors. We therefore used heteroscedasticity-consistent estimates derived with the Davidson-MacKinnon HC3 technique, as implemented in the SPSS RALM macro (v1.01) developed by Darlington and Hayes (2016). Mediation analysis was performed using the SPSS PROCESS macro (Hayes, 2017). To assess the relationship between TCK-status and diversity beliefs as mediated by the four intercultural competences, we fitted a parallel mediation model using the PROCESS tool model 4, with HC3 heteroscedasticity-consistent standard errors, and bootstrapped confidence intervals with 10.000 bootstrapped samples. The size of the samples resampled with replacement during the bootstrapping procedure was equal to N = 1454.

Descriptive statistics
Table 2 provides the means of the self-reported intercultural competences and diversity beliefs for all participants, and for TCKs and non-TCKs separately. Table 2 also provides the t-test statistics that test the differences between TCKs and non-TCKs in the marginal distribution. TCKs scored higher on three of the four intercultural competences as follows: TCKs scored higher than non-TCKs on intercultural sensitivity, t (1452) = 5.50, p < .001, d = 0.29; on intercultural communication, t (1452) = 5.83, p < .001, d = 0.31; and on building commitment, t (1452) = 3.30, p = .001, d = 0.17. TCKs also scored higher on diversity beliefs, t (1452) = 2.98, p = .003, d = 0.17. Correlations among TCK-status, demographic characteristics, intercultural competences, and diversity beliefs are provided in Table 3. Women scored substantially higher on the scale for diversity beliefs than men, β = .34, t = 6.97, p < .001.

Hypothesis testing
Hypothesis 1 stated that TCKs would score significantly higher than non-TCKs on diversity beliefs. We already reported that TCKs scored higher on diversity beliefs on average (see under results -descriptive statistics). TCKs showed significantly higher diversity beliefs than non-TCKs after controlling for gender, for being of Dutch nationality and for the level of education, β = .16, t = 3.02, p = .003. These findings are in line with hypothesis 1 (Table 4).
Hypothesis 2 stated that the relationship between TCK-status and diversity beliefs would be mediated by the four intercultural competences, namely intercultural sensitivity (hypothesis 2a), intercultural communication (hypothesis 2b), building commitment (hypothesis 2c), and managing uncertainty (hypothesis 2d). Results are depicted in Table 5.
Second, intercultural sensitivity, β = .18, t = 2.52, p = .012, building commitment, β = .57, t = 7.04, p < .001, and managing uncertainty, β = .51, t = 7.72, p < .001, showed a positive association with diversity beliefs when controlling for TCK-status, gender, education and having a Dutch passport, and holding the respective other competences constant. The total effect of TCK-status on diversity beliefs, both directly and indirectly through intercultural competences, was significant, β = .08, bootstrapped 95 % CI [.031; .132], with a partially standardized effect of β SD = .09, bootstrapped 95 % CI [.035; .147]. Note that the partially standardized effect, β SD , is calculated by dividing the given effect β by the standard deviation of the outcome variable, in this case diversity beliefs. It therefore represents the expected average change in the number of standard deviations of the outcome variable that is indicated by a one-unit change of the respective independent variable, when holding all other variables in the model constant. *p ≤ .05; ** p ≤ .01; *** p ≤ .001. a 0 = no, 1 = yes. b Alpha reliabilities in the current sample (N=1454). c ≥2 languages = speaks two or more languages. d Level of education was treated as a continuous variable ranging from lower to university education. Because of this, 104 participants who indicated "Other" as the highest attained level of education were excluded from this analysis. Note. Standard errors and p-values are heteroscedasticity-consistent and were derived with the Davidson-MacKinnon HC3 estimator. a Level of education was treated as a continuous variable ranging from lower to university education. Because of this, 104 participants who indicated "Other" as the highest attained level of education were excluded from this regression analysis. b 0 = no, 1 = yes. With respect to the mediating variables, being a TCK had a significant indirect effect on diversity beliefs through intercultural sensitivity (hypothesis 2a), β = .02, bootstrapped 95 % CI [.002; .035], β SD = .02, bootstrapped 95 % CI [.003; .039] and building commitment (hypothesis 2c), β = .05, bootstrapped 95 % CI [.014; .079], β SD = .0 5, bootstrapped 95 % CI of β SD [.016; .089]. In sum, these two competences mediated the relationship between one's TCK-status and diversity beliefs. That is, the positive effect of being a TCK on diversity beliefs can be partly explained by the positive effect of being a TCK on intercultural sensitivity and building commitment. These findings are therefore in line with hypothesis 2a and c. In contrast, we did not find evidence for hypothesis 2b and d relating to the competency of intercultural communication and managing uncertainty.

Discussion
The purpose of this study was to explore whether one's early cross-cultural experiences would lead to positive diversity beliefs by comparing Third Culture Kids with non-Third Culture Kids. We found that TCKs reported more positive diversity beliefs than non-TCKs. We also found that TCKs showed higher intercultural sensitivity, intercultural communication, and building commitment than non-TCKs, and that intercultural sensitivity and building commitment in turn mediated the relationship between TCK-status and diversity beliefs. During their developmental years, TCKs have encountered a variety of perspectives more often while living abroad than have most non-TCKs, who grew up in one sole culture, implying more positive diversity beliefs.
Intercultural sensitivity as a mediator of the relationship between TCK-status and diversity beliefs confirms earlier findings on the importance of this competence for intercultural effectiveness. Mol, Born, and Van der Molen (2005) found intercultural sensitivity to be a predictor for expatriate performance. Hechanova, Beehr, and Christiansen (2003) found it to predict expatriate adjustment. Intercultural sensitivity refers to having an active interest in others' perspectives, as well as their needs and beliefs. For most TCKs an active interest in others from a different cultural background, can be regarded as a direct result of growing up cross-culturally.
Building commitment carries the strongest part of the mediation effect found. As described earlier, building commitment refers to the ability to influence one's social environment, based on a concern for integrating different cultural perspectives to be able to achieve results together (Brinkmann & Van Weerdenburg, 2014). A possible explanation for this finding is that people who grow up more mono-culturally as non-TCKs, need less effort to integrate different perspectives than growing up as a TCK within a culturally diverse context. Building commitment furthermore implies feeling confident while dealing with cultural differences. A possible explanation of the mediation effect by building commitment may thus also be that TCKs have more confidence in intercultural encounters because they have experienced such encounters in different settings. Said differently, TCKs might thus have developed more confidence to deal with cultural differences compared to non-TCKs as a result of growing up cross-culturally. TCKs at times even are perceived as 'arrogant' by their non-TCK peers because of the ease with which they refer to experiences in other cultures and countries (Pollock, Van Reken & Pollock, 2010). The emphasis on achieving results together lies at the heart of building commitment. TCKs thus believe that working together adds value.
A possible explanation for the finding that intercultural communication does not mediate the relationship between TCK-status and diversity beliefs, can be found in one of the intercultural communication aspect of actively listening and adapting one's communicative style to the other. By itself, however, being able to communicate well with people from other cultural backgrounds does not necessarily increase one's diversity beliefs. For example: speaking another person's language, even non-verbally, does not automatically mean appreciating what that person brings to the encounter. Similarly, Li states that "While learning a second language, one may develop cognitive and linguistic competence. (…) [one's] cultural competence is often underdeveloped" (…) They [Migrants] may speak the shared language fluently, but they do not share the cultural model or the collective value with native speakers of the host society" (2012, p. 108).
The relationship between TCK-status and diversity beliefs appeared not to be mediated through managing uncertainty. Managing uncertainty refers to the ability to appreciate uncertainty and complexity of culturally diverse environments as an opportunity for personal development. A possible explanation for this comes from stress management research. From this domain it is known that growing up with much stress does not necessarily mean that one will become more competent in stress management. On the contrary, one will be less competent in this (Skinner & Zimmer-Gembeck, 2016). A parallel could be the case here with managing uncertainty; having dealt with a lot of uncertainty, may not have made the respondents better in managing uncertainty.

Limitations of this study
We recognize this study has some limitations. First, the diversity beliefs scale of Homan et al. (2010) that we utilized, until now has been used in only a limited number of studies. Second, all data were based on self-reports. Self-perception does not include the perspectives of others about oneself, and feedback about one's self-perception is often missing. Third, the study design was cross-sectional in nature, not allowing for causal inference per se. However, whether someone is a TCK or a non-TCK is a position that is likely influenced by exogenous variables (such as one's parents' job) and therefore it precedes the intercultural competences and diversity beliefs causally.
As a fourth limitation, the choice to label the respondents as Third Culture Kids was based on the fact that all respondents were 17-19 years of age and had lived more than one year abroad. For the reason that, as mentioned before, the broader term CCK captures all groups of people who have had cross-cultural life experiences during their formative years -whether as refugees, migrants, TCKs, members of cultural minority groups, as children of parents with different cultural backgrounds, or as international adoption children (Pollock & Van Reken, 1999) -there might have been a number of CCKs (yet non-TCKs) in our database. We unfortunately did not have further information available about our respondents' background. For instance, intergenerational effects of especially maternal multicultural experiences on their children's tolerance for children from other cultures (Tadmor, Berger, Brenick, Abu-Raiya, & Benatov, 2017) would be interesting to investigate in relationship to Third Culture Kids.
No information, which is a fifth limitation, was available about respondents' native language, thus we could not identify how many respondents would regard English as their first or as their second language. Lastly, intercultural experiences do not necessarily only occur abroad. An example of such experiences within one's own culture would be when people from diverse cultural backgrounds come together in a classroom. Hence, not every student needs to be a TCK to be able to benefit from cross-cultural experiences.

Conclusion
Among 17-19 year olds, TCKs show stronger diversity beliefs than non-TCKs, which relationship is partially mediated by TCKs' intercultural competences, namely being interculturally sensitive and being able to build commitment. We concur with Matthewman (2011) who indicates the growing importance of TCKs for the workforce in the future. The future thus seems bright for TCKs who have their working lives still ahead of them, as they have the essential beliefs and competences that are needed in this globalized world. Alpha reliabilities in the current sample (N=1454). c ≥2 languages = speaks two or more languages. d Level of education was treated as a continuous variable ranging from lower to university education. Because of this, 104 participants who indicated "Other" as the highest indicated level of education were excluded from this analysis.
The magnitude and direction of these correlations indicate that the two measuring instruments, namely the IRC and MPQ, measure constructs which are similar to a certain degree and related to intercultural personality and behavior, which provides support for the construct validity of the IRC. The deviation of these correlations from values close to 1 is to be expected as the IRC measures competences, while the MPQ measures personality traits. The predictive validity of the IRC was tested using data from a large group of respondents (19,946 for IS, BC and MU; 4908 for IC). IRC-scales correlated significantly with an 8-item International Orientation scale (p < .001; with medium to large effect sizes; Brinkmann & Van Weerdenburg, 2014).