International Journal of Intercultural Relations 35 (2011) 17–30
Contents lists available at ScienceDirect
International Journal of Intercultural Relations journal homepage: www.elsevier.com/locate/ijintrel
No place to call home: Cultural homelessness, self-esteem and cross-cultural identities Raquel C. Hoersting, Sharon Rae Jenkins ∗ University of North Texas, USA
a r t i c l e
i n f o
Article history: Received in revised form 26 September 2010 Accepted 30 October 2010 Keywords: Cultural homelessness Cultural identity Cross-cultural moves Self-label Self esteem Third culture kid Sojourners
a b s t r a c t This study examined relations between a cross-cultural geographically mobile childhood and adulthood cultural homelessness, attachment to cross-cultural identities, and self esteem. Cross-cultural identities are loosely defined identities (e.g., third culture kids, military brats, missionary kids) that describe some individuals’ childhood cross-cultural experience. The 475 participants spent at least two years before age 18 in a country different from their parents’ home culture, then returned to the latter. They completed an online survey which included general demographic information regarding cross-cultural experiences in childhood, as well as the Cultural Homelessness Criteria, the Rosenberg Self Esteem Scale, and items that evaluated the strength of affirmation, belonging, and commitment to a self-labeled cross-cultural identity. Cultural homelessness was related to lower self esteem scores; higher affirmation, belonging and commitment to any cross-cultural identity was related to higher self esteem and lower cultural homelessness. Furthermore, such affirmation, belonging, and commitment buffered the cultural homelessness–self esteem association, whereas just having a cross-cultural identity did not. © 2010 Elsevier Ltd. All rights reserved.
1. No place to call home: cultural homelessness, self-esteem and cross-cultural identities Childhood experiences are important for consolidating identity development and culture membership, especially those experiences that involve identifying with cultural, ethnic, and racial groups. During childhood, individuals grow to understand the world around them and learn important social rules and behaviors that are appropriate for survival in their cultural environment (Erikson, 1950/1963; Phinney, 1990; Piaget, 1972; Pollock & Van Reken, 1999; Vygotsky, 1934/1986). However, when there is divergent information about one’s cultural membership, such as from cross-cultural experiences, a multicultural local environment, or being born to a multi-ethnic or multi-racial family, a person might experience difficulty attaining a solid cultural, ethnic, or racial identity (LaFromboise, Coleman, & Gerton, 1993; Vivero & Jenkins, 1999). As current and future trends lead toward globalization, it is more common for children to spend their developmental years exposed to many cultural environments, especially as international moves become more frequent (Bandura, 2002) and further away (Schachter, 2001). The development of faster modes of travel, facility in relocation, international schools, and the openness and willingness of families to live abroad have led societies to become increasingly global. People who experience cross-cultural geographical moves may find that behaviors that were socially appropriate in one context may not be accepted in a new country or culture (LaFromboise et al., 1993). Thus, the motivation to establish group
∗ Corresponding author at: University of North Texas, Psychology Department, 1155 Union Circle #311280, Denton, TX 76203-5017, USA. Tel.: +1 940 565 4107; fax: +1 940 565 4682. E-mail address:
[email protected] (S.R. Jenkins). 0147-1767/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.ijintrel.2010.11.005
18
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
membership in the new country may lead to changes in the individual’s self in an effort to become more integrated into new social groups and establish cultural membership in that new location (Hogg & Mullin, 1999; Sussman, 2000), as can be observed with immigrants, refugees and sojourners (Dana, 2005; LaFromboise et al., 1993; Phinney, 1990; Ryder, Alden, & Paulhus, 2000). Furthermore, a strong group identification helps maintain an individual’s sense of belonging (Lewin, 1948; Tajfel & Turner, 1979). Repeated cross-cultural moves at a young age may have consequences for development of cultural identity (McCaig, 1996; Pollock & Van Reken, 1999; Useem, 1999). However, people who do not belong to typical social categories, well-defined cultural groups, or ethnic enclaves, and who are also exposed to multiple cultural frameworks in their families of origin or by cross-cultural travel at a young age, may experience difficulty finding groups and a culture to which they feel a sense of belonging and being understood (Sussman, 2000), and may develop cultural homelessness (CH; Vivero & Jenkins, 1999). The purpose of this study is to examine how adults who experienced a cross-cultural geographical move during childhood have negotiated their cultural identity, focusing on CH (Vivero & Jenkins, 1999) and its negative association with adult self esteem (Navarrete & Jenkins, 2010), and whether forming a cross-cultural identity (CCI) with others who share an internationally mobile childhood might influence self esteem, buffering the expected association with CH. A corollary purpose is to replicate Navarrete’s (1999) findings relating CH to various parameters of cross-cultural experience in a broad sample of adults who reported widely varied childhood cross-cultural experiences before the age of 14. Understanding how these individuals deal with shifting worldviews and exposure to different cultures in their developmental years, and how these may or may not be internalized as an identity, can be useful for the globalized world. People who are raised cross-culturally may experience difficulty with finding a sense of cultural belonging (Hogg & Mullin, 1999; Sussman, 2000; Vivero & Jenkins, 1999), which may be important for psychological well-being (LaFromboise et al., 1993). 1.1. Acculturation: adapting to belong Changing aspects of one’s self-identity to accommodate new information within an unfamiliar cultural context is one aspect of acculturation, and can sometimes involve alteration in the individual’s sense of self (Ryder et al., 2000). The large literature on acculturation (see Berry, 2005; Chun, Organista & Marin, 2003; LaFromboise et al., 1993 for reviews) has been useful in explaining how individuals exposed to more than one cultural context (i.e., home culture of origin and host culture or cultures) negotiate their cultural identity development. Most of this literature has assumed a single permanent cross-cultural geographic move by an immigrant or refugee population. The present study focused on the children of sojourners, a population who leave their parents’ home culture for an extended stay in a host culture, then return to the home culture. Children of sojourners, like those of immigrants and refugees, share several commonalities. Living part of their developmental years outside their home country, they do not have a choice in geographical moves because of their position in their family. These transitions can make it difficult to establish a cultural identity (Sussman, 2000). However, whereas families of immigrants and refugees generally aim to create a new life in a new country, sojourner families’ aim is to return to the home culture (Useem & Cottrell, 1996). Children of sojourners, like other culture-crossing individuals, may adopt different acculturation strategies. In the early acculturation literature Berry (1990) proposed four such strategies. In different circumstances and contexts, those who come into contact with differing cultural frameworks may value and maintain strong links to both host and home cultures (integration), develop relationships within the host culture at the expense of home culture contact (assimilated), hold strong relationships with their home culture and not with the host culture (separated), or even employ strategies in which identity is based on neither host or home cultures (marginal; see Berry, 1990 for a further review and Rudmin, 2009 for a critique). Successful bicultural identity integration appears to depend on the individual’s perception that the cultural identities involved are compatible and not conflicting (Benet-Martínez & Haritatos, 2005; Haritatos & Benet-Martínez, 2002). When those living in multiple cultural frameworks are able to learn the norms of the two cultures available to them and interact with two distinct cultural frameworks in a successful manner, they are likely to develop bicultural competence (reviewed by LaFromboise et al., 1993), consistent with Rudmin’s (2009) view of acculturation as second-culture acquisition. Bicultural competence results from individuals’ valuing and maintaining strong links with both home and host cultural frameworks, such as those who are integrated according to Berry’s (1990) framework. Biculturally competent individuals demonstrate flexibility at understanding language and communication of each culture (Benet-Martínez, Lee, & Leu, 2006) and can flourish in each cultural setting. When there is a positive attitude toward the two cultures, biculturally competent individuals reap the psychological benefits of having better interpersonal adjustment and socio-cultural adaptation (reviewed by LaFromboise et al., 1993). High bicultural identity integration is related to culturally congruent responding, but those low in bicultural identity integration may respond in culturally incongruent ways (Benet-Martínez, Leu, Lee, & Morris, 2002), depending on the valence of the situation (Cheng, Lee, & Benet-Martínez, 2006). There is evidence of benefits from culture-crossing experiences such as cultural sensitivity and adaptability (LaFromboise et al., 1993; Navarrete, 1999; Pollock & Van Reken, 1999; Useem & Cottrell, 1996), tolerance of diversity (Gerner, Perry, Moselle, & Archbold, 1992; Viser, 1978), feeling at home with others (Greenholtz & Kim, 2009), and a ready understanding of cultural rules (Navarrete, 1999; Pollock & Van Reken, 1999). They are more often bi- or multi-lingual (Navarrete & Jenkins, 2010; Pollock & Van Reken, 1999) and more interested in learning other languages than are those of monocultural upbringing. They adopt other cultural frames of reference readily and report being adept at cross-cultural codeswitching (Navarrete, 1999; Vivero & Jenkins, 1999).
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
19
The benefits of a cross-cultural upbringing may come at the cost of feeling culturally rootless (Pollock & Van Reken, 1999) and suspended between cultures (Useem & Cottrell, 1996). The contradictions of home and hosts’ cultural values that the child may experience growing up may lead to confusion and ambivalence regarding his or her cultural identities in adulthood (Pollock & Van Reken, 1999; Viser, 1978), much like CH individuals. When returning to their home country, they may grieve over relationships and environments that belong to their host country and then experience loneliness and isolation in their home country (Sorti, 1997). About 75% (Useem, Useem, Cottrell, & Jordan, 1993) to 90% (Useem & Cottrell, 1996) of adults who experienced a cross-cultural childhood felt culturally different from individuals without overseas experience. Remarkably, they reported feeling different from others, but did not report feeling isolated or adrift (Useem & Cottrell, 1996) as do marginalized individuals in some acculturation literature (Berry, 1990; LaFromboise et al., 1993) or CH persons (Vivero & Jenkins, 1999). Ryder et al. (2000) proposed that individuals’ multiple cultural identities shift in strength depending on the cultural context to which they are exposed. During a person’s developmental years, when one is consolidating a cultural identity, one may learn several cultural frameworks which, taken together, not only lead to a multicultural identity (Hong, Morris, Chiu, & Benet-Martinez, 2000), but also can produce confusion about which identities to internalize (Vivero & Jenkins, 1999). Sussman’s (2000) sojourner identity model specifically addresses sojourners’ psychological responses to home culture re-entry in four types of cultural identity shifts. Sojourners who begin to feel less at home with their home culture’s values during their host culture stay were labeled as experiencing a subtractive identity shift. Such individuals were likely to report more self-concept ambiguity, self esteem instability and distress when home. Sojourners with an additive identity shift begin to feel more similar to their host cultures and feel that host and home cultures’ values, norms and behaviors are similar. Re-entry was just as distressing for such individuals as for those who adopted a subtractive identity shift. Those sojourners who experienced an affirmative cultural identity shift had low adaptation and little shift in identity during their host culture stay, and generally experienced a positive repatriation. Sojourners with an intercultural identity shift initially have a degree of cultural identity awareness, adapt to their new host environment, and are able to maintain high self-concept. At re-entry they are at ease in their home environment and open to future cultural experiences. They can manage varying cultural frameworks at the same time and shift according to what is contextually appropriate. The latter identity Sussman (2000) has called a Global Identity. Like multicultural identities (Hong et al., 2000), integrated (Berry, 1990), and biculturally competent individuals, they are often able successfully to manage multiple cultural frameworks and reap the benefits of the varying cultural frameworks surrounding them. Thus, while the acculturation literature may be helpful in understanding how individuals cope with cultural encounters, it may not offer a complete understanding for children who experience frequent geographical moves or whose familial goal is to return eventually to their home country. Children of sojourners may employ varied, accumulative, or even different acculturation strategies than those used by children of immigrants and refugees. Furthermore, adopting and maintaining mainly one acculturation strategy may be especially problematic because of the multiple cultural frameworks in which they are exposed, as by definition they generally experience at least two cross-cultural moves, first leaving their parents’ home culture and then returning (although in some cases they are born to families during their overseas assignment, which would involve only one move, that of returning “home”). For example, attaining bicultural competence may be more problematic for children who have spent most of their youth outside their home country because they may have little or no direct experience of their parents’ home culture. Intercultural identity shifts (Sussman, 2000) may not be an option if a child or adolescent does not possess initial cultural awareness, which often is not acquired until later adolescence and early adulthood (Erikson, 1950/1963).
1.2. Cultural homelessness and cultural uncertainty CH is a framework proposed by Vivero and Jenkins (1999) that describes individuals who feel a lack of cultural or ethnic group membership, emotional detachment from any cultural group, and a need for a cultural home. Individuals who experience multiple cultural frameworks before age 14 may be at risk for developing CH (Navarrete, 1999). This includes not only individuals whose family moves geographically during childhood, but also racially or ethnically mixed individuals. CH individuals may have multiple minority status, making their cultural, ethnic or racial identity different even from that of their parents (Navarrete, 1999; Navarrete & Jenkins, 2010; Vivero & Jenkins, 1999) and presenting challenges for racial identity integration (Cheng & Lee, 2009). CH individuals, like those whose acculturation strategy is marginalizing (Berry, 1990) or who have low bicultural identity integration (Benet-Martínez & Haritatos, 2005), live in a framework that may include experiences, feelings, and thoughts that do not belong to any specific cultural reference group (Vivero & Jenkins, 1999). CH is associated with both positive and negative aspects of this uniqueness. The positive experiences associated with CH are positive feelings about differences, being multilingual, and self-reported cross-cultural competence. The negative aspects are those associated with not belonging to any group, feeling alone in one’s differences, and feeling disoriented by the unlabeled experiences attendant on frequent codeswitching (Navarrete, 1999). Not belonging to and not being accepted by any group move them consistently into outgroups, “always a minority wherever they go” (Vivero & Jenkins, 1999, p. 12). Because CH individuals lack an ethnic enclave or a community with which to identify, they lack a cultural home. They may experience a strong yearning to “go home,” but home is no one place (Vivero & Jenkins, 1999).
20
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
Like those individuals who experience subtractive or additive cultural identity shifts (Sussman, 2000), individuals who undergo cross-cultural childhood moves may experience cultural identity shifts that move toward a state in which their cultural identity is not completely defined by either home or host culture. Although CH may be conceptually similar to marginalized groups (Berry, 1990) or some of Sussman’s (2000) cultural identity shifts, it differs in that conceptually it addresses the psychological needs of belonging and an individual’s desire for a cultural home. 1.3. Cross-cultural identities (CCIs) Several related anecdotal literatures have emerged that describe similar phenomena of cultural rootlessness that are distinguished by their self-labels as cross-cultural identities (Greenholtz & Kim, 2009; Pollock & Van Reken, 1999). Useem (1999) identified children of sojourners in India as developing a unique cultural identity that they called “third culture kids” (TCKs) in which children synthesized aspects of their home and host cultures into one that differed from both, an “interstitial culture.” Pollock and Van Reken (1999) described TCKs’ experience of incomplete relatedness to multiple cultures, along with a sense of belonging to a TCK cultural group rather than to (parental) home or host cultures. Sociologists and those working closely with TCK and adult TCK groups have observed the emergence of a cross-cultural subculture in the TCK population (Greenholtz & Kim, 2009; Pollock & Van Reken, 1999; Useem & Cottrell, 1996). Similar terms have been developed to describe children of specific sojourner groups, such as military brats, missionary kids, Norway’s “suitcase children” (Pirinen, 2002), and Japan’s “kaigai/kikoki-shijo” (Podolsky, 2004), global nomads (McCaig, 1996; Thompson, 2009), cross-cultural children, internationally mobile adolescents, and international school students (Gerner & Perry, 2000). We refer to these self-labels as cross-cultural identities (CCIs). Although most published information about cross-culturally mobile children forms a rich anecdotal literature, there are a few very specific studies (Gerner & Perry, 2000; Gerner et al., 1992; Pollock & Van Recken, 1999; Useem et al., 1993), which are mostly related to children of missionaries or military personnel (Jones & Austin, 1987; Schultz, 1986; Stringham, 1993; Viser, 1978). Furthermore, there is little research that focuses on cross-culturally mobile children that is specific to children of sojourners, and the research on TCKs and similar cross-culturally mobile children is not consolidated or unified (Isogai, Hayashi, & Uno, 1999; Szkudlarek, 2010). A central question for this study is whether having a CCI buffers the negative effects of cultural rootlessness and uncertain cultural identity that are associated with CH. Three aspects of CCIs are of theoretical and practical interest. First, the minimal case of a CCI is having an identifiable social category for one’s experience that describes one’s unique frame of reference. Having a label for this experience, such as CH or being a TCK, may help to normalize it by connecting it to the experiences of similarly situated others by a linguistically mediated cognitive process of social categorization (Tajfel, 1982). Second, having a feeling of attachment, belonging, and commitment to the self-label for this category of experiences allows for an emotional substrate for this cognitive process. Third, having a social–interpersonal network of close family members or peers who also have considerable cross-cultural experience would allow for direct face-to-face social support for sharing the otherwise unique aspects of this experience, not only normalizing it but also providing an immediate experience of belonging and a forum for developing a shared understanding and appreciation of both the strengths and strains involved in acquiring such an aspect of social identity. 1.4. Present study and hypotheses The literature on cross-cultural childhood experiences supports the argument that such experiences can lead to feeling suspended between cultures, cultural membership uncertainty, and CH. This can be problematic because it is associated with low self esteem, sensing less perceived control over one’s life (Hogg & Abrams, 1999), and unmet needs for belonging and attachment (Baumeister & Leary, 1995). Because there is so little systematic research in this area, we first replicated Navarrete’s (1999) central findings on a college student sample with our larger and more diverse sample of communitydwelling adults selected for their breadth of early cross-cultural experience. Replicating Navarrete’s findings is important not only for the substantive findings as these bear on the theory behind CH, but also for confirmation of the measurement validity of the CH Criteria Scales (Navarrete & Jenkins, 2010). Furthermore, supportive replication would ensure that CH is not merely a transient artifact of a specific group of 18–25 year old college students with relatively little cross-cultural experience and relatively lower risk for CH. Although adults with cross-cultural childhood experiences may not feel full membership in any specific culture, having a CCI might help to reduce or avert CH by providing a self-label that represents a cultural home shared with others who have similar histories, including similar experiences of confusion and disorientation. Even a tenuous attachment may be helpful against the sense of isolating differentness that characterizes CH. Furthermore, having a sense of affirmation, belonging, and commitment to this label theoretically increases the likelihood that this self-label serves some of the functions of a cultural home, as described by Vivero and Jenkins (1999), further reducing the chances of CH. In addition, having friends and/or family members who have had similar experiences might provide an actual face-to-face community that normalizes the negative aspects of cross-cultural identity strains, enhancing the sense of attachment and belonging and reducing the negative effects of CH. Distinguishing these three aspects of CCIs allows examination of the cognitive, emotional, and social–interpersonal aspects of CCIs separately. Finally, each of these aspects of CCIs may buffer the negative association of CH with self esteem,
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
21
such that those who experience CH but have a CCI, and/or feel attachment, belonging, and commitment to it, and/or have a close social–interpersonal network with members who have cross-cultural childhood histories, may benefit from a higher level of self esteem than those who experience CH and do not have a CCI. Having a strong positive ethnic identity confers feelings of attachment and belonging to a stable and identifiable multigenerational social group with a shared heritage, traditions, and a common frame of reference (Phinney, 1992). The sense of pride and attachment that often follows provides a foundation for self esteem in the family of origin socialization context. Lengthy sojourns outside of that culture are likely to attenuate this sense of attachment and belonging, perhaps lowering self esteem, especially for younger children. In essence, to the extent that CH resembles ethnic homelessness (Vivero & Jenkins, 1999), we are asking whether a CCI might serve some of the functions of a healthy ethnic identity, chiefly that of providing a bit of cultural home. Thus, the first set of hypotheses replicated and expanded on Navarrete (1999) and Navarrete and Jenkins (2010) by testing the following associations: Hypothesis 1(a).
Participants who report higher CH would report lower self esteem relative to those reporting lower CH.
Hypothesis 1(b). Participants who report higher CH would more often report being biracial or multiracial than would those reporting lower CH. Hypothesis 1(c). Participants who report higher CH would report more cross-cultural experience (younger age of first cross-cultural geographic move, more such moves, spending more total time abroad, speaking more languages, holding more citizenships) relative to those reporting lower CH. The second set of hypotheses related CH to the various aspects of CCIs: Hypothesis 2(a). Participants who report higher CH would be less likely to report having a CCI self-label than would those reporting lower CH. Hypothesis 2(b). Participants who report higher CH would report lower levels of affirmation, belonging and commitment toward a CCI relative to those reporting lower CH. Hypothesis 2(c). Participants who report higher CH would report lower levels of cross-cultural experience in their close social–interpersonal network relative to those reporting lower CH. The last hypothesis explored the buffering effect of having a CCI on the association between CH and self esteem: Hypothesis 3. Participants who report higher CH would report lower levels of self esteem unless they (a) have a CCI self-label; (b) report high affirmation, belonging and commitment to that label; and/or (c) report that their close social–interpersonal network is high in cross-cultural experience. 2. Method 2.1. Participants The online survey attracted 520 individuals who provided data that were downloaded into SPSS directly from the online survey program (ZOPE) available at the authors’ institution. The survey was available online between April 2008 and June 2008. Of the 520 adults, 475 met the selection criteria for this study: that as children (younger than age 18) they had spent more than two years on one or more occasions in a country that was not a parent’s home culture and later returned to their parent’s home culture. Excluded were 19 individuals who spent less than two years abroad, 22 whose first moving age was 16 or older (thus not completing two years abroad before the age of 18), and four who had significant missing data (>80%). All moves occurred cross-nationally, with over 100 countries reported that spanned 6 continents (Africa, Asia, Europe, South and North America and Oceania). Because of the family role status of participants during their childhood (as children or adolescents), it can be assumed that they had limited choice about moves. All participants at some point experienced reentry to their home country or close contact with their home country because of their parents and family. Since childhood they may have continued at their cultural home environments, moved to their cultural host country, or settled elsewhere as adults. One benefit of using an online survey is that participants may complete the survey regardless of geographic location. The mean age of the final sample of 475 was 42.6 (SD = 14.7). There were 301 females (63.4%) and 174 males (36.6%). The majority were married (59.9%; n = 280) and had completed an undergraduate or graduate degree (64.4%; n = 307). Most were monoracial (90.1%; n = 428) and of those, most were White or of European descent (90.9%; n = 389). Family occupation, which is a main reason for a family’s cross-cultural relocation, included mostly families from the U.S. military forces (i.e., U.S. Air Force, Army, Marines, Navy; 39.6%; n = 188) and missionary organizations (29.1%; n = 138). See Tables 1 and 2 for a summary of demographic information. The mean age reported for the first intercultural move was 5.6 years (SD = 4.6, n = 474). The mean number of international moves was 2.4 (SD = 1.5). Of the 475 participants, 93.9% (n = 446) provided information concerning the time they spent in a foreign country during their childhood. The mean time spent abroad during childhood was 9.5 years (SD = 5.2). Most participants spoke at least two languages before the age of 18 (76.6%; n = 362) and possessed one citizenship (83.4%; n = 396).
22
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
Table 1 Demographic statistics describing cross-cultural experiences. Demographic variables
M
SD
Range
Age Age of first move Number of international moves Number of languages spoken Number of citizenships Time spent abroad (in years)
42.6 5.6 2.4 2.2 1.2 9.5
14.7 4.6 1.5 1.0 .5 5.2
18–78 0–16 1–6 1–8 1–5 2–18
2.2. Procedure Most of the participant recruitment occurred through a snowball-like design. Emails for participation were sent by the authors and research team to individuals whom they knew to meet, likely meet, or know others who met the criteria for the study; and to email discussion lists for which the description made it seem likely that they would meet the criteria. Recipients were asked to circulate the email to others as they deemed appropriate. Members of organizations that worked directly with international sojourners were contacted and asked to post a blog or publish the survey information in newsletters or websites Table 2 Demographic frequencies. Variables and levels Marital status Married Single Divorced Living with partner Engaged, widowed or separated Education completed Less than secondary Secondary By exam/GED Some University Undergrad degree Masters degree Doctorate Family occupation abroad Missionary Military Government agency Non-Governmental agency Diplomatic Corps International business Other Number of citizenships 1 2 3 5 Number of languages 1 2 3 4 5 6 or more Number of moves 1 2 3 4 5 or more Race (monoracial) White or European descent Asian Black or African descent Native American or indigenous to America Race biracial Race multiracial
N
% of sample
280 134 31 13 17
59.9 28.2 6.5 2.7 3.7
6 26 3 133 163 144 29
1.3 5.5 .6 28.0 34.3 24.0 6.1
138 188 20 14 26 46 43
29.1 39.6 4.2 2.9 5.5 9.7 9.1
396 62 13 1
83.9 13.1 2.8 .2
107 218 95 36 8 5
22.8 46.5 20.3 7.7 1.7 1.0
107 218 95 36 13 428 389 20 7 5 41 4
22.8 45.9 20.3 7.7 2.7 90.1 90.9 4.7 1.6 1.2 8.6 .8
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
23
(such as www.membercare.org, Ecolint alumni Newsletter, www.tckid.com, and Among World blog). The researcher posted announcements on www.facebook.com interest groups, and military and international school alumni organizations were contacted. Individuals and organizations who received the recruitment email voluntarily forwarded the survey’s web address at their discretion. The recruitment email and webposts stated: “Have you lived part of your childhood or adolescence in a foreign country? If you have lived over 2 years before the age of 18 in a country outside the culture of your parents, we invite you to participate in our study about the consequences of living a cross-cultural childhood. You will have the opportunity to enter one of eight $50 gift certificate raffles from Amazon.com.” Potential participants were given a website and password to the site, which invited them to complete an anonymous online survey in English. It was available at the researcher’s institution’s online server, https://web2survey.unt.edu/. The initial page provided for informed consent. It briefly explained the study, overviewed confidentiality, participant rights, and possible risks. It was anticipated that individuals completing the survey would experience minimal risk. Although no participants have contacted the investigator with this concern, they were encouraged to do so if they experienced discomfort as a result from completing the survey. Their consent to participate was required before continuing on to the survey. An opportunity was given to those participants who wished to enter a drawing as an incentive for participation. Participants emailed their contact information to an account that was created specifically for this purpose. Their contact information was kept separate from their responses for confidentiality purposes and the email account deleted after the gifts were awarded. Eight participants were selected at random and sent $50 gift certificates from www.amazon.com, an online retailer. 2.3. Measures The website allowed data collection for both self-rating scales and open-ended questions. Although several other measures were included in the survey, for the purpose of the present study, only measures of demographic information, cross-cultural experience, CCI variables, CH, and self esteem were used. 2.3.1. The General Demographics Questionnaire Self-descriptive information and demographic data included gender, age, marital status, family occupation while abroad, and highest level of education attained, as well as open-ended questions regarding racial, ethnic, and cultural background. They were asked about geographical moves, age of first move, number of international moves, time spent in each location, number of languages spoken, and number of citizenships. 2.3.2. Self Label Identity Measure (SLIM) Items to assess affirmation, belonging, and commitment to a self-labeled CCI were based on the Multigroup Ethnic Identity Measure (Phinney, 1992). This new scale was named Self Label Identity Measure (SLIM). The Multigroup Ethnic Identity Measure was originally developed to assess ethnic identity and consisted of two factors: ethnic identity search and affirmation, belonging, and commitment. Only this second factor was used for the creation of the SLIM. Cronbach’s alphas for the Multigroup Ethnic Identity Measure of affirmation, belonging, and commitment ranged between .81 and .92 in past studies (Ponterotto, Gretchen, Utsey, Stracuzzi, & Saya, 2003). The directions and seven questions from the affirmation, belonging and commitment scale were adapted to allow participants to spontaneously name a self-label or CCI rather than directing them to name an ethnic group as in the original Multigroup Ethnic Identity Measure or prompting them with a list of cross-cultural labels (such as TCK, global nomad, or military brat). The directions for this scale were changed to: “Most people belong to one or more social groups, communities, or networks. There might be several ways to describe people who have had a variety of cross-cultural experiences. Is there a particular label that you feel best describes a group that encompasses your childhood cross-cultural experience? If so, what is that?” The response to this open-ended question was used as the indicator of having a CCI label (or not). Only participants who answered “yes” to having a CCI label answered the seven SLIM items. Those who answered “no” were gated past these items. A sample item is, “I feel a strong attachment towards the group I named above.” Items were rated on a four-point Likert-type scale ranging from 1 (strongly disagree) to 4 (strongly agree). A Cronbach’s alpha of .90 was found for the seven items in this study. Interitem correlations ranged from .38 to .75. The mean of all seven items was used to reflect the affirmation, belonging and commitment to the cross-cultural self-label. 2.3.3. Cross-cultural experience network score The cross-cultural experience network score was created to evaluate whether participants’ closest family and peers (close social networks) had undergone similar cross-cultural experiences as themselves. The cross-cultural experience network score was derived from the question, “List initials and relationship (i.e., mother, brother, friend, spouse) of those who you feel are closest to you” and “Now go back and indicate on a scale from 1 to 5, what level of cross-cultural/multicultural
24
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
Table 3 Descriptive statistics and reliability for continuous measures. Measure
M
SD
Range
Self Label Identity Measure (SLIM) Cross-cultural experience score Cultural homelessness (CH) Criterion I Cultural homelessness Criterion II Cultural homelessness Criterion IIIa Cultural homelessness mean of 3criteriab Rosenberg Self-esteem Scale
3.5 3.8 2.6 2.6 3.6 2.9 4.4
.53 1.0 .93 .99 1.1 .71 .65
1–4 1–5 1–5 1–5 1–5 1–4.7 1–5
a b
˛ .90 – .83 .85 – – .89
Inter-item r .38–75 – .27–75 .32–65 – – .24–74
CH Criterion III is comprised of 1 item. Reliability for CH is calculated separately for each CH criterion.
experiences you perceive they have.” Participants were asked to rate three individuals with whom they felt close. Crosscultural experience scores were 1 (limited), 2 (somewhat limited), 3 (lived for a brief time in another country), 4 (lived for an extended period of time in another country during adulthood) and 5 (lived for an extended period of time in another country during their childhood or adolescence). The average of these three ratings was taken as a summary score. Because this is an index, not a scale, internal consistency reliability is not appropriate (Streiner, 2003). 2.3.4. Cultural Homelessness Criteria (CH; Navarrete & Jenkins, 2010) This 14-item theory-based self-report index measures three specific domains of cultural belonging: struggles to determine ethnic group membership, lack of emotional attachment to any particular cultural group, and feelings of not belonging to any group (Navarrete, 1999). The three criteria are Lack of Ethnic/Cultural Group Membership and Attachment, Lack of a Cultural Home, and Need for a Cultural Home. Sample items for Criterion I, Lack of Ethnic/Cultural Group Membership and Attachment, included “When I think which ethnic or cultural group I mostly act or think like, I cannot find one,” and “No one ethnic group label accurately describes me.” Sample items for Criterion II, Lack of a Cultural Home, are “I struggle to determine where I belong ethnically or culturally,” and “I don’t feel culturally ‘at home’ anywhere I go.” The third criterion, Need for a Cultural Home, was comprised of one item, “Finding a cultural home is important to me.” Individuals were considered CH when each of the three criteria was met to a moderate degree (higher than a mean score of 3 on a 1–5 Likert-type scale). Ratings ranged from 1 (strongly disagree) to 5 (strongly agree). Navarrete and Jenkins (2010) reported Cronbach’s alphas of .71 and .84 for Criteria I and II, respectively, and overall interitem correlations of .33–.78. Inter-criterion correlations ranged from .33 to .46 (see Table 3). Reliability analysis for this study was conducted on the three criteria separately, as appropriate for such indices (Streiner, 2003). Criterion I, Lack of Ethnic/Cultural Group Membership and Attachment, had a Cronbach’s alpha of .83, with interitem correlations ranging from .27 to .75 (see Table 3). Criterion II, Lack of a Cultural Home, had a Cronbach’s alpha of .85, with interitem correlations ranging from .32 to .65. Criterion I and Criterion II were strongly correlated (r = .77, p < .001). The single-item Criterion III, Need for a Cultural Home, was correlated with Criterion I and Criterion II at −.07 (p = ns) and .10 (p < .05), respectively. As CH criteria are conceptualized as an index, not a scale, high inter-criterion correlations are not necessarily expected nor required (Streiner, 2003). Each criterion is central to the theory. Following Navarrete and Jenkins’ (2010) scoring of CH, the criteria were combined in two ways for analysis, one yielding a presence–absence categorization and one giving a continuous mean score. The categorical approach defined presence of CH as a score of 3 or above on all three criteria. The continuous CH score was calculated by taking the mean of the scores of all three criteria so that less variance is lost than when using a categorical variable. The point biserial correlation between the categorical CH status and the continuous CH score was .68 (p < .001). 2.3.5. Rosenberg Self-esteem Scale (Rosenberg, 1989) The Rosenberg Self-esteem Scale is a widely used 10-item self-report scale that measures participants’ feelings regarding their self-worth. Five positively worded and five negatively worded items are rated on a Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree), and the scores for the latter items are reversed before averaging them. In past studies, test–retest reliability for this scale has been found to be between .82 and .88 and the internal consistency reliability was Cronbach’s alpha of .77–.99 (Blascovich & Tomaka, 1991). For the present study it was .89 (Table 3). 2.4. Descriptive analysis The variable distributions were examined for outliers and normality. Except for the SLIM and self esteem, which revealed ceiling effects, there were no remarkable patterns of responses. Concerning CCI self-labels, 71.4% (n = 339) reported that they had a phrase or label that described their childhood cross-cultural experiences and responded to the SLIM, with a mean score of 3.5 (SD = .53; see Table 3). The distribution of means was negatively skewed (skewness = −1.4) and leptokurtic (kurtosis = 2.0), revealing a ceiling effect; about 50% of participants scored between 3.7 and 4. The sample mean for the cross-cultural experience network score was 3.8 (SD = 1.0; see Table 3). About 50% of cases fell between 4 and 5.
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
25
Table 4 Correlations among variables. Variable
1
2
3
Age 2. Gendera , b 3. Highest level of schooling 4. Number of races 5. Time spent abroad 6. Number of moves 7. Number of citizenships 8. Number of languages 9. Cultural Homelessness 10. Having a CCI self-labela , c 11. SLIM 12. Cross-cultural experience score 13. Self-esteem
– −.06 .26*** −.10* −.35*** −.01 −.23*** −.13** −.38*** .04 .15** −.10* .28***
– −.08 −.03 .13** .06 .08 .05 .01 .00 .05 .05 −.11*
– −.11** −.05 .00 −.02 .07 .00 .04 −.14* −.10* .12**
a b c d * ** ***
4
5
6
7
8
9
10
– .15** .24*** −.01 −.03 .11* −.08 .02
– .06 .18*** −.10 .02 .00 −.02
– .22*** −.03 −.02 −.01 .01
– −.05 −.20*** .09 −.49***
– –d −.02 .00
11
12
– .10* .09* .05 −.04 .15* −.05 −.02 .01 −.12*
– .40*** .20*** .20*** .19*** .03 .02 .04 −.11*
– .13** .22***
– −.02
Correlations are point biserial. Coded 1 = Male, 2 = Female. Coded 0 = no, 1 = yes. Only participants with a CCI answered SLIM items. p < .05. p < .01. p < .001.
For the categorical scoring of the CH Criteria, 22.3% (n = 106) of participants were identified as CH because they scored 3 or higher on all three CH criteria, as compared to 7.9% for Navarrete and Jenkins (2010). The mean score for the CH Criterion I, Lack of Ethnic/Cultural Group Membership and Attachment, was 2.6 (SD = .93), for Criterion II, Lack of a Cultural Home, was 2.6 (SD = .99), and the mean for Criterion III, Need for a Cultural Home, was 3.6 (SD = 1.1). The CH mean score was 2.9 (SD = .71) compared to Navarrete and Jenkins’s 1.41, about a half-point higher than for Navarrete and Jenkins (2010) given that they used a 0–4 scale. The mean response on the Rosenberg Self-esteem Scale was 4.4 (SD = .65). The distribution was negatively skewed (skewness = −1.3) and leptokurtic (kurtosis = 1.5), revealing a ceiling effect. The majority of the scores fell at a higher range of this scale and showed little variability among high scores, indicating that correlations with this variable reflect primarily lower self esteem, with little practical distinction between those with moderate and high levels. Bivariate correlations were run to identify relationships among variables and any demographic confounding variables. Significant relationships were found among demographics, cross-cultural experiences, and measures of interest (see Table 4). Age was significantly related to all variables except gender, number of moves and having a CCI, and was thus controlled in relevant hypothesis tests. 2.5. Analysis of data A power analysis indicated that for this study, small effect sizes (r = .10; Cohen, 1992) could be detected at p = .05 with high power levels (power = .80). Because of the associations among age, SLIM, CH, and self esteem scores, univariate Analysis of Covariance (ANCOVA) with age controlled, rather than t-test, was the correct test for group differences in Hypothesis 1(a)Hypotheses 1(a)-(c) and 2(b) for the categorical variable CH status. Similarly, partial correlations controlling for age, rather than bivariate correlations, tested Hypothesis 1(a)Hypotheses 1(a)-(c) and 2(b) for the continuous CH mean. To test Hypothesis 3, Baron and Kenny’s (1986) approach to evaluating moderator variables was used to evaluate three possible moderating variables affecting the association between CH (independent variable) and self esteem (dependent variable). These three represented the CCI construct in cognitive, emotional, and social forms. Having a CCI or not (cognitive) was represented by a dichotomous, categorical variable, and SLIM (emotional) and cross-cultural experience network scores (social) were both continuous moderating variables. Therefore, two different approaches were taken to test for moderation depending on whether the moderator was categorical or continuous, following Baron and Kenny’s (1986) procedures. Baron and Kenny (1986) emphasized that before running a moderator analysis, the independent and dependent variables should be uncorrelated with the moderating variable so as to provide support for the moderation effect. Having a CCI self-label and cross-cultural experience network scores were not correlated with either CH or self esteem. Significant correlations were found between SLIM and self esteem scores (r[473] = .22, p < .001) and SLIM scores and CH scores (r[337] = −.20, p < .001). Because age was significantly correlated with SLIM, CH, and self esteem (r[337] = .15, p = .006; r[471] = −.38, p < .001; r[471] = .28, p < .001, respectively), it might be a spurious cause of these relationships. Thus, age was entered into the regression equation as a first step in all three moderator analyses. For the categorical moderator strategy, two separate multiple regressions were conducted; one for participants having a CCI self-label and one for those without. The difference in the standardized regression coefficients was tested using a Fisher’s z transformation and test. For the continuous moderator strategy, to create the multiplicative interaction term, SLIM scores, cross-cultural experience network scores, and CH scores were standardized and recentered on 10. The interaction variable
26
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
Table 5 Summary of Hypothesis 1(a)Hypotheses 1(a)-2(c). Variable
CH M
Self-esteem Number of races Age of first move Number of international moves Time spent abroad (in years) Number of languages spoken Number of citizenships CCI labelb , c SLIMe Cross-cultural experience score a b c d e * ** ***
3.8 1.2 4.6 2.5 10.8 2.5 1.3 – 3.4 3.9
Non-CH
F
SD
n
M
SD
n
.76 .42 4.4 1.6 5.3 1.1 .52 – .55 .84
105 103 105 105 101 103 103 106 78 104
4.6 1.1 5.8 2.4 9.1 2.2 1.7 – 3.6 3.8
.51 .32 4.7 1.5 5.1 1.0 .47 – .51 1.1
368 368 367 368 348 364 367 369 259 362
79.3*** 3.2* 11.7*** .05 30.9*** 5.6** 13.3*** –d 6.8*** 1.6
2
ra
.25 .01 .05 – .12 .02 .05 – .05 –
−.43*** .12** −.04 −.01 .07 .19*** .10* .05 −.15** .09
Partial correlations with age controlled. Correlations are point biserial. Coded 1 = yes, 0 = no. Ф = .04, p = ns. Only for those having a CCI self-label. p < .05. p < .01. p < .001.
(CH × moderator) was calculated by multiplying the standardized and recentered independent variable (CH) by the similarly transformed moderating variable. These were then entered into the regression analysis as detailed below. 3. Results 3.1. Hypothesis testing: replication hypotheses
Hypothesis 1(a).
Participants who report higher CH would report lower self esteem relative to those reporting lower CH.
Results from an ANCOVA with age as a covariate revealed that those identified as CH (n = 105) scored statistically significantly lower (F [1,471] = 79.3, p < .001) on self esteem (M = 3.8, SD = .76) than did those who belonged in the non-CH group (M = 4.6, SD = .51, n = 368) at a large effect size (partial 2 = .25). For the continuous form, the partial correlation with age controlled was negative and statistically significant, r(466) = −.43, p < .001. This hypothesis was supported with a medium to large effect size. These findings support the hypothesis that CH is associated with low self esteem (see Table 5). Hypothesis 1(b). Participants who report higher CH would more often report being biracial or multiracial than would those reporting lower CH. Results from an ANCOVA with age as a covariate revealed that those identified as CH status (n = 103) significantly more often reported being biracial or multiracial (F[1,469] = 3.2, p < .05; M = 1.2, SD = .42) than did those who belonged in the nonCH group (M = 1.089, SD = .32, n = 368) at a small effect size (partial 2 = .01). For the continuous form, the partial correlation with age controlled was positive, small, and statistically significant, r(466) = .12, p < .01. These findings support the hypothesis that CH is associated with biracial and multiracial status (see Table 5). Hypothesis 1(c). Participants who report higher CH would report more cross-cultural experience (younger age of first cross-cultural geographic move, more such moves, spending more total time abroad, speaking more languages, holding more citizenships) relative to those reporting lower CH. Results from five ANCOVAS with age as a covariate revealed that those identified as CH had their first geographic crosscultural move at a younger age (F[1,470] = 11.7, p < .001; 2 = .05), spent more time abroad (F[1,447] = 30.9, p < .001; 2 = .12, a moderate effect size), spoke more languages (F[1,465] = 5.6, p = .004; 2 = .02) and held more citizenships (F[1,468] = 13.3, p < .001; 2 = .05), all small effect sizes except as noted, compared with those who belonged in the non-CH group. There was no statistically significant difference in the number of geographic cross-cultural moves between those who belonged to the CH and the non-CH group. When using the continuous form of CH, the partial correlation with age controlled revealed that those who report higher CH also report speaking more languages (r[462] = .19, p < .001) and possessing more citizenships (r[465] = .10, p = .02) at small effect sizes, but not moving at younger ages, moving more often, or spending more time abroad (see Table 5). These findings support the hypothesis that participants who report more intense CH differentially report some forms of cross-cultural experience, especially speaking more languages and possessing more citizenships, but not others.
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
27
Table 6 Summary of hierarchical regression analysis for SLIM scores as a moderator of the CH–self-esteem association (N = 475). Variable Step 1 Age Step 2 Age Cultural homelessness (CH) Self Label Identity Measure (SLIM) Step 3 Age Cultural homelessness (CH) Self Label Identity Measure (SLIM) SLIM × CH interaction * ** ***
Zero-order r
SE × B
B
ˇ
.33***
.22
.03
.33***
.33*** .54*** .37***
.08 −.29 .07
.03 .04 .03
.12* −.44*** .11*
.33*** .54*** .37*** .55***
.09 −1.25 −.89 .10
.03 .31 .32 .03
.14* −1.88*** −1.33** 1.84**
R2
R2
.11
.11
.29
.18
.31
.02
p < .05. p < .01. p < .001.
3.2. Hypothesis testing: CCI hypotheses
Hypothesis 2(a). Participants who report higher CH would be less likely to report having a CCI self-label than would those reporting lower CH. A correlation (Ф, phi) between these binary-coded variables revealed that those who reported higher CH status were just as likely to have a CCI as were those who reported lower CH. For the continuous CH form, the point biserial correlation was not statistically significant (see Table 5). These findings do not support the hypothesis that participants who report higher CH are less likely to report a CCI self-label than those reporting lower CH. Hypothesis 2(b). Participants who report higher CH would report lower levels of affirmation, belonging and commitment toward a CCI relative to those reporting lower CH. Results from an ANCOVA with age as a covariate revealed that among those who gave a CCI, those identified as CH status reported lower levels of affirmation, belonging and commitment toward that label, F[1,334] = 6.8, p = .001; M = 3.4, SD = .55, than did those in the non-CH group (M = 3.6, SD = .51), with a small effect size (2 = .05). A partial correlation with age controlled showed a statistically significant negative relationship supporting this hypothesis, r(332) = −.15, p = .001, with a small effect size (see Table 5). These findings support the hypothesis that CH is related to less CCI affirmation, belonging and commitment. Hypothesis 2(c). Participants who report higher CH would report lower levels of cross-cultural experience in their close social–interpersonal network relative to those reporting lower CH. Results from a one-way ANOVA revealed no significant difference between participants who reported higher CH (M = 3.9, SD = .84, n = 104) and those with lower CH (M = 3.8, SD = 1.0, n = 362). For the continuous CH calculation, a Pearson’s correlation did not reveal any significant relationship. These findings did not support the hypothesis that CH is associated with participants’ cross-cultural experience in their close social–interpersonal network. 3.3. Moderation analysis
Hypothesis 3. Having a CCI (operationally defined in three ways using three different moderator variables: (a) having a CCI self-label; (b) affirmation, belonging and commitment to that label [SLIM]; and (c) close social–interpersonal network’s crosscultural experience [cross-cultural experience network score]) would have a buffering effect on the association between CH and self esteem. Of the three moderator variables, only SLIM scores moderated the relationship between CH and self esteem. Having a CCI self-label or higher cross-cultural experience network scores did not moderate the relationship between CH and self esteem. A hierarchical multiple regression analysis was used to evaluate the moderating effects of the continuous moderating variables (SLIM and cross-cultural experience network scores) with self esteem as the dependent variable. Age was entered at the first step, then on the second step, both the independent variable (CH) and the moderating variable (either SLIM or cross-cultural experience network score) were entered, and, finally, the multiplicative interaction variable on the third step. For the SLIM, the standardized regression coefficient for the interaction variable (SLIM × CH) produced a significant 2% increase in variance (see Table 6) supporting this hypothesis. The higher participants scored on the SLIM, the higher their level of self esteem, even when they were also high in CH.
28
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
4. Discussion The purpose of this study was to look at how adults who had a cross-cultural geographically mobile childhood and adolescence experienced their cultural identities in adulthood and how this might be related to their self esteem. This study took three perspectives on assessing cross-cultural identity, addressing cognitive, emotional, and social components. Five out of seven hypotheses were supported in this sample of community-dwelling adults, including those that replicated Navarrete’s (1999) findings with college students (Hypothesis 1(a)Hypotheses 1(a)-(c)). Those indicated that high CH individuals had lower self esteem, were more likely to be bi- or multiracial, had longer cross-cultural experiences at younger ages, spoke more languages, and held more citizenships. Of the hypotheses generated for this study, low CH persons were more likely to express emotional attachment to a CCI self-label (SLIM) if they had one, compared to those who scored higher (Hypothesis 2(b)), but they were not more likely to have a cognitive component (a cross-cultural self-label), or a social one (more cross-culturally experienced close social network). Strength of affirmation, belonging and commitment to a CCI (as measured by SLIM scores) buffered the negative association between CH and self esteem (Hypothesis 3(b)), but just having a CCI self-label or high cross-cultural experience network scores did not. Thus, it seems that emotional attachment is more central than the cognitive or social component to the role of CCIs in psychological well-being as represented by self esteem. Age was controlled for most hypothesis testing as it was found to be a confounding variable. These findings support Vivero and Jenkins’s (1999) theory and Navarrete’s (1999) findings that adults who led a crossculturally mobile childhood and adolescence are at risk for CH, which may not only provide adaptive cognitive and social benefits but also impact an individual’s emotional security and self esteem. About one-fourth of this sojourner sample met criteria for CH (based on Navarrete’s scale criteria) compared to Navarrete’s 7.9% in a college student sample. Although this was an expected finding given the selection criteria for participants’ cross-cultural experience, it helps to explain the rise and popularity of an internet-based culture of websites and discussion groups related to CCIs. In addition, meeting criteria (M > 3) for all three CH criteria placed individuals at even higher risks for lower self esteem than meeting criteria for only one or two. This study aimed to find possible CCI moderators of this relationship. Two important findings support the theory that strong group identification helps maintain an individual’s sense of belonging (Lewin, 1948; Tajfel & Turner, 1979) and that membership in social groups goes beyond self-labeling (Cox & Gallois, 1996). Strength of affirmation, belonging, and commitment to a CCI, the emotional component of group membership, was related to both CH and self esteem, and also acted as a moderator of the association between CH and self esteem, neither of which was true for the purely cognitive or social components. Thus, those who experienced CH but expressed greater commitment to any cross-cultural label or identity tended to have higher self esteem than those who did not exhibit such a commitment but were CH. One particular benefit of using the SLIM was that it allowed participants to spontaneously label a CCI using their own words. Then for those given labels, a scale of affirmation, belonging and commitment based on Phinney’s (1992) Multigroup Ethnic Identity Measure was created. Having a continuum of commitment to any CCI was more informative than the dichotomous categorization of having a label or not. 4.1. Limitations Because of the recruitment method, this convenience sample was composed of people who were comfortable with the internet and were somewhat connected to an international community, as well as having experienced a cross-cultural mobile childhood. This might explain their relatively high educational level as noted also by Useem (1999). Thus, the results may not generalize to people who experience cross-cultural childhood mobility but are not connected to an international community. Such individuals are likely to be even more vulnerable to CH due to their distance from a community of common experience to which they might belong. The fact of their omission may have restricted the range of the social network’s cross-cultural experience score, limiting the possibility of support for Hypotheses 2(c) and 3(c). If a future study is better able to include such persons, these hypotheses should be tested again. The sample had a large percentage (82%) of racially White individuals (10% multi-racial, 7% non-white). This may be due to the sampling procedure. Future studies should include a more culturally and racially diverse sample, perhaps by outreach to countries where international businesses are emerging. The survey was written in English, requiring English reading fluency. Traditionally, the USA and Europe have acted as hosts for organizations that send employees abroad. Most organizations and individuals contacted to distribute the survey were based in the USA (such as the U.S. Military) or Europe (such as MemberCare). However, as other countries are quickly developing larger international business economies (such as India, Brazil, Russia, and China; O’Neill, 2001), this demographic should change. One important consideration in this study is that the design was cross-sectional, not longitudinal. Although a strong relationship was found between CH and self esteem and a moderate relationship between CH and childhood cross-cultural experience variables, causal relationships among them cannot be assumed. Cross-cultural experiences are certainly not the only influence on the association of CH and self esteem, as CH is surely not the only form of cultural alienation or correlate of low self esteem. Statistical considerations are also a limitation. First, results of the self esteem measure and the SLIM revealed ceiling effects, as is typical in nonpatient samples. When individuals in one group were found to have lower self esteem than other groups, these “low” groups still had a mean self esteem score in the moderate to high range. Second, the cross-cultural
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
29
experience network score was a new index developed for this study, and in hindsight, including more than the person’s three closest relationships may have yielded different results. This measure did not specify the amount of cross-cultural experience had by close others, nor did it specify the age range used to screen participants, all of which may have yielded a less precise measure. Which parameter is most important remains unclear; perhaps the number rather than closeness of cross-culturally experienced network members has greater power, or the sense of community that develops among groups. Age was a significant confounding variable in all analyses. Although this could be an artifact of sampling, it also makes developmental sense. Younger individuals may still be consolidating a personal and cultural identity, independent of their childhood experiences, which might explain the relationship between age, CH and self esteem. Older participants may have succeeded in consolidating an identity of sufficient complexity to encompass their childhood cross-cultural experiences, or may have otherwise come to terms with CH. 4.2. Future research Given the above limitations of the cross-cultural experience network score, Hypothesis 2(c) merits testing again with a more refined strategy. One approach would reverse the structure of the cross-cultural experience network score by first asking whether the person knows anyone who had cross-cultural experiences before age 18, then ascertaining the relationships between the participants and such persons and finally scoring their subjective closeness. An expanded item set could include items designed to assess participation in internet communities. Future researchers may also wish to explore the cross-cultural variables such as CH by asking participants to recall their experiences retrospectively. For example, rather than ask participants how they think or feel in the present as adults about their cross-cultural experiences, they may wish to ask participants to recall how they thought and felt as children. Future researchers may wish to study the impact of cross-cultural moves as these impact specific developmental years. One question of practical value might be to study whether there are any specific ages in which a cross-cultural move would be psychologically beneficial or detrimental. They may also wish to explore how long a child needs to live in their host or home country in order to develop a sense of cultural identity and knowledge. In the USA and Europe, educators working with established re-entry programs, such as those that try to reduce reverse culture-shock and facilitate sojourners’ return to their home culture, may wish to incorporate into their curriculum strategies for individuals to explore their cultural identity in their own words, rather than impose labels or definitions. Although a comparison group might make this study’s design stronger, the hypotheses replicating Navarrete (1999) allow for comparison with a college student sample many of whom had never traveled abroad. Thus, participants varied widely in their exposure to cross-cultural experiences and within both samples, that variation in degree matters. However, in the future, researchers may want to make a comparison study in which both groups’ data are collected at a closer timeframe and with similar methodologies (internet vs. paper and pencil). As social networking sites reach farther into the global world, researchers might explore the impact of these on cultural identity formation. As the use of technology and social networking increases, the way that culture is learned might change. However, when it comes to interpersonal communication, electronic channels primarily convey declarative knowledge (information about) rather than procedural knowledge (understanding of how to do). The latter is a matter of daily interactive performance with face-to-face reinforcements and punishments that have an immediate emotional impact, connecting individuals to their cultures emotionally. Although having a CCI may be helpful in understanding oneself, it is not the label or definition itself that is involved in increased self esteem but the feeling of group belonging and commitment to the self-label of the individual’s choosing. Acknowledgments This study was supported by a student research grant award from the American Psychological Association Division 19, Military Psychology. The authors acknowledge Jennifer L. Wilson for her contribution to the development and data collection of this project and Patrick McLeod and Robert Peacock for their technical assistance. The authors wish to thank the Editor and anonymous reviewers for their suggestions for future research. References Bandura, A. (2002). Social cognitive theory in cultural context. Applied Psychology: An International Review, 51, 269–290. Baron, R. M., & Kenny, D. A. (1986). The moderator–mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173–1182. Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497–529. Benet-Martínez, V., & Haritatos, J. (2005). Bicultural Identity Integration (BII): Components and psychosocial antecedents. Journal of Personality, 73(4), 1015–1050. Benet-Martínez, V., Lee, F., & Leu, J. (2006). Biculturalism and cognitive complexity: Expertise in cultural representations. Journal of Cross-Cultural Psychology, 37(4), 386–407. Benet-Martínez, V., Leu, J., Lee, F., & Morris, M. (2002). Negotiating biculturalism: Cultural frame switching in biculturals with oppositional versus compatible cultural identities. Journal of Cross-Cultural Psychology, 33(5), 492–516. Berry, J. W. (1990). Psychology of acculturation: Understanding individuals moving between cultures. In R. W. Brislin (Ed.), Applied cross-cultural psychology (pp. 232–253). Thousand Oaks, CA: Sage.
30
R.C. Hoersting, S.R. Jenkins / International Journal of Intercultural Relations 35 (2011) 17–30
Berry, J. W. (2005). Acculturation: Living successfully in two cultures. International Journal of Intercultural Relations, 29, 697–712. Blascovich, J., & Tomaka, J. (1991). Measures of self esteem. In L. S. Wrightsman (Ed.), Measures of personality and social psychology attitudes (pp. 115–160). San Diego, CA: Academic Press. Cheng, C., & Lee, F. (2009). Multiracial identity integration: Perceptions of conflict and distance among multiracial individuals. Journal of Social Issues, 65(1), 51–68. Cheng, C., Lee, F., & Benet-Martínez, V. (2006). Assimilation and contrast effects in cultural frame switching: Bicultural Identity Integration and valence of cultural cues. Journal of Cross-Cultural Psychology, 37(6), 742–760. Chun, K. M., Organista, P. B., & Marin, G. (2003). Acculturation: Advances in theory, measurement and applied research. Washington, DC: American Psychological Association. Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155–159. Cox, S., & Gallois, C. (1996). Gay and lesbian identity development: A social identity perspective. Journal of Homosexuality, 30, 1–30. Dana, R. H. (2005). Multicultural assessment: Principles, applications, and examples. Mahwah, NJ: Erlbaum. Erikson, E. H. (1950/1963). Childhood and society. New York: Norton. Gerner, M. E., Perry, F., Moselle, M. A., & Archbold, M. (1992). Characteristics of international mobile adolescents. Journal of School Psychology, 30, 197–214. Gerner, M. E., & Perry, F. (2000). Gender differences in cultural acceptance and careers orientation among internationally mobile and non-internationally mobile adolescents. School Psychology Review, 29, 267–284. Greenholtz, J., & Kim, J. (2009). The cultural hybridity of Lena: A multi-method case study of a TCK. International Journal of Intercultural Relations, 33, 391–398. Haritatos, J., & Benet-Martínez, V. (2002). Bicultural identities: The interface of cultural, personality, and socio-cognitive processes. Journal of Research in Personality, 36(6), 598–606. Hogg, M. A., & Abrams, D. (1999). Social identity and social cognition: Historical background and current trends. In D. Abrams, & M. A. Hogg (Eds.), Social identity and social cognition. (pp. 1–25). Oxford, UK: Blackwell. Hogg, M. A., & Mullin, B. A. (1999). Joining groups to reduce uncertainty: Subjective uncertainty reduction and group identification. In D. Abrams, & M. A. Hogg (Eds.), Social identity and social cognition (pp. 179–249). Oxford, UK: Blackwell. Hong, Y. Y., Morris, M. W., Chiu, C. Y., & Benet-Martinez, V. (2000). Multicultural minds: A dynamic constructivist approach to culture and cognition. American Psychologist, 7, 709–720. Isogai, T. Y., Hayashi, Y., & Uno, M. (1999). Identity issues and reentry training. International Journal of Intercultural Relations, 23, 493–525. Jones, B. V., & Austin, C. N. (1987). Reentry among missionary children: An overview of reentry research from 1934 to 1986. Journal of Psychology and Theology, 14, 315–325. LaFromboise, T., Coleman, H. L., & Gerton, J. (1993). Psychological impact of biculturalism: Evidence and theory. Psychological Bulletin, 114, 395–412. Lewin, K. (1948). Resolving social conflicts. New York: Harper. McCaig, N. M. (1996). Understanding global nomads. In C. Smith (Ed.), Strangers at home (pp. 99–120). New York: Aletheia Press. Navarrete, V. (1999). Ethnically mixed individuals: Cultural homelessness or multicultural integration? Unpublished thesis, University of North Texas, Denton. Navarrete, V., & Jenkins, S. R. (2010). Cultural homelessness, multiminority status, ethnic identity development, and self esteem. Unpublished manuscript, University of North Texas, Denton. O’Neill, J. (2001, November 30). Global Economics Paper No. 66. GS Global Economics Website. Retrieved from www.researchselectfund. com/ideas/brics/building-better-doc.pdf. Phinney, J. S. (1990). Ethnic identity in adolescents and adults: A review of research. Psychological Bulletin, 108, 499–514. Phinney, J. S. (1992). The Multigroup Ethnic Identity Measure. Journal of Adolescent Research, 7, 156–176. Piaget, J. (1972). Intellectual evolution from adolescence to adulthood. Human Development, 15(1), 1–12. Pirinen, M. (2002). In place, out of place: Meaning of place and question of belonging for adult-Third Culture Kids. Unpublished master’s thesis, University of Joensuu, Finland. Podolsky, M. K. (2004). Cross-cultural upbringing: A comparison of the ‘Third Culture Kids’ framework and ‘Kaigai/Kikoku-shijo’ studies. Contemporary Society (Kyoto Women’s University Bulletin), 6, 67–78. Pollock, D., & Van Reken, R. (1999). Third culture kids: The experience of growing up among worlds. Yarmouth, ME: Intercultural Press. Ponterotto, J. G., Gretchen, D., Utsey, S. O., Stracuzzi, T., & Saya, R. (2003). The Multigroup Ethnic Identity Measure (MEIM): Psychometric review and further validity testing. Educational and Psychological Measurement, 63, 502–515. Rosenberg, M. (1989). Society and the adolescent self-image. Middletown, CT: Wesleyan University Press. Rudmin, F. (2009). Constructs, measurements and models of acculturation and acculturative stress. International Journal of Intercultural Relations, 33(2), 106–123. Ryder, A., Alden, L., & Paulhus, D. (2000). Is acculturation unidimensional or bidimensional? A head-to-head comparison in the prediction of personality, self-identity, and adjustment. Journal of Personality and Social Psychology, 79, 49–65. Schachter, J. (2001, May). Geographical mobility: Population characteristics. Current Populations Report. Retrieved from http://www.census. gov/prod/2001pubs/p20-538.pdf. Schultz, D. M. (1986). A study of third culture experience in relation to the psycho-cultural adjustment of returned Church of Christ missionaries families. Unpublished doctoral dissertation, University of Nebraska, Lincoln. Sorti, C. (1997). The art of coming home. Yarmouth, ME: Intercultural Press. Streiner, D. L. (2003). Being inconsistent about consistency: When coefficient alpha does and doesn’t matter. Journal of Personality Assessment, 80, 217–222. Stringham, E. M. (1993). Reacculturation of missionary families: A dynamic theory. Journal of Psychology and Theology, 21, 66–73. Sussman, N. M. (2000). The dynamic nature of cultural identity throughout cultural transitions: Why home is not so sweet. Personality and Social Psychology Review, 4, 355–373. Szkudlarek, B. (2010). Reentry: Review of the literature. International Journal of Intercultural Relations, 34, 1–21. Tajfel, H. (1982). Social psychology of intergroup relations. Annual Review of Psychology, 33, 1–39. Tajfel, H., & Turner, J. C. (1979). An integrative theory of intergroup conflict. In W. Austin, & S. Worchel (Eds.), The social psychology of intergroup relations (pp. 33–47). Monterey, CA: Brooks/Cole. Thompson, M. (2009). Appreciation: Passage of a global nomad, Norma M. McCaig, 1945–2008. Foreign Service Journal, 86(2), 36–37. Useem, R. H. (1999). Third Culture Kids: Focus of major study—TCK “mother” pens history of field. Retrieved from www.tckworld.com/useem/artl.htlm. Useem, R. H., & Cottrell, A. B. (1996). Adult third culture kids. In C. Smith (Ed.), Strangers at home: Essays on the effects of living overseas and coming home to a strange land (pp. 22–35). Bayside, NY: Alethea. Useem, R. H., Useem, J., Cottrell, A. B., & Jordan, K. A. F. (1993). Research on adult TCKs. In NewsLink. Princeton, NJ: International Schools Services. Viser, W. C. (1978). A psychological profile of missionary children in college and the relationship of intense group therapy to weekly group therapy in the treatment of personality problems as reflected by the by the Minnesota Multiphasic Personality Inventory. Unpublished doctoral dissertation, Southwestern Baptist Theological Seminary, Fort Worth, TX. Vivero, V. N., & Jenkins, S. R. (1999). Existential hazards of the multicultural individual: Defining and understanding “cultural homelessness”. Cultural Diversity & Ethnic Minority Psychology, 5, 6–26. Vygotsky, L. S. (1934/1986). Thought and language. Cambridge, MA: MIT Press.