Transnational Relationships and Cultural Identity of Older Migrants

Social Identity Theory and Cultural Identity

Social identity theory (Tajfel & Turner, 1986) and the closely allied self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) have roots in social interactionism in which the self is a product of social interaction (Mead, 1934). Both theories explain self-identity in terms of group processes and intergroup relationships. Broadly speaking, social identity theory concerns the affective-evaluative element of intergroup relationships, while self-categorization theory comprises the cognitive-perceptual component. In this paper both are subsumed within the “social identity approach.” This approach suggests that the degree of social identification with a particular group in society is based on perceived shared characteristics, beliefs, and/or behaviors. Groups may be based on nationality, political affiliation, religion, sex, geographical location, social status, and so on, or based on combinations thereof. Social identities “both describe and prescribe one’s attributes as a member of that group [...] that is, what one should think and feel and how one should behave” (Hogg, Terry, & White, 1995, p. 206). Self-categorization into a group (e.g., based on certain attributes such as economic status, sex, and age) underscores the extent to which one feels one belongs to that group. It provides a sense of self-identity and self-verification (Hogg et al., 1995).

A social identity approach focuses on intergroup relationships. Self-categorization within a particular group entails differentiation from other groups by pronouncing or accentuating positive attributes of the ingroup and negative attributes of the outgroup(s), resulting in self-enhancing outcomes such as improved self-esteem for members of the ingroup (Stets & Burke, 2000). The strength of comparison with other groups are influenced by subjective belief structures concerning the perceived legitimacy and stability of the status that the group holds (Hogg et al., 1995).

A social identity approach can account for the complexity of multiple, context-specific social identities, because people may have as many distinct self-defined identities as there are groups whose opinions matter to them (James, 1890/1950). Phinney and Ong (2007) noted that “a commitment, or sense of belonging, is perhaps the most important component” of cultural identity (p. 272). Thus, we operationalize cultural identity as the extent to which migrants feel they belong to the country in which they reside (the UK), the country of origin of their family, or ethnic group. In this article, we take a social identity approach to examine the potential for different influences on cultural identity arising from group membership based on cultural heritage, social class, support network, and global citizenship (transnational relationship type) (Roccas & Brewer, 2002).

Cultural Heritage and Cultural Identity

Cultural heritage differs from cultural identity inasmuch as the former is “made in the present but has recourse to the past” (Aronsson & Gradén, 2013, pp. 9–10) and is based on the singularity or plurality of ethnic group membership over time, while the latter is based on an individual’s sense of belonging (Phinney & Ong, 2007). Thus, an individual’s cultural heritage is forged within the culture of the family of origin and the trappings that accompany that culture, such as the dress code, food, traditions, and values (UNESCO, 2003). However, cultural heritage is also influenced by the culture and accoutrements of the place or places in which one lives or has lived, so that inheritance of culture from generation to generation can take on new as well as familiar forms. Cultural heritage is to a certain extent governed by one’s circumstances and a common cultural heritage would be dependent on shared histories. For example, many Indians living in the UK have never lived in India, but in fact were expelled from Africa (e.g., from Uganda, Kenya and Tanzania) during the Africanization of labor in the 1970s (Burholt, 2004a) and are likely to have a mutual cultural heritage. In this article, we use the terms “ethnic group” and “length of residence in the UK” to indicate a common historical cultural heritage.

Cultural identity is expressed through self-categorization into cultural group(s) because the person feels that they belong (van Oudenhoven, Ward, & Masgoret, 2006). Thus, cultural identity is a personal experience of representation and “part of an individual’s self-concept which derives from knowledge of membership of a social group (or groups) together with the value and emotional significance attached to that membership” (Tajfel, 1981, p. 255). Some research suggests that long-term exposure to racism (Karlsen & Nazroo, 2013) or identity denial (whereby individuals from ethnic groups are seen as less prototypically British because of their language, appearance, attitudes or behavior; Cheryan & Monin, 2005) may decrease a sense of British identity. However, the majority of research suggests that emotional ties to and identification with places and communities strengthen over time (Burholt, Curry, Keating, & Eales, 2014). Thus, our first hypothesis posits that (H1a) there will be differences between ethnic groups with regard to the time spent in the UK and (H1b) longer residence in the UK will be associated with a stronger British identity.

Social Class and Cultural Identity

Many commentators suggest that Britain is no longer a “deeply class conscious society” (Savage, 2000, p. 40). Nowadays, although class inequalities persist, social class rarely confers membership of a particular group in the UK. Despite a perceived detachment of cultural identity from social class (Bottero, 2004), this may not hold true for all groups in society. Older migrants may identify with a particular social class and be influenced by the perceived or actual power accorded to it. A few studies have found that higher social class is linked to stronger ethnic identity, possibly because better education produces heightened ethnic awareness (Khanna, 2016). For example, Asian and Black Americans (with partners from a different ethnic group) are more likely to label children as Asian or Black, respectively (Qian, 2004; Xie & Goyette, 1997). Conversely, a majority of theorists suggest that the stigmatization and devaluing of lower status groups may reinforce ethnic solidarity and identity (Khanna, 2016). For example, in the USA many Black youths develop positive ethnic identities rejecting negative and low-status stereotypes (Cross & Fhagen-Smith, 2001). For some older migrants in the UK (e.g., older Sylheti men), patterns of settlement, hypercyclical employment, and unemployment may contribute to shared socioeconomic class experience and decreased British identity (Burholt, 2004a). Given the weight of evidence concerning the direction of influence of social class and ethnic identity, we hypothesize that (H2a) migrants with lower socioeconomic status will have a stronger ethnic identity. However, with a focus on British identity, we hypothesize that (H2b) migrants with lower socioeconomic status will have a weaker British identity than those with higher socioeconomic status.

Social Support Networks and Cultural Identity

Each older migrant is embedded within a social network: a configuration of family, friends and neighbors to whom they have social ties. The social-support network is a vehicle for potential support and also may be responsible for determining individual behavior. In this respect, the selection of group membership (a social-support network) is likely to be influenced by the extent to which members of the group reinforce or give credibility to a commitment to a particular social identity that has been developed over time (Hitlin & Elder, 2007). Research has found that cultural diversity within networks (versus monocultural uniformity) contributes to multicultural identities (Mao & Shen, 2015). Thus, we would expect older people with diverse social-support networks to express their cross-cultural interconnectedness through the strength of plural cultural identities. While family-focused monocultural networks are likely to have strong ethnic identities, there is no reason to expect them to be any stronger or weaker than for those people with diverse networks. On the other hand, we hypothesize that (H3) older people with diverse networks will have stronger British identities than those with family focused networks.

Transnational Family Relationships and Cultural Identity

Transnationalism is defined as the process by which migrants forge and sustain multistranded social relations that link their societies of origin and settlement through the creation of cross-border and intercontinental networks (Vertovec, 1999). Transnationalism is facilitated by “space- and time-compressing technology” (Portes, Guarnizo, & Landolt, 1999, p. 222) within an infrastructure that encompasses new communication technology and relatively easy long-distance cross-border travel. Ten years ago, research with South Asian elders found that different migrant groups in the UK were more or less disposed to maintaining transnational social bonds with relatives through telephone contact, sending remittances, or gifts and visiting (Burholt, 2004b; Burholt & Wenger, 2004). Transnational contact using information communication technology (ICT) such as email and voice or video calls via the internet (e.g., Skype) was less common at that time. Some authors have suggested that globalization through the use of ICT has generated movement toward a “homogeneous global village” in which national identities are jeopardized (Neyestani & McInturff, 2006, p. 91). In the present study we expect specific types of transnational family behavior to impact on the formation of national cultural identities. Specifically, we posit that (H4a) transnational relationships typified by frequent visiting and other forms of traditional transnational contact between family members living abroad are associated with a stronger cultural identity with the family’s country of origin and a stronger ethnic identity; and that (H4b) transnational relationships characterized by contact through ICT are associated with a weaker cultural identity with the family’s country of origin, ethnic identity and British identity.

We would expect to see associations between the types of transnational relationships (based on frequency and type of contact) and cultural identity for specific dyadic relationships (e.g., parent-child, siblings, and relatives). For example, research showed that parent-child relationships are more frequently described in terms of functional exchanges of support. Relationships with siblings are significant throughout life, and most siblings have a commitment to maintaining the relationship. There is also evidence to indicate that sibling relationships may become more affectionate in later life (Wenger & Burholt, 2001). However, there are differences between cultures in their expectations concerning functional reciprocity (Burholt & Dobbs, 2010). The extent to which transnational relationship dyads enact or deviate from these affectionate or functional norms may impact older people’s sense of ethnic identity. Because there is insufficient evidence on transnational relationships to support hypotheses, in this article we explore the associations between transnational relationship types and cultural identity separately for (1) parent-child, (2) sibling, and (3) other relative dyads.

Procedure

We identified sampling points for each ethnic group in England and Wales using the Postcode Address File (PAF) that divides the United Kingdom into postcode districts comprising around 30,000 postcodes (Royal Mail, 2009). Information from the UK Census 2001 was used to construct a data file for populations of each ethnic group organized at the PAF locality level. The PAF localities were ordered according to the population size of each ethnic group, separately for each ethnic group in England and each ethnic group in Wales (12 lists in total); systematic random sampling was used to select sampling points on each list.

The interview schedule was compiled in English and included items from a project conducted in South Asia and Birmingham (Burholt & Dobbs, 2014) and from the Survey of Household Carers 2009–2010 (National Health Service Information Centre, 2010). Questions were comparable to the original English items in terms of appropriateness and meaning and considered conceptual matters, cultural relevance, and the subtle connotations of words and phrases (Hunt, 2003). For the majority of survey questions, conceptual and functional equivalence was straightforward: these items were translated directly from English to Punjabi, Gujarati, Hindi, Mandarin Chinese, Bengali, Somali, Yoruba, or Urdu during the course of the interview. This method of translation is standard practice for the research group (e.g., Grant & Bowling, 2011). Twenty questions were translated into the respective eight languages using front/backtranslation methods (Koller et al., 2007).

Interviews were conducted by a market research group (Ethnic Focus) between October 2011 and April 2012 in the respondent’s native language or in English and, wherever possible, in the respondents’ own homes. Potential participants approached by Ethnic Focus who did not identify with one of the six ethnic groups were not eligible for inclusion in the study and were not interviewed.

Sample

Overall, the response rate was 40% (Black African: 41%; Black Caribbean: 41%; Indian: 48%; Pakistani: 44%; Bangladeshi: 41%; and Chinese: 27%). The net final size was N = 1,206. This article is based on a subsample of older people aged 55 years or more. The subsample comprised n = 815 people (Black African: n = 143; Black Caribbean: n = 162; Indian: n = 141; Pakistani: n = 141; Bangladeshi: n = 134; Chinese: n = 94).

Each participant was asked to provide information on relationships with all living children and siblings, plus other relatives with whom the participant was in contact (in the last year). Rosters were constructed for children, siblings and other relatives. For example, a respondent with two children would have data on the first two lines of the parent-child roster. From the roster datasets of parent-child, sibling and relative dyads n = 1736 transnational family relationships emerged, that is, where the child (n = 317), sibling (n = 747) or other relative (n = 672) lived overseas.

Descriptive statistics of the sample by ethnic group are presented in Table 1. Black Caribbean participants were on average the oldest participants (M = 68 years) (F(5, 809) = 2.27, p < .05) and were more likely than other participants to be never married, divorced or separated. In all other ethnic groups, a majority of participants were married, and between one-fourth and one-third were widowed (χ² = 90.98, df 15, p < .001). There were no differences between ethnic groups in the sex of the participants, with the male to female ratio roughly 50:50.

Table 1 Demographic characteristics and cultural identity of the sample by ethnic group and for all participants

		Black Caribbean	Black African	Indian	Pakistani	Bangladeshi	Chinese	All
		Mean (SD) Range	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)	Mean (SD)
		[N = 162]	[N = 143]	[N = 141]	[N = 141]	[N = 134]	[N = 94]	[N = 815]
Note.*p = .05, **p ≤ .01, ***p ≤ .001. Analysis of variance for age, years living in the UK, and cultural identity using posthoc group comparisons Tukey HSD test: numbers that appear in bold italics constitute subsets with the highest values; numbers that appear in italics constitute subsets with the lowest values.
Age*		68.0 (6.8) 55–85	66.2 (5.5) 55–78	55.8 (6.4) 55–92	67.1 (5.4) 55–82	66.1 (4.5) 55–78	67.5 (6.7) 55–89	67.0 (6.0) 55–92
Years living in the UK***		51.3 (5.3) 21–67	31.1 (12.7) 4–72	42.1 (9.6) 7–65	39.5 (8.5) 11–62	34.9 (6.6) 11–49	38.5 (9.1) 13–64	40.0 (11.1) 4–72
Sex
	Male	50.6%	51.7%	48.9%	46.1%	48.5%	43.6%	48.6%
	Female	49.4%	48.3%	51.1%	53.9%	51.5%	56.4%	51.4%
Marital Status***
	Never married	7.4%	4.2%	2.8%	0%	0%	3.2%	3.1%
	Married/partnership	49.4%	53.1%	64.5%	68.8%	65.7%	64.9%	60.5%
	Divorced/separated	27.2%	16.8%	2.8%	6.4%	3.7%	12.8%	12.0%
	Widowed	16.0%	25.9%	29.8%	24.8%	30.6%	19.1%	24.4%
NS-SEC classification of occupation***
	Managerial and professional	16.1%	14.3%	14.3%	9.2%	9.0%	18.1%	13.3%
	Intermediate	23.0%	7.1%	17.9%	10.6%	12.7%	22.3%	15.4%
	Routine and manual	41.0%	36.4%	46.4%	33.3%	38.1%	36.2%	38.8%
	Never worked and long-term unemployed	19.9%	42.1%	21.4%	46.8%	40.3%	23.4%	32.5%
Network type***
	Multigenerational Household: Older Integrated Network	16.7%	16.1%	22.7%	21.3%	17.9%	21.3%	19.1%
	Multigenerational Household: Younger Family Network	29.6%	54.5%	46.8%	58.2%	68.7%	36.2%	49.1%
	Middle-Aged Friends Network	23.5%	12.6%	14.2%	8.5%	6.7%	20.2%	14.2%
	Restricted Nonkin Network	30.2%	16.8%	16.3%	12.1%	6.7%	22.3%	17.5%
Cultural identity
	Ethnic identity***	4.4 (1.1) 1–5	4.4 (1.0) 1–5	4.8 (0.6) 1–5	4.5 (0.9) 1–5	4.5 (0.8) 2–5	4.0 (1.1) 1–5	4.5 (0.9) 1–4
	With family’s country of origin***	4.2 (1.2) 1–5	4.3 (1.1) 1–5	4.7 (0.7) 1–5	4.4 (0.9) 2–5	4.5 (0.7) 3–5	4.0 (1.0) 2–5	4.4 (1.0) 1–5
	British identity***	3.4 (1.5) 1–5	3.1 (1.3) 1–5	4.4 (1.0) 1–5	3.6 (1.1) 1–5	3.4 (1.0) 1–5	3.4 (1.0) 2–5	3.6 (1.3) 1–5

Table 1 Demographic characteristics and cultural identity of the sample by ethnic group and for all participants

View as image HTML

Measures

Data Analyses

The Statistical Package for the Social Sciences (SPSS) Version 22 was used to analyze data. We compared cultural identity using Wilcoxon signed rank test to contrast participants’ ethnic identity, cultural identity with the families’ country of origin and British identity. For each form of cultural identity, three multiple regression models were run to determine the relative contribution of the independent variables (nine regression models in total). Each model included sex, socioeconomic class, number of years living in the UK, ethnic group, network type, and a count of transnational relationships types. The latter varied in each of the three models because we wanted to explore the associations for parent-child, sibling, and relative dyads separately. The inclusion of all counts of transnational relationship dyads in one model would identify the strongest predictive relationship with cultural identity, but would not necessarily identify all significant associations. Model 1 included the count of each type of transnational relationship for relative dyads, Model 2 included the count of each type of transnational relationship for parent-child dyads only, and Model 3 included the count of each type of transnational relationships for sibling dyads only. A variance inflation factor (VIF) was calculated for each predictor in the models, with values greater than 10 indicating high levels of multicollinearity (Hair, Anderson, Tatham, & Black, 1995).

Regression Analysis

Adjusted R² values were small indicating that the models only explained a small amount of variance. However, each model had a significantly better fit to the data than the intercept only model and contained statistically significant predictors from which we draw important conclusions about cultural identities (Table 2). The VIF value for all predictors in each model was < 3, indicating that there was not a high degree of multicollinearity (Hair et al., 1995).

Table 2 Regression results predicting cultural identity through ethnic identity, cultural identity with family’s country of origin and British identity

Strength of belonging to:		Ethnic group			Family’s country of origin			UK
		b		SE	b		SE	b		SE
Note: Models and coefficients not marked with asterisks were not significant. Reference groups for categorical variables: ethnicity: Bangladeshi; support networks: multigenerational: older integrated. *p < .05, **p = .01, ***p = .001.
Model 1
	Sex	0.19**		0.07	0.15*		0.07	0.06		0.09
	Socioeconomic status	–0.04		0.03	–0.05		0.04	–0.19***		0.04
	# years living in the UK	0.00		0.00	–0.00		0.00	0.02**		0.01
Ethnicity
	Black Caribbean	–0.01		0.13	–0.17		0.13	–0.34*		0.16
	Black African	–0.02		0.11	–0.13		0.12	–0.19		0.15
	Indian	0.36**		0.12	0.24*		0.12	0.83***		0.15
	Pakistani	0.01		0.11	–0.08		0.11	0.16		0.15
	Chinese	–0.47***		0.13	–0.50***		0.13	–0.10		0.16
Network type
	Middle-Aged Friends	–0.38***		0.11	–0.29*		0.12	–0.22		0.15
	Restricted Nonkin	–0.16		0.11	–0.08		0.11	–0.04		0.14
	Multigenerational: Younger Family	–0.05		0.09	–0.04		0.09	–0.18		0.11
# Transnational Relationship: Relatives
	Infrequent Digital Communicators	–0.12*		0.06	–0.10		0.06	–0.08		0.07
	Infrequent Telephone Communicators	0.00		0.02	–0.01		0.02	–0.01		0.02
	Highly Connected Regular Benefactors	0.06		0.03	0.07*		0.03	0.08		0.04
	Occasional Bilateral Bounteous Visitors	0.08		0.06	0.08		0.06	0.01		0.08
Adj R2			.08***			.06***			.14***
Model 2
	Sex	0.18**		0.07	0.15*		0.07	0.06		0.09
	Socioeconomic status	–0.03		0.03	–0.04		0.04	–0.18***		0.04
	# years living in the UK	–0.00		0.00	–0.00		0.00	0.01**		0.01
Ethnicity
	Black Caribbean	–0.05		0.13	–0.21		0.13	–0.37*		0.16
	Black African	–0.03		0.11	–0.14		0.12	–0.19		0.15
	Indian	0.36**		0.12	0.25*		0.12	0.86***		0.15
	Pakistani	0.02		0.11	–0.07		0.11	0.18		0.14
	Chinese	–0.50***		0.13	–0.51***		0.13	–0.10		0.16
Network type
	Middle-Aged Friends	–0.38***		0.11	–0.29*		0.12	–0.23		0.15
	Restricted Nonkin	–0.16		0.11	–0.08		0.11	–0.05		0.14
	Multigenerational: Younger Family	–0.05		0.09	–0.04		0.09	–0.19		0.11
# Transnational Relationships: Children
	Infrequent Digital Communicators	0.24		0.20	0.32		0.21	0.56*		0.26
	Infrequent Telephone Communicators	–0.06		0.05	–0.08		0.05	–0.06		0.06
	Highly Connected Regular Benefactors	0.08		0.07	0.11		0.07	0.08		0.09
	Occasional Bilateral Bounteous Visitors	0.12		0.28	0.12		0.29	0.21		0.36
Adj R2			.08***			.06***			0.14***
Model 3
	Sex	0.18**		0.07	0.14*		0.07	0.06		0.09
	Socioeconomic status	–0.05		0.03	–0.05		0.04	–0.19***		0.05
	# years living in the UK	0.00		0.00	–0.00		0.00	0.01***		0.01
Ethnicity
	Black Caribbean	–0.01		0.13	–0.16		0.13	–0.35*		0.17
	Black African	–0.02		0.11	–0.14		0.12	–0.20		0.15
	Indian	0.34***		0.12	0.28*		0.12	0.85***		0.15
	Pakistani	0.02		0.11	–0.07		0.12	0.17		0.15
	Chinese	–0.46***		0.13	–0.48***		0.13	–0.10		0.17
Network type
	Middle-Aged Friends	–0.37***		0.11	–0.28*		0.12	–0.21		0.15
	Restricted Nonkin	–0.17		0.11	–0.10		0.11	–0.05		0.14
	Multigenerational: Younger Family	–0.05		0.09	–0.04		0.09	–0.18		0.11
# Transnational Relationships: Siblings
	Infrequent Digital Communicators	–0.10		0.09	–0.08		0.09	–0.08		0.12
	Infrequent Telephone Communicators	0.04		0.03	0.02		0.03	0.01		0.04
	Highly Connected Regular Benefactors	0.20*		0.09	0.17		0.10	0.13		0.12
	Occasional Bilateral Bounteous Visitors	0.04		0.20	0.12		0.10	–0.01		0.13
Adj R2			.08***			.06***			.13***

Table 2 Regression results predicting cultural identity through ethnic identity, cultural identity with family’s country of origin and British identity

View as image HTML

Sociodemographic variables and ethnicity were associated with cultural identity across all nine models. A stronger ethnic identity and cultural identity with the family’s country of origin were predicted by being male and Indian, while a weaker ethnic identity and cultural identity with the family’s country of origin were predicted by being Chinese. A stronger British identity was predicted by higher socioeconomic status and a greater number of years living in the country and being Indian. A weaker British identity was related to the Black Caribbean ethnic group. The results were consistent with H1b, whereby longer residence in the UK was associated with a stronger British identity; and H2b inasmuch as migrants with lower socioeconomic status had a weaker British identity than those with higher socioeconomic status. However, H2a was not supported, and we failed to reject the null hypothesis, that is, there is no association between social class and ethnic identity.

Considering the influence of the role of support networks in cultural identity, having a Middle-aged Friends Network was associated with a weaker ethnic identity (Models 1–3) and cultural identity with the family’s country of origin (Models 1–3), but network type was not associated with British identity. This was inconsistent with hypothesis H3 that stated that older people with diverse networks would have a stronger British identity than those with family focused networks.

The effect of transnational family relationships differed between models and for each type of cultural identity. First, Highly Connected Regular Benefactor transnational relationships were typified by frequent visits and traditional forms of contact with relatives abroad. There was a positive association between the number of Highly Connected Regular Benefactor relationships that older participants had with siblings and ethnic identity (Model 3). Every additional Highly Connected Regular Benefactor transnational relationship with a sibling (e.g., two transnational Highly Connected Regular Benefactor sibling relationships instead of one) increased the ethnic identity score by 0.28. Similarly, there was a positive relationship between the number of Highly Connected Regular Benefactor relationships with other relatives and cultural identity with the family’s country of origin (Model 1). Every additional Highly Connected Regular Benefactor relationship with other relatives abroad strengthened cultural identity with the family’s country of origin by 0.29. The results supported hypothesis H4a, demonstrating that transnational relationships that are typified by frequent visiting and other forms of traditional transnational contact were associated with a stronger ethnic identity (although this was true only for relationships with siblings), and a stronger cultural identity with the family’s country of origin (although this was true only for relationships with other relatives).

Second, Infrequent Digital Communicator transnational relationships were characterized by contact through network digital communication such as email and Skype. The analysis showed that a greater number of Infrequent Digital Communicator relationships with other relatives was associated with a weaker ethnic identity (Model 1). Every additional Infrequent Digital Communicator transnational relationship with other relatives decreased the strength of ethnic identity by 0.12. On the other hand, the positive association between the number of Infrequent Digital Communicator relationships between older adults and children overseas and British identity suggested that these types of relationships may bolster the sense of belonging to the place of residence (Model 2). Every additional Infrequent Digital Communicator with a child overseas increased the strength of British identity by 0.56. The results partially supported hypothesis H4b and demonstrated that transnational relationships with other relatives (but not children or siblings) characterized by contact through ICT were associated with a weaker sense of ethnic identity. However, the results did not demonstrate a negative association between these types of relationship and cultural identity with the family’s country of origin or British identity. With regard to the latter, the converse was true and Infrequent Digital Communicator transnational relationships with children increased the strength of British identity.

Limitations

There are limitations to the research reported here. First, gaining access to participants from Black and Minority Ethnic groups is challenging. We achieved our target sample for five out of our six groups, but were unable to recruit sufficient Chinese participants, only achieving 75% of our target. In addition, our overall response rate was 40%. Focusing on five ethnic groups (excluding the Chinese), the response rate increases to 43%, which is broadly in line with “nongovernment” sponsored surveys.

Second, the research highlights the need for caution in reporting generalizations based on an amalgamation of ethnicities, cultures and personal histories. The results draw attention to the possibility of heterogeneous ethnic identities for Chinese elders in the UK (e.g., Hong Kong Chinese, Malaysian Chinese, and Singaporean Chinese). Elsewhere we have shown that important differences exist between “South Asians” (Burholt, 2004b), and subsequently research has moved toward demarcation between migrant groups. Older Chinese migrants should be accorded the same consideration as members of the same broad ethnic classification may have different ethnic identities (see also Fouad & Brown, 2000).

Third, our analyses are limited because of (1) the cross-sectional nature of data and (2) the low variance that the models accounted for. Longitudinal studies with ethnic groups in the UK (which presently do not exist) could provide opportunities to test causal pathways explaining variance. However in order to better understand cultural identity, the theorizing behind the relationship between predictors and the dependent variables was explicit. The unexplained variance is not an indicator of the quality or explanatory power of the social identity approach to studying cultural identity. However, it may be an indication that such models would benefit from refining. We speculate that the explanatory power may be rationalized to a large degree by the complexities of operationalizing cultural heritage. Cultural heritage is not yet crystallized into a coherent theory, and it is likely that individual experiences (e.g., migration history and settlement patterns) were insufficiently captured by ethnic group and length of residence in the UK. Moreover, our models did not include personality characteristics (Padilla & Perez, 1993) which may influence individuals to associate with, or reject (more or less strongly) different forms of cultural identity. However, understanding the link between individual characteristics, group membership (social class, social-support networks), transnational family behavior and cultural identity offers the advantage of a “generalizable framework” and may inform future investigations.

Implications

Overall, our analysis shows that migrants have complex multiple cultural identities based on cultural heritage, social class, social-support networks, and transnational relationships. Our novel focus on transnational family relationships demonstrated that it is not just contact with relatives abroad that is important to cultural identity, but the type of relationship and who the relationship is with (i.e., all relatives, children, or siblings). We demonstrated that transnational family relationships provide migrants with a range of alternatives identities into which they can self-categorize and identify with, or with which they can contrast to their group identity. This is important because to date research on transnational family ties has not examined the complexity underpinning these relationships.

Transnational relationships with relatives can reinforce a sense of belonging to a country of origin through high levels of investment in communities overseas. They can strengthen a sense of ethnic identity through emphasizing similarities with siblings, yet strengthen British identity through accentuating differences between the Britishness of parents when contrasted with the cultural identity of children living abroad. Our findings also suggest that one form of identity does not necessarily come at the expense of another (Nandi & Platt, 2015). Indeed, the plurality of cultural identity suggests that migrants can maintain multiple identities that are not necessarily oppositional (Verkuyten, 2007). Further research is warranted to determine whether the plurality of cultural identity is most likely to contribute to social harmony.