Introduction
Population ageing and international migration are generally studied separately. Although the number of immigrants in many western countries has risen substantially in the last decades; until recently immigrant older people were expected to return to their home-countries (Warnes et al. Reference Warnes, Friedrich, Kellaher and Torres2004). It is, however, becoming increasingly apparent that the majority of immigrant older people in host countries like The Netherlands will stay. In the near future, a rising proportion of older people in many western societies will be of immigrant origin, with several implications for care and family relations. This article focuses on one aspect of the immigrant older person's experience and situation, their perceptions about filial obligations. It is often assumed that older immigrants are supported by families with strong interdependent ties and filial obligations (Bolzman et al. Reference Bolzman, Poncioni–Derigo, Vial and Fibbi2004). To what extent such family relationships continue after the migration and the passage of time is still largely unexplored.
In most studies, filial obligation norms are examined from the point of view of the younger generation. Studies on changes in family relations and related perceptions of filial obligations have mainly focused on how and to what extent immigrant youth adapts to the host society, while attitudes and beliefs from the perspective of older people have received less attention. Immigrants in western countries are likely to experience old age differently from native older people, and to have different beliefs and attitudes about their children's obligations, but little is known about ethnic-group variations in either family forms of attitudes and expectations (Seelbach and Die Reference Seelbach and Die1988). North-American studies have found that parents’ expectations of their children differ by ethnic group (Burr and Mutchler Reference Burr and Mutchler1999; Fuligni and Pedersen Reference Fuligni and Pedersen2002; Lee, Peek, and Coward Reference Lee, Peek and Coward1998; Rosenthal Reference Rosenthal1986). Immigrant families are thought to adhere to the traditions in their countries of origin, in particular in the private domain, but it remains unclear how and to what extent cultural factors account for inter-ethnic differences. Previous studies have suggested that differences arise from the contrasting demographic and socio-economic characteristics of natives and immigrants (Glick and Van Hook Reference Glick and Van Hook2002; Mitchell, Wister and Gee Reference Mitchell, Wister and Gee2004), and that ethnic differences in family relations are subject to the migration experience itself and may not be as large as assumed (Silverstein and Waite Reference Silverstein and Waite1993; Schans and Komter Reference Schans and Komter2006).
In this paper, we examine the relative influence of migration history and socio-demographic factors on the perceptions of filial obligation among native Dutch and four immigrant groups of older people in The Netherlands. A large comparative survey included the native Dutch, Turkish and Moroccan guest workers, and immigrants from the (former) Dutch colonies of Suriname and the Dutch Antilles. In this article, the four latter groups are described as ‘immigrants’ because the majority migrated to The Netherlands as young adults. The current population of The Netherlands includes many different immigrant groups and their descendants. Most immigrant groups apart from those included in the survey have a young age structure, however, and do not yet include many older people. The objectives of this paper are therefore threefold: to provide evidence of the perceptions of filial obligation among older people with different ethnic backgrounds in The Netherlands; to investigate if and how these attitudes differ among the various ethnic groups; and to examine whether differences in attitudes towards filial obligation are attributable to ethnic background or other factors including socio-demographic background and acculturation.
Research on filial obligations
Filial obligation refers to a societal attitude that prescribes a duty of (adult) children to meet the needs of their ageing parents (Seelbach and Die Reference Seelbach and Die1988; Walker et al. Reference Walker, Pratt, Shinn and Jones1990). Rossi and Rossi (Reference Rossi and Rossi1990) defined norms of filial responsibility as culturally-defined rights and duties that specify both the ways in which family members are expected to behave toward each other and the obligations to exchange and provide support to one another. It is argued that an individual's expectations of and attitudes toward filial obligation develop during socialisation, through personal experiences as well as by observing relationships between family members of different generations (Goldscheider and Lawton Reference Goldscheider and Lawton1998; Burr and Mutchler Reference Burr and Mutchler1999). Most research has examined filial obligations from the view point of the young generation. The few studies that focus on the beliefs and attitudes of older people have found relationships between various socio-demographic characteristics and levels of adherence to filial obligations. Lee, Netzer and Coward (Reference Lee, Netzer and Coward1994) and Lee, Peek and Coward (Reference Lee, Peek and Coward1998) found that elderly parents of low socio-economic status expected to receive more help from their adult children than did those in higher status groups: the same applied to unmarried and unhealthy parents. Likewise, Seelbach (Reference Seelbach1981) found that, as parents age and grow poorer, they expect to receive more support from their children.
There is little empirical research on the influence of ethnic background on attitudes to filial obligations. Most of the available studies have been of a single minority group or have methodological limitations, such as small unrepresentative samples (Burr and Mutchler Reference Burr and Mutchler1999; Gans and Silverstein Reference Gans and Silverstein2006). Their findings nevertheless indicate that norms of filial responsibility vary by cultural background, although are inconclusive about the direction of causation. Lee and Aytac (Reference Lee and Aytac1998) found that Black elderly parents in the United States had higher expectations of filial obligations than Whites, but Seelbach (Reference Seelbach1981) and Hanson, Sauer and Seelbach (Reference Hanson, Saur and Seelbach1983) did not corroborate the difference. Lee, Peek and Coward (Reference Lee, Peek and Coward1998) argued that among Blacks, extensive support networks developed between generations to combat the negative effects of prejudice and discrimination, and that they had a ‘cultural aversion’ to formal services, seeing them as exploitive rather than supportive. In the case of Hispanics in the United States, it is also argued that the immigration experience creates strong links to traditional family-oriented cultures in the countries of origin (Bean and Tienda Reference Bean and Tienda1987). As Arjouch (Reference Ajrouch2005) reports in her study on Arab-American older persons, many immigrated before experiencing the burden of caring for their own older parents, which encouraged attitudes closer to traditional norms than those moderated by lived experiences.
Burr and Mutchler (Reference Burr and Mutchler1999) suggested that ethnic beliefs and attitudes differed depending on the specific norm. They showed that older Blacks and Hispanics were more likely than non-Hispanic Whites to agree that parents and adult children should co-reside if necessary. At the same time, few group differences were found when attitudes towards financial aid were considered. Rossi and Rossi (Reference Rossi and Rossi1990) found that Blacks reported less filial obligation than Whites to primary kin including parents, but showed greater commitment to norms of filial obligation towards distant kin. In The Netherlands, we are not aware of any previous quantitative studies on ethnic differences in filial responsibility expectations. Qualitative research among some immigrant groups (Niekerk Reference Niekerk1991; Yerden Reference Yerden2000) has suggested, however, that levels of filial obligations among immigrants are higher than among the native Dutch.
The backgrounds of immigrants in The Netherlands
Large-scale immigration to The Netherlands started in the 1960s, since when there have been distinct successive flows. The earliest immigrants were from former Dutch colonies, like Suriname and the Antilles in the Caribbean. Many older people in The Netherlands of Surinamese and Antillean origin either came for higher education or immigrated to join their adult children already resident in the country (Schellingerhout Reference Schellingerhout2004a; Vermeulen and Penninx Reference Vermeulen and Pennix2000). Many Surinamese and Antillean immigrants were familiar with Dutch society and had some command of the language. The second large flow, as in many other western European countries, was of (predominantly male) unskilled labour migrants from southern Europe and the Mediterranean (particularly Turkey and Morocco). Many of their families (wife and children) came to The Netherlands during the 1980s and 1990s (Vermeulen and Penninx Reference Vermeulen and Pennix2000; De Valk et al. Reference De Valk, Liefbroer, Esveldt and Henkens2004). The third migration phase, of asylum seekers from very many countries, became substantial during the 1990s and comprised mainly young people (and therefore does not concern this paper).
Among the Turks, Moroccans, Antilleans and Surinamese, a substantial number are now reaching old age. In 2005, around 12 per cent of the 1.1 million persons with a Turkish, Moroccan, Antillean, or Surinamese background were aged 50 or more years, but few are yet over 65 years-of-age (Schellingerhout Reference Schellingerhout2004a; Statline 2005). Recent studies of elderly immigrants and their use of welfare and health-care provision have established that they have lower socio-economic status than native Dutch older people – the Turkish and Moroccan have the lowest education and income (Ament and Lautenbach Reference Ament and Lautenbach2002; Nationaal Instituut voor Budgetvoorlichting 2004). This is no surprise if one takes into account their employment history, for many had low-status, unskilled jobs in The Netherlands, and many were unable to work for extended periods through enforced unemployment and sickness, and became dependent on social welfare. Almost one-half of Turkish and Moroccan men aged between 55 and 64 years claim to be unable to work because of health problems (Schellingerhout Reference Schellingerhout2004a). The socio-economic position of the elderly Surinamese is closer to the Dutch; on average they had had more education and have higher mean incomes than the Turks or Moroccans. The socio-economic position of Antilleans lies in between the Surinamese and the Turks or Moroccans (Vermeulen and Penninx Reference Vermeulen and Pennix2000). Rates of employment among both men and women of Surinamese and Antillean origin are higher than in the other two immigrant groups. Although immigrant older people make less use of the formal care provided by the Dutch welfare state, those from Suriname and the Antilles are more aware of their entitlements than other groups, which has been attributed to their greater knowledge of Dutch society and the language (Dagevos, Gijsberts and van Praag Reference Dagevos, Gijsberts and Van Praag2003). The Turks and Moroccans are least knowledgeable about formal welfare provision (Schellingerhout Reference Schellingerhout2004b).
Hypotheses
Ethnicity and family relations
Theories of differences in ‘family systems’ suggest that family relations and the related expectations reflect the importance attached to kinship in a society. Several authors have argued that in more collectivistic societies, kinship ties take centre stage (Todd Reference Todd1985; Inglehart and Baker Reference Inglehart and Baker2000; Kagitcibasi Reference Kagitcibasi1996; Nauck Reference Nauck, Trommsdorff and Kornadt2007; Reher Reference Reher1998). Kagitcibasi (Reference Kagitcibasi1996) referred to these societies as cultures of relatedness. The description applies to Turkey and Morocco, especially in the more remote rural areas. In Suriname and the Antilles also, the family is of great importance and intergenerational ties are strong. It can be assumed that many immigrant older people in The Netherlands grew up in kinship-oriented societies, where intergenerational interdependence was a prerequisite for a family's material wellbeing. In such societies, children contribute to the family wellbeing both while young (as by working in the fields and contributing to the family economy), and when adult by providing old-age support and financial security for their parents.
As in many western countries, Dutch society is characterised by individualism and the independence or autonomy of parent and child. According to Kagitcibasi (Reference Kagitcibasi1996), family relations in The Netherlands exemplify the culture of separateness. In these societies, support is mainly provided by the (welfare) state although the emotional bonds between parents and children are generally strong. Some Dutch older people even experience support from their children as a form of control (Komter and Vollebergh Reference Komter and Vollebergh2002). We therefore hypothesised that immigrant elders would agree more strongly than Dutch elders that children should support their parents (H1).
Acculturation theories suggest that over time immigrants adjust their perceptions and orientations to the cultural patterns of the country of residence (Alba and Nee Reference Alba and Nee1997). This does not imply that norms and values from the country of origin are totally abandoned but rather that the immigration experience prompts their revision (Kagitcibasi Reference Kagitcibasi1996). Living in the host society for a long period increases exposure to new values. Besides the effect of duration of residence in the new society, orientation (or receptiveness) to change is influential (Berry Reference Berry and Padilla1980). Acculturation studies have shown the importance of language proficiency as an indicator of orientation to the host culture (Van Tubergen Reference Van Tubergen2006; Chiswick and Miller Reference Chiswick and Miller2002). Based on these notions, we expect that immigrant elders who had been resident in The Netherlands for a longer period, and those who are proficient in the Dutch language, will be less of the opinion that children should support their parents, than those with short residence and little language proficiency (H2).
Socio-demographic characteristics and the need for support
Previous research has shown that variations in socio-demographic attributes and wellbeing account for different filial norms. In order to distinguish between the ethnic influence and other factors, it is important to control for socio-demographic position and the need for support (Lee, Netzer and Coward Reference Lee, Netzer and Coward1994; Lee, Peak and Coward Reference Lee and Aytac1998). Educational attainment, marital status, and having children influence opinions and perceptions regarding the family (Kalmijn Reference Kalmijn2004; Roschelle Reference Roschelle1997; Rossi and Rossi Reference Rossi and Rossi1990; Spitze and Logan Reference Spitze and Logan1990). The highly educated can more readily purchase private care, which might reduce their sense of filial obligations. Co-resident married people can provide care to each other rather than depend on others, which again might reduce the expectation of filial obligation. Conversely, those who have children have the option to depend on them to provide help and care. This reasoning leads to the hypothesis that those who are (a) married, (b) childless or (c) highly educated will be less of the opinion that children should support parents than the unmarried, those with children and the less educated (H3a–c).
We have also studied the impact of current physical health on perceptions of filial obligations (De Jong Gierveld and Van Tilburg Reference De Jong Gierveld and Van Tilburg1999; Gierveld Reference Gierveld2003). It can be assumed that the current need for support is related to a person's views about filial obligation. A current need may raise expectations of filial obligation, whereas those without care and support needs may not be aware of the importance of children helping out. We expect that older people who have physical health problems will be more of the opinion that children should support their parents than those who do not (H4). All hypotheses have been examined after controlling for gender, age and level of urbanisation.
Data and measures
The data for the analysis are from the first round of the Netherlands Kinship Panel Study (NKPS) (Dykstra et al. Reference Dykstra, Kalmijn, Knijn, Komter, Liefbroer and Mulder2005). The main sample (Dutch respondents) and the migrant sample (including Turkish, Moroccan, Surinamese, and Antillean respondents) were used. The main sample is a nationally representative sample of about 8,000 Dutch respondents. The migrant sample was drawn from 13 Dutch cities in which 50 per cent of the migrants from the four main ethnic groups live. It includes 1,400 migrants with a Turkish, Moroccan, Surinamese or Antillean origin. The topics covered in the main and the migrant questionnaires were similar, and provided comparable data. The respondents were interviewed in their homes, in most cases by an interviewer of the same ethnic background. All interviews followed a structured questionnaire in Dutch that was also available in Turkish or Arabic. The response rate among the migrants was in the same range as that of the Dutch, from 41 per cent for the Surinamese, to 52 per cent for the Turks. The present analysis compares immigrants with Dutch respondents living in the 13 cities in which the immigrants were sampled. The sample is confined to the respondents aged 50–80 years. After the exclusions, the sample comprised 70 Turkish, 73 Moroccan, 125 Surinamese, 59 Antillean, and 469 Dutch respondents.
Dependent variables
The first analysis compares the levels of agreement with three views about filial obligation among the Dutch and immigrant elders. The respondents were asked whether they agreed with the statements that: ‘children who live nearby should visit their parents at least once a week’, ‘children should care for their sick parents’, and ‘if parents are old, children should provide co-residence for them’. The answers were requested on a five-point Likert scale, from ‘1’ for ‘fully agree’ (group oriented) to ‘5’ for ‘fully disagree’ (individualistic). A later multivariate analysis tries to disentangle immigrant-group specific effects on perceived filial obligations. The average scores of agreement on the three statements were calculated, and a continuous scale constructed. Higher scores on the scale indicate a more individualistic orientation.Footnote 1
Independent variables
Immigrant group
The ethnic background of the respondents was defined according to country of birth and that of the respondent's parents. Those born abroad or with at least one parent born abroad were assigned to one of the four ethnic minority or immigrant groups. For each group, a separate dummy variable was created to compare them to the Dutch.
Physical health
Respondents were asked to rate their general physical health on a five-point scale that ranged from ‘1’ for ‘very well’ to ‘5’ for ‘very bad’. This variable was entered into the analyses as a continuous variable.
Educational level
The educational level of the respondent was measured as the highest educational level to which the respondent had been enrolled (with or without completion or accreditation) and three levels were distinguished, from low ‘1’ to high ‘3’.
Marital status
Respondents who are married at the time of the interview (coded ‘1’) were compared with those who were divorced, widowed or never married (coded ‘0’).
Children
Respondents who had one or more surviving children (own or adopted) (coded ‘1’) were compared with those without children (coded ‘0’). Given that childlessness is much less common among the older cohorts of immigrants than their Dutch counterparts, a measure was designed of the effects of acculturation on the perceptions of immigrant elders. The dichotomy distinguishes those who had no more than the average number of children for their group of origin (‘0’) from those who had more than the average (‘1’).
Gender
A dichotomous variable (men ‘0’, women ‘1’).
Age
The respondents' age (in years) at the time of the interview.
Urbanisation
Respondents who lived in one of the four major cities in The Netherlands (Amsterdam, Rotterdam, The Hague and Utrecht) were coded ‘1’ and compared with those who lived elsewhere in The Netherlands (coded ‘0’). In the second part of the analyses, on the immigrant-group specific effects on perceived filial obligations, two additional variables were included.
Years of residence in The Netherlands
Respondents with an immigrant background were asked to report the year in which they migrated to The Netherlands, from which the duration of residence was calculated and used as a continuous variable in the analyses.
Ability in the Dutch language
Fluency in the Dutch language was assessed by the language in which the interview was conducted. Respondents for whom the interview was primarily conducted in Dutch (coded ‘1’) were compared to those who had limited or no Dutch language command and were interviewed in their mother tongue (coded ‘0’).
Methods
Bivariate analyses provided the first insight into the variations in the respondents' perceptions of filial obligations. The replies to each of the three statements about filial obligations were analysed separately, and differences between ethnic groups tested with Fisher's least significant difference (LSD) post hoc test (Fisher Reference Fisher1966; Prochan Reference Prochan1997). Secondly, ordinal logistic regression was used to study the effects of ethnic origin, socio-demographic variables and need for support on perceived filial obligations. Finally, ordinary least-squares (OLS) regression was used to study the effects of acculturation on the perceptions of the respondents with Mediterranean (Turkish and Moroccan) and Caribbean (Surinamese and Antillean) origins.
Results
Perceptions of filial obligations: an impression
The socio-demographic profiles of the respondents are presented in Table 1, and Figure 1 summarises the level of agreement with the three statements of filial obligation. It is clear that around one-half of the Turkish and Moroccan elders agreed with the statement that children who live nearby should visit their parents at least once a week, that 36 per cent of the Surinamese agreed with this statement, as did 16 per cent of the Antilleans and the Dutch. By contrast, three to four per cent of Turks and Moroccans did not agree with the statement as against 26 per cent of the Dutch. Testing these group differences shows that the opinions of the Turks and Moroccans differed significantly from the other three groups but not from each other. The Surinamese and Antilleans did not differ from each other but their opinions differed from all other groups. Finally, the Dutch agreed least with this statement and their views differed from each of the immigrant groups.
Figure 1. Level of agreement regarding three types of filial obligation by (immigrant) group of origin.
Notes: T: Turkish, M: Moroccan, S: Surinamese, A: Antillean and D: Dutch. Source: Netherlands Kinship Panel Study 2002–03 (main and migrant samples).
Table 1. Socio-demographic profile of the respondents by ethnic group
Notes: SD: standard deviation. 1. 1=‘good’ to 5=‘bad’. 2. 1=married. 3. 1=yes. 4. 1=‘low’ to 3=‘high’. 5. 1=G4. 6. Dichotomy for above average number of children.
Source: Netherlands Kinship Panel Study 2002–03 (main and migrant sample).
The levels of agreement with the statement that children ‘should care for sick parents’ showed similar patterns of responses. More than one-half of the Turks (57%) and Moroccans (51%) agreed, compared to 11 per cent of the Dutch. Again the Surinamese and Antilleans took an intermediate position: 27 and 19 per cent respectively did not think this was the children's duty. The results of an analysis of variance showed that three groups could be distinguished by the level of agreement: the Turks and Moroccans, the Surinamese and Antilleans, and the native Dutch differed from each other. Finally, 44 per cent of the Moroccans and 34 per cent of the Turks held the opinion that children should co-reside with elderly parents, but only 16 per cent of the Surinamese had this opinion, and only three per cent of the Dutch. The LSD comparisons showed that, except for the Surinamese and Antilleans, all the groups’ level of agreement differed significantly from one another. The Moroccans agreed most with the statement, followed by the Turks, the Surinamese and Antilleans, and lastly the Dutch.
Filial obligations: variations by ethnic groups and socio-demographic factors
Table 2 shows the results of the ordinal logistic regressions for each of the three dependent variables. A larger proportion of the Turks, Moroccans and Surinamese than the Dutch believed that children should visit their parents at least once a week. Furthermore, it was shown that level of education influenced the level of agreement with the statement. In line with our hypothesis, the more highly educated were less in agreement, but controlling for educational attainment did not explain the differences between the immigrant groups and the Dutch. Contrary to our expectation (H3), none of the other socio-demographic characteristics were significantly associated with the level of agreement that children have a duty to visit parents.
Table 2. Level of acceptance of three statements of filial obligation by ethnic group
Notes: B: Ordinal logistic regression coefficients; negative coefficients indicate stronger agreement. s.e.: standard error.
Significance levels: * p<0.05, ** p<0.01, *** p<0.001.
Source: Netherlands Kinship Panel Study 2002–03 (main and migrant samples).
With regard to the obligation to provide care to older people, the analyses show that all the immigrant groups (the Turks and Moroccans most obviously) more strongly agreed that children should care for ill parents than the Dutch, in line with the first hypothesis. Again educational level was associated with the level of agreement, with the highly educated holding a more liberal view. Contrary to hypotheses H3 and H4, however, none of the other socio-demographic characteristics nor need-for-support related to upholding this filial obligation. Of the control variables, only gender had an effect, with women being less of the opinion that children should care than men.
For the third filial obligation statement, on providing co-residence to parents in old age, there were again differences between Dutch and immigrant older people. The immigrants more often agreed with the view than the Dutch, and educational level again had a liberalising effect. Although the other socio-demographic and support factors were not of importance, all three control variables were. Greater age and being a woman were associated with less agreement with this filial obligation. Those living in the largest cities more often agreed with the statement than those living elsewhere. The effects of self-reported overall wellbeing and disabilities were examined in separate analyses, but neither factor was associated with the level of agreement with the three filial obligations (analysis not shown). We also tested whether the effects of the socio-demographic factors and need-for-support had similar strengths among the immigrant groups and the Dutch by examining the interaction terms separately for each group. There was no indication that the variables operated differently for particular migrant groups or for the Dutch.
Filial obligations: acculturation
The second part of the analyses examined the effects of acculturation, more specifically the duration of residence in The Netherlands and variations in Dutch-language ability. Given the similarity in the levels of agreement with the three filial obligation statements and the small samples, two groups of immigrants were studied: those with Mediterranean origins (Turks and Moroccans), and those with Caribbean origins (Surinamese and Antilleans). The results of the OLS regression, which controlled for background characteristics, are shown in Table 3. Although we expected that a longer period of residence would change perceptions of filial obligations (H2), no such effect was found. Incidentally, those whose period of residence was unknown were less of the opinion that children should support elderly parents. The second indicator of acculturation, Dutch-language ability, was significantly associated with the perception of filial obligations, but only among those of Mediterranean origin. Thus, among those from the Caribbean, level of urbanisation was a significant influence, with those living in the larger cities much more assertive of filial obligation. Finally, it should be noted that those from the Mediterranean with an above-average number of children more strongly agreed with the statements than those with smaller families.
Table 3. Regression coefficients for filial obligations by immigrant origin
Source: Netherlands Kinship Panel Study 2002–03 (main and migrant samples).
Significance levels: * p<0.05, ** p<0.01, *** p<0.001.
Discussion
The first aim of the paper was to determine the relative importance of ethnic origin and other socio-demographic background characteristics on the strength of older people's normative views about filial obligations. It was found, in line with hypothesis H1, that ethnic background was the principal influence. Turks and Moroccans most strongly adhered to the opinion that children have obligations towards their aged parents. Whereas all immigrant groups expressed stronger filial obligations than the Dutch, the extent of the differences varied with the statement or aspect, and were greatest with reference to the prescription that elderly parents should be able to share an adult child's home. These findings suggest that the respondents' views about filial obligations were rooted in the opinions and norms into which they were socialised. Opinions on intergenerational relations and support are core values that, contrary to more practical domains of life, are not easily adjusted in a different society. Today's older immigrants in The Netherlands are the ‘first generation’, and most grew up in their countries of origin. We can expect that in the second generation, more will have been socialised in The Netherlands, and will hold views more similar to those of the Dutch.
There was only limited support for the hypotheses of the influence of socio-demographic background and the need for support (H3 and H4). Only educational attainment had clear and consistent effects on normative views. The more highly educated were less of the opinion that children have an obligation to care for elderly parents. This finding suggests that, as the educational level of immigrants rises, their opinions and perceptions change and perhaps diversify. Women agreed less than men that children have caring obligations, including caring for ill parents and providing co-residence for older parents. Separate analyses for the individual ethnic groups, however, did not find gender differences among the immigrant groups. It appears that the effect was principally because Dutch women agreed less with the norms than men, (cf. Schans and Komter (Reference Schans and Komter2006). Finally, those living in the largest cities most strongly supported the norm of co-residence with children, which might be associated with big-city housing costs and shortages or the unfavourable reputation of local nursing homes.
The second aim of our study was to assess the role of acculturation on filial obligation norms. It was hypothesised (H2) that longer residence in The Netherlands would weaken the adherence to filial norms, but no such effects were found among the immigrant groups. On the other hand, in accordance with hypothesis H2, the Mediterranean respondents who were fluent in the Dutch language had relatively weak agreement with the filial obligation norms. Among the Caribbean-origin respondents, this effect was not found, possibly because of their high level of language competence, a legacy of their countries' colonial past. It seemed that for the Caribbeans, Dutch-language ability was not an indicator of acculturation, although living in the largest cities was associated with relatively strong agreement with traditional filial obligation norms, which may be interpreted as an acculturation and selection effect. The more acculturated Caribbean middle class tend to leave the largest cities for the suburbs, and therefore to live in neighbourhoods dominated by the native Dutch. The effects of acculturation are probably better measured by studying contacts in the host society rather than the duration of residence. Furthermore, we can conclude that acculturation processes vary by the immigrant group. Future studies should be aware of the diversity of immigrant groups and take the different acculturation pathways into account.
Although this study has provided new insights into the filial norms of immigrant older people, some limitations should be noted. First, it has examined exclusively the perceptions of the children's obligations. It is as yet unclear whether the actual behaviour differs as much by ethnic group (Pyke Reference Pyke1999). Children's actual support may depend on attributes of the elderly parent and of the child as well as the availability of formal, subsidised (or welfare state) support. Conflicts may arise between the support expectations and the actual conditions, as reported by Ajrouch (Reference Ajrouch2005) among Arab-American older people in Detroit, Michigan. Future research should further explore the link between filial responsibility norms and actualities. Second, the causal relationship between the filial obligation norms and the current situation of the older person is difficult to disentangle (Gans and Silverstein Reference Gans and Silverstein2006). Cognitive dissonance theory (Finley, Roberts and Banahan Reference Finley, Roberts and Banahan1988) suggests that filial expectations may be adjusted in an attempt to reconcile the gap between the ideal and revealed practice. More understanding of these causal relationships requires longitudinal data that traces the ways in which filial obligations change. Third, it should be remembered that, as Pyke (Reference Pyke1999) has shown, asserting a norm of filial responsibility does not itself empower aged parents in their intergenerational relations nor provide an entitlement to care from their children, at least not without a loss of regard or power. The associations between filial responsibilities, the quality of intergenerational relations and within-family power dynamics were beyond this article's scope. More research is needed to develop understanding of the interplay of these factors among immigrant families. Finally, this study has only examined the older generations' views, and future work should examine those of the children and grandchildren.
To summarise, given the growing number of older migrants in western societies, we strongly argue that minority and immigrant families should be included in large surveys as a matter of standard practice. Their inclusion will enable comparison of family ideals, attitudes and practices across diverse groups and generations. Filial obligations are just one aspect of intergenerational support systems that, in turn in welfare-states like The Netherlands, are just one aspect of old-age security. The interplay between filial responsibilities and formal assistance might be of special importance for immigrant families and should be taken into account by policymakers and family sociology scholars alike.
Acknowledgements
We thank Jenny Gierveld and two anonymous reviewers for their thoughtful comments and stimulating suggestions. An earlier version of this paper was presented at the European Population Conference 2006, Liverpool June 2006. The Netherlands Kinship Panel Study is funded by the Major Investment Fund of The Netherlands Organization for Scientific Research (grant 480-10-009), and by The Netherlands Interdisciplinary Demographic Institute, Utrecht University, the University of Amsterdam and Tilburg University.
