Introduction
The long-standing gerontological interest in family care-giving for older adults started with a primary focus on tasks or behaviours: what does the care-giver do? Only more recently has attention turned to filial responsibility, or the attitudes expressed by adult child care-givers regarding parent support. Part of the interest in attitudes of care-giving, also referred to as care-giving responsibility, derives from the assumption that internalisation of care-giving norms within individual attitudes will lead to actual care-giving behaviours. For example, in traditional Asian culture, children are explicitly taught they should care for their parents during old age, as part of the cultural norm of xiao or filial piety. As such, Zhan observes that xiao ‘has long been held responsible for the practice of familial caregiving for older parents’ (2004: 123); indeed, that filial responsibility attitudes are the foundation for ‘the whole system of familial eldercare in China’ (2004: 127). In Western society the social norm is not as explicit, but there is nevertheless a general social expectation that children will respect, love, and provide care for their parents (Dai and Dimond Reference Dai and Dimond1998). Some, such as Silverstein and Bengston (Reference Silverstein and Bengston1997), criticise the strength of individual attitudes as inadequate in Western contexts, resulting in an undermining of the family care system.
However, over three decades of gerontological research on care-giving behaviours (Chappell Reference Chappell1993; Montgomery, Borgatta and Borgatta Reference Montgomery, Borgatta, Borgatta, Liu and Kendig2000) has demonstrated that in virtually all countries, the majority of children care for their elderly parents when in need, despite differences in the strength of cultural norms regarding the filial provision of care. Much less research has demonstrated that personal attitudes of filial responsibility are necessary in order for children to provide this care (Chen, Bond and Tang Reference Chen, Bond and Tang2007). Existing research is contradictory, with some reporting a relationship between care attitudes and behaviours (Zhan Reference Zhan2004) and some not (Ho Reference Ho and Bond1996; Kauh Reference Kauh1999; Yu Reference Yu1983). Others report structural factors (such as co-residence) as stronger predictors of care provision than attitudes (Keefe Rosenthal and Beland Reference Keefe, Rosenthal and Beland2000). This raises the question of the importance of culture, as manifested within individual attitudes of filial responsibility, for the provision of care to older adults. This paper reports the results of an empirical examination of the relationship between attitudes of filial responsibility and care-giving behaviours to parents among three cultural groups of filial care-givers: Caucasian-Canadians, Chinese-Canadians and Hong Kong-Chinese. Caucasian-Canadian and Hong Kong-Chinese samples were included because cultural norms of filial responsibility for these groups have long been viewed as different, in their strength and/or nature. The Chinese-Canadian sample was included because they are a diasporic group (i.e. exposed to both sets of cultural norms). These three cultural groups, potentially influenced by differing cultural norms, may have differing attitudes of filial responsibility. The purpose of this paper is to examine the relative role of attitudes of filial responsibility in the provision of parental care when controlling for a variety of other cultural and structural factors.
Review of the literature
While the term ‘filial responsibility’ lacks clear definition, it is often conceptualised as a norm or cultural schema (Holroyd Reference Holroyd2001); its translation into individual attitudes (i.e. ‘learned predispositions to respond in a favourable or unfavourable manner to a particular person, behaviour, belief or thing’ (Feldman, Collins and Green Reference Feldman, Collins and Green2001: 513), including various emotions towards the object, beliefs and thoughts about it, and intentions or predispositions to act in a certain manner regarding it) is far from straightforward (Stein et al. Reference Stein, Wemmerus, Ward, Gaines, Freeberg and Jewell1998). Much has been written about the closely related concept of filial piety in traditional Chinese culture, and recently about whether and how these attitudes and norms are changing within modern-day China (Cheung and Kwan Reference Cheung and Kwan2009a, Reference Cheung and Kwan2009b; Lee and Hong-Kin Reference Lee and Hong-Kin2005b). Historically, children, especially first-born sons, were strongly obligated to provide for their ageing parents. Daughters were ‘transferred’ to their husband's family and expected to assist in the care of his parents, not their own (Liu and Kendig Reference Liu and Kendig2000). While the notion of filial piety was officially attacked as feudal during the rapid and tumultuous political, social and cultural change beginning in the mid-20th century under the leadership of Chairman Mao (Ikels Reference Ikels2004; Miller Reference Miller and Ikels2004), in the late 20th century it was re-embraced as a virtue. Women are now exhorted to support their own parents out of gratitude rather than purely obligation (Wang Reference Wang and Ikels2004).
Debate remains, however, over whether and to what extent the cultural norm of filial piety is changing in Asia. Some believe it remains strong (Whyte Reference Whyte and Ikels2004; Wong and Chau Reference Wong and Chau2006) and some research finds strong adherence among young adults (Yue and Ng Reference Yue and Ng1999). Sung (Reference Sung2001), however, suggests that the norm is now generally perceived as more reciprocal, affection-based, and voluntary in nature. Research by Ho, Hong and Chiu (Reference Ho, Hong and Chiu1989), Chuang and Yang (Reference Chuang, Yang, Yang and Hwang1990) and Ikels (Reference Ikels2004) supports the view that the strength of filial piety attitudes is fading among younger generations. Indeed, traditional commitment to filial obligation may co-exist alongside the value of independent decision-making, as contradictory internalised discourses (Traphagen Reference Traphagen2008).
The discursive ideal of filial piety and obligation espoused in Chinese culture contrasts with the ideology of individualism and a focus on individual autonomy in the West that is perhaps better known for ageism than reverence of older adults (Liu Reference Liu, Liu and Kendig2000). In Western society it has been argued that there is no normative consensus about family obligations, individuals have only ‘a vague mental awareness’ of filial responsibility often activated more fully after a triggering event (Fry Reference Fry, Binstock and George1996; Donorfio Reference Donorfio1996). The individualisation thesis (Beck and Beck-Gernsheim Reference Beck and Beck-Gernsheim2001) posits a Western tendency to eschew the concept of obligation in favour of affection, as the primary frame for meaning within family relationships. Obligations are viewed as conditional, varying depending on personal circumstances and perceived parental deservedness.
Nevertheless, North American research suggests that adult children tend to feel some form of filial obligation, irrespective of socioeconomic background or generation (Dellman-Jenkins and Brittain Reference Dellmann-Jenkins and Brittain2003). For instance, adult children may feel responsible to provide emotional support and discuss important matters such as resources with parents (Hamon and Blieszner Reference Hamon and Blieszner1990). Others report a strong internalised moral obligation to provide care, sometimes qualified for parents who are more aggressive or incontinent (Wolfson et al. Reference Wolfson, Handfield-Jones, Cranley-Glas, McClaran and Keyserlingk1993), and simultaneously embracing the idea of personal choice and parental autonomy (Funk Reference Funk2008). Lastly, critical and feminist researchers point to the strong emphasis on familialism and family responsibility in existing rhetoric and discourse (Armstrong and Kits Reference Armstrong and Kits2001).
Thus, despite some apparent differences at the normative (historical, cultural, socio-political) level, between North America and China, there is much uncertainty about whether these broader norms are echoed at the personal (attitudinal) level in contemporary times and whether these attitudes are related to care provision to parents (behaviours). In China, there is evidence that actual care provision frequently begins before there is parental health need, as a demonstration of respect due to age. While decreasing numbers of children co-reside with parents (e.g. Chen Reference Chen1998; Whyte Reference Whyte and Ikels2004; Zhang Reference Zhang and Ikels2004), they still do so more frequently than in the West. The roles of older adult spouses as well as daughters in care provision also appear to be increasing, while daughters-in-law are providing less care to their parents-in-law. Sons continue to provide much support (Chappell Reference Chappell2003; Lee and Hong-Kin Reference Lee and Hong-Kin2005a; Whyte Reference Whyte and Ikels2004) but daughters provide more hands-on and personal care than sons, and many older adults now reside with a daughter rather than a son (Zhang Reference Zhang and Ikels2004). Financial support to parents is not uncommon (Lee and Hong-Kin Reference Lee and Hong-Kin2005a, Reference Lee and Hong-Kin2005b). In sum, in mainland China and Hong Kong, adult children continue to care for their ageing parents in ways that reflect historic patterns but also suggest new emerging patterns. They do so within a context in which formal service support is very limited.
In North America, despite relatively less explicit moral rhetoric admonishing filial responsibility (Lee and Sung Reference Lee and Sung1997), research since the 1970s has demonstrated that families do provide care for their older members (Chappell Reference Chappell2003; Pyke Reference Pyke2000). Spouses are the primary source of care, followed by daughters (Montgomery, Borgatta and Borgatta Reference Montgomery, Borgatta, Borgatta, Liu and Kendig2000). Sons tend to provide care when daughters are not available, or offer primarily monetary assistance and/or supervision (Frederick and Fast Reference Frederick and Fast1999). The diasporic Chinese in North America experience a transnational identity, integrating aspects of both Chinese and Western cultures (Van Ziegert Reference Van Ziegert2002). Among Chinese-Canadian older adults, traditional Chinese practices are evident in a strong pattern of co-residence with both adult children and spouses (Chappell and Kusch Reference Chappell and Kusch2007). However, joint living arrangements are less prominent than in China. The son and daughter-in-law unit tends to be more involved in care than might be expected of Western children (Chappell and Kusch Reference Chappell and Kusch2007), while both daughters and spouses provide more care than might be prescribed within the traditional Chinese notion of filial piety. Chinese-Canadian families, that is, appear to adopt care patterns that reflect a blending of those found in China and in the West.
Research examining both filial attitudes and care-giving behaviours of adult child care-givers is inconsistent when non-Western groups are examined. Some report that traditional attitudes are related to care provision (Chen, Bond and Tang Reference Chen, Bond and Tang2007; Kwok Reference Kwok2006; Zhan Reference Zhan2004). However, in Korea, respect for elders may be more symbolic than real; children express obligations to care for their parents, but often do not enact them (Kauh Reference Kauh1999). Some research suggests the equal or greater role of structural factors compared with cultural ones. For example, Zhan and Montgomery (Reference Zhan and Montgomery2003) reveal a dual influence of cultural norms and family resources/structural factors (availability of siblings, employment status, receipt of pension) on the provision of care in China. Such findings parallel research on family care-giving among other non-Western cultural groups (e.g. African Americans, Hispanics, American Indians, and Asians/Pacific Islanders) that acknowledge the role of culturally rooted attitudes of filial responsibility and the equal if not greater role of structural and socio-economic factors such as family and household structure, and poverty (Mangum et al. Reference Mangum, Garcia, Kosberg, Mulllins and Barresi1994; Peek, Coward and Peek Reference Peek, Coward and Peek2000).
Turning to research examining both filial attitudes and behaviours in Western contexts, Lowenstein and Daatland (Reference Lowenstein and Daatland2006) report that filial obligation is only moderately related to support provision in five European countries. Yet Stein et al. (Reference Stein, Wemmerus, Ward, Gaines, Freeberg and Jewell1998), Piercy (Reference Piercy1998), and Klein Ikkink, van Tilburg and Knipscheer (Reference Klein Ikkink, van Tilburg and Knipscheer1999) all find that attitudes of filial responsibility are related to greater filial support. However, Hamon and Blieszner (Reference Hamon and Blieszner1990) point out that despite an established gender difference in care provision (e.g. daughters provide more assistance, with more intensive and less discretionary tasks), most attitudes of filial obligation do not vary by gender. In other research, the enactment of filial norms into behaviour is stronger among daughters than sons (Silverstein, Gans and Yang Reference Silverstein, Gans and Yang2006). That is, links between attitudes and behaviours are not unequivocal.
Further, among employed adult children, structural factors, particularly gender and co-residence, are stronger predictors of parental care than filial obligation (Keefe, Rosenthal and Beland Reference Keefe, Rosenthal and Beland2000). Other structural correlates of care provision from past care-giving literature, not necessarily focusing on adult children, include: lower socio-economic status, being married, and being older (Keefe, Rosenthal and Beland Reference Keefe, Rosenthal and Beland2000).
The purpose of this paper is to investigate associations between attitudes toward care provision to parents (filial responsibility) and care-giving behaviours among three groups of care-givers: Caucasian-Canadians, Chinese-Canadians and Chinese in Hong Kong. Are personal attitudes supportive of filial responsibility associated with filial care-giving behaviours and if so, of which types of care provision (emotional support, companionship, assistance with basic (ADL) or instrumental (IADL) activities of daily living, or with finances)? Further, is cultural group related to care-giving behaviours and if it is, does the relationship between attitudes and behaviours vary between cultural groups? The literature reporting a relationship between filial responsibility and care-giving behaviours would suggest that this relationship will be strongest among Hong Kong-Chinese and weakest among Caucasian-Canadians among the three groups studied here; Chinese-Canadians would fall between. However, others suggest that other factors such as co-residency, ill health and socio-economic status will be greater correlates of care-giving behaviour than attitudes, and that this will be found for each of the cultural groups.
Methods
Data were collected in the autumn/winter of 2007 in Hong Kong, and in the winter/summer of 2008 in Canada (Victoria/Vancouver, British Columbia). Face-to-face interviews were conducted with filial care-givers, using measures validated in both English and Chinese. The interview schedule was constructed in English then translated and back translated into/from Chinese. Interviews were approximately one and a half hours in length on average; those with Caucasian-Canadians and Chinese-Canadians were conducted in English (all Chinese-Canadians chose English, rather than Mandarin or Cantonese). Those in Hong Kong were conducted in Cantonese.
Eligibility criteria included: geographic proximity (access by car) and providing care to a parent age 60+ for three or more hours per week. Care or support included, for example: errands, phone calls, attending appointments, linking parents to formal services, cooking, cleaning, and emotional support.
The Hong Kong sample was selected from a random sample of those aged 60+, originally obtained from a list of households in the region representing land-based, non-institutionalised elderly residents. All in the original study who said they received care from an adult child were approached for this study until sufficient numbers were obtained. Both Caucasian- and Chinese-Canadians were convenience samples recruited through advertising in local papers and newsletters, announcements at meetings of local family care-giver networks and multicultural societies, and through word of mouth. These samples are not representative; the findings focus only on relationships between variables.
A brief telephone interview explained the study and screened for eligibility. Caucasian was defined as white or light skinned (ancestrally: Canadian, British, American, Scandinavian, Icelandic, and European). Chinese-Canadians had to self-identify as Chinese and, to ensure comparability with the Chinese sample from Hong Kong, had to be descended from Hong Kong people. Hong Kong respondents had to self-identify as Chinese.
The primary variables of interest are attitudes of filial responsibility and care-giving behaviours. Attitudes of filial responsibility were measured using two scales. The Filial Expectancy Scale (Kim and Lee Reference Kim and Lee2003; Stein et al. Reference Stein, Wemmerus, Ward, Gaines, Freeberg and Jewell1998) is used extensively in North American and Asian samples (alpha=0.69–0.80; for this study, alpha=0.81). The scale consists of five items such as ‘Children should live close to their parents’. Liu and Kendig's (Reference Liu and Kendig2000) six-item measure of filial piety includes items such as: ‘How much would you agree or disagree that, when it comes to elderly parents adult Chinese have the obligation to look after them?’ or ‘assist them financially if needed?’ Alphas are strong (0.81–0.88) among two generations of Chinese immigrants in New Zealand; in this study, alpha=0.83.
Five different types of care-giving behaviours were measured as outcome variables. Assistance with ADL and IADL were measured by asking whether the care-giver provided help with a list of these activities, which were summed in a count of the number of activities the adult child assisted with and analysed separately (ADL: bathing, dressing, using the toilet, feeding, getting in/out of bed, getting about the house; IADL: shopping, meal preparation, light housework, heavy housework, transportation, finances, negotiating service systems). Alphas: ADL=0.90; IADL=0.92. Respondents were asked whether they provided emotional support (‘How well do you feel that you meet your [mother/father]'s needs for emotional support, on a scale where 0 represents “not at all” and 4 represents “very much so”?’), companionship (‘How well do you feel that you meet your [mother/father]'s needs for companionship and visiting?’) and financial support (‘Do you provide support for your [mother/father] financially – giving them money if they needed it?’).
Cultural group (Caucasian-Canadian, Chinese-Canadian, and Hong Kong-Chinese) was another primary independent variable of interest in this study, and was entered into the multivariate regression as a dummy variable, with the Chinese-Canadian group as the reference category. In addition, among diasporic Chinese-Canadians only, the Asian American Multidimensional Acculturation Scale (AAMAS) consisting of four domains (Chung, Kim and Abreu Reference Chung, Kim and Abreu2004) was administered. The Food Consumption Factor for that scale consists of two questions (e.g. ‘How often do you actually eat Chinese food?’); alpha=0.87. The Cultural Knowledge Factor consists of three questions (e.g. ‘How knowledgeable are you about the culture and traditions of Hong Kong or mainland China?’); alpha=0.89. The Language Factor consists of four questions (e.g. ‘How fluent are you in speaking Cantonese or other Chinese dialect?’); alpha=0.95. The Cultural Identity Factor consists of six questions (e.g. ‘How much do you feel you are a part of Hong Kong or mainland China?’); alpha=0.80. All are six-point Likert- type scales ranging from ‘not very much’ to ‘very much’.
Two questions were also asked to measure the perceived quality of the child's relationship with their parent: ‘How enjoyable is the time you and your [mother/father] spend together on the same scale of 1 (not at all) to 5 (very greatly)?’ and ‘Lastly, to what extent can you confide in your [mother/father] on the same scale of 1 (not at all) to 5 (very greatly)?’ Structural factors included care-givers’: gender (male/female); age (continuous); education (continuous); employment status (full-time, part-time, not at all); income (categorical); marital status (married, divorced/separated, not married); number of children; persons in household; distance from care-receiving parent; number of siblings; friends and others in network. Parental structural characteristics included: gender (male/female); age (continuous); and current marital status (married, widowed).
Questions were also asked about the child's perceptions of their own health (a single indicator) and the parent's health: need for assistance in ADLs (summed items); need for assistance in IADLs (summed items); short-term memory difficulties (yes, no); perception of parent's overall health (poor to excellent); and diagnosis of dementia or other forms of cognitive impairment (yes, no).
First, descriptive characteristics of the sample as a whole are estimated (frequency distributions). Then bivariate Pearson correlations examine the relationship between attitudes and each of the five care-giving behaviours for the total sample. Multiple regression analyses identify correlates of each care-giving behaviour while controlling for other factors, with attitudes of filial responsibility and cultural group as independent variables, and care-giver and parent socio-demographic and health characteristics and perceived relationship quality as control variables. All data were checked for multicollinearity, linearity and homoscedasticity. The two attitudinal measures of filial responsibility were multicollinear and therefore were entered into separate equations. The multivariate analyses were repeated for each of the three cultural groups separately.
Results
Sample description
In total, 125 Hong Kong-Chinese, 90 Chinese-Canadians and 100 Caucasian-Canadians were interviewed (N=315); the majority of the sample is female (73.7%) (see Table 1). Few care-givers live alone (8.6%), reflecting their marital status; 21.3 per cent are single, 64.8 per cent are married or common-law. Approximately half have only one parent who is still alive (47.6%) and 75.7 per cent are caring for their mother. While the majority is over age 40 – i.e. middle-aged or older – a substantial minority (20.5%) are younger (age 20–39). Most are employed; 55.9 per cent work full-time; 15.6 per cent work part-time.
Table 1. Sample description (frequencies for selected characteristics) – total sample
Bivariate correlations
Attitudes (filial expectancy and filial piety) are unrelated at the bivariate level to the number of ADLs with which assistance is provided (0, 1, 2, 3, etc.) (see Table 2). Attitudes are correlated with the provision of IADLs. Those scoring lower on both filial expectancy and filial piety are more likely to provide assistance with more IADLs, a direction opposite to what one might expect. That is, assistance with more instrumental tasks is more likely if the adult child espouses less support for filial responsibility, suggesting that attitudes in this area should not be assumed to lead to care behaviours. However, though filial expectancy is unrelated to emotional support, those scoring higher on filial piety are more likely to provide emotional support to their parent. Further, those scoring higher on both expectancy and piety are more likely to provide companionship and to provide financial support.
Table 2. Bivariate correlations between attitudes and behaviours – total sample
Notes: ADL: basic activities of daily living. IADL: instrumental activities of daily living.
Significance levels: NS: not significant, * p<0.05, ** p<0.01.
Multivariate analyses (entire sample)
Table 3 presents the results of the multiple regression analyses for each of the five measures of care-giving behaviours, for the total sample. Only statistically significant associations are shown in Table 3. All of the regressions are relatively successful in explaining variance, with total variance explained ranging between 18 per cent for companionship and 58 per cent for IADL assistance. Notably, however, the filial responsibility measures (both expectancy and piety) are significant in predicting only one type of care-giving behaviour – emotional support (in a positive direction). Even here, it is not among the strongest correlates. These multivariate analyses indicate the relative lack of relationship between filial attitudes and subsequent behaviours.
Table 3. Ordinary least-squares regressions – total sample (standardised beta and significance level)
Notes: 1. Filial piety was multicollinear with filial expectancy and was therefore entered in a separate analysis. ADL: basic activities of daily living. IADL: instrumental activities of daily living. CG: care-giver. CR: care recipient. ADL: R 2=0.38; F=37.46; df=5+301; p<0.000. IADL: R 2=0.58; F=83.38; df=5+306; p<0.000. Companionship: R 2=0.18; df=3+307; p<0.000. Emotional support: R 2=0.29; df=5+297; p<0.000. Finances: R 2=0.34; df=4+306; p<0.000.
Significance levels:* p<0.05, ** p<0.01, *** p<0.001.
However, cultural group is a significant and strong correlate of the support provided. Caucasian-Canadian respondents are more likely, in comparison to the other groups, to provide more care with ADLs, IADLs and emotional support; and significantly less likely to provide any financial assistance to their parents. Hong Kong-Chinese respondents are significantly more likely (than the other groups) to provide support with ADLs, but not IADLs or companionship.
The perceived relationship between the child and parent is also predictive of care-giving behaviours. Being able to confide in the parent is related to the provision of more companionship, more emotional support and more financial assistance. Enjoying time with the parent is related to greater emotional support but less financial assistance. In sum, perceived relationship and culture emerge as important in predicting care provision. It is not the filial attitude component of culture that is related, but other unspecified aspects reflected in the cultural group variable.
Several other findings are noteworthy. As might be expected, parental ill health is associated with more assistance with ADLs and less emotional support. Several structural variables are also significant. Care-givers are likely to provide more assistance with ADLs if they are unemployed and if they live with the parent, and with IADLs if they have lower levels of formal education. In addition, older children tend to provide more assistance with IADLs and emotional support, but less assistance with finances.
In sum, whereas attitudes of filial responsibility are not strongly associated with care-giving behaviours (i.e. only with emotional support), cultural group predicts all five care-giving behaviours examined. Further, confiding in the care recipient (one of the indicators of relationship quality) is associated with three of the behaviours (companionship, emotional support, and financial support), whereas enjoying time together (another indicator of relationship quality) emerges for two (emotional and financial support). These findings suggest the overall importance of both relationship quality and cultural group in care-giving practice and suggest relatively less of a role for attitudes of filial responsibility. Several structural factors (co-residence, employment, age, education) also emerge as related to behaviours, but none do so for all five behaviours consistently. Furthermore, with the exception of age, the structural factors are related to assistance with tasks (ADL and/or IADL). In contrast, cultural group and relationship quality emerge as the important predictors of emotional support, companionship and financial support. Such findings suggest that cultural and structural influences may operate differentially, depending on the type of care-giving behaviour examined.
Multivariate analyses (separate subsample analyses)
Because cultural group emerges as one of the strongest correlates of each of the five care-giving behaviours, multivariate analyses were conducted for each of the groups separately (not shown here but available on request), with all of the controls noted above. This approach controls for cultural differences and allow for the comparison of differences in the association between filial responsibility attitudes and care-giving behaviours across cultural groups. Among the Hong Kong-Chinese and Chinese-Canadian groups, but not the Caucasian group, those scoring higher on filial expectancy and filial piety are more likely to provide more emotional support. In addition, in the Hong Kong-Chinese group only, those scoring higher on filial expectancy are more likely to provide companionship. In the Chinese-Canadian group only, those scoring higher on filial expectancy and filial piety are more likely to provide financial assistance. Care-giving attitudes are not related to care-giving behaviours among the Caucasian-Canadians.
Discussion and conclusions
The dominant assumption that filial care-giving attitudes predict care provision persists despite empirical evidence that adult children from all cultural groups care for their parents. Given the longevity of this assumption, this paper examines the empirical relationships between attitudes of filial responsibility and actual care-giving behaviours among adult child care-givers. Three different cultural groups were included that reflected differing cultural norms (the West, Hong Kong and a diasporic group). When examined in multivariate analyses in the sample as a whole, attitudes by and large do not predict care-giving behaviours, with the exception of emotional support – and even then, it is not among the strongest predictors.
That is, filial attitudes are generally not predictive of care behaviours. This finding supports earlier research demonstrating that filial care is provided irrespective of specific cultural or social prescriptions (Hamon and Blieszner Reference Hamon and Blieszner1990; Montgomery, Borgatta and Borgatta Reference Montgomery, Borgatta, Borgatta, Liu and Kendig2000). Multivariate analyses conducted within each cultural group, however, qualify this finding. Filial attitudes are unrelated to any of the five care-giving behaviours among Caucasian-Canadian respondents, but are related to the provision of emotional support in the Hong Kong-Chinese and Chinese-Canadian subgroups, as well as financial support among Chinese-Canadian respondents and companionship among Hong Kong-Chinese respondents. Such findings suggest some support for the idea that socialisation into filial piety within Chinese culture may have relevance to actual care-giving when parents age (Chen, Bond and Tang Reference Chen, Bond and Tang2007; Zhan Reference Zhan2004). However, given the important role of filial piety within Chinese culture historically, it is surprising that attitudes are only significantly correlated with the provision of emotional support, and to a lesser extent financial support and companionship. In other words, filial responsibility attitudes are unrelated to the care-giving behaviours that are most commonly examined in gerontological research, namely ADL and IADL assistance, suggesting a much weaker role for such attitudes than is often assumed.
A possible explanation for the relative lack of strong relationship between filial attitudes and care-giving behaviours is the nature of the selected measures. Psychological research suggests that attitudes and behaviours tend to be correlated when sufficient specificity is obtained in the measures. In other words, asking about attitudes in a more abstract, less personal way (e.g. about adult children ‘in general’), and about care for parents in general rather than about the specific types of behaviours, may be less likely to result in significant correlations (Palys and Atchison Reference Palys and Atchison2008). In the present study, the two measures of care-giving attitudes have a general referent (e.g. ‘adult children should’) but include some particular care behaviours. The expectancy measure included: living close to parents, co-residence, taking care of sick parents, financial support, and visiting. The piety measure included ‘looking after’ parents, financial support, retaining contact, but also respecting parents, taking their direction, and pleasing them/making them happy.
While there is some overlap (e.g. between ‘making parents happy’ and emotional support, or between ‘taking care of parents’ and ADLs/IADLs), the only item directly specific to the care-giving behaviours used in these analyses is financial support, asked in both scales. Even here, attitudes are related to this behaviour only among Chinese-Canadian respondents and notably not among the Hong Kong-Chinese. Furthermore, these scales have been validated as representing the normative dimensions of filial responsibility in both cultures, that is, they reflect the notion espoused as part of the cultural norm. Overall, then, these findings suggest similar patterns among non-Western and Western cultures: specifically, that attitudes of filial responsibility, even where they have traditionally been highly espoused cultural norms, are not strongly associated with commonly studied care-giving behaviours (with the exception of emotional support) – and certainly not as strongly as much of the literature (and policy rhetoric) suggests. Whether this represents change in Hong Kong and among the diasporic Chinese in Canada is not known. Past research tends not to compare these groups, not to empirically assess both care-giver attitudes and behaviours, nor to study the five behaviours included here. The findings could represent changing attitudes with globalisation as suggested by Sung (Reference Sung2001), Ikels (Reference Ikels2004) and others. Without adequate past research for comparison, it is difficult to say.
The multivariate analyses do, however, confirm the importance of culture, over and above filial attitudes in predicting care-giving behaviours. Among the two Chinese groups but not among the Caucasian-Canadians, attitudes towards providing care to parents is related to the provision of emotional support but only among Hong Kong-Chinese is it related to companionship and among Chinese-Canadians is it related to financial support. Cultural variation among Chinese groups has been reported by others. For example, Chen, Bond and Tang (Reference Chen, Bond and Tang2007) found mainland Chinese scored higher on filial behaviours than those living in Hong Kong (but found no differences between the two on filial attitudes). In the present research, Hong Kong-Chinese and Chinese-Canadians are similar in some respects (suggesting the Westernisation of Hong Kong may have some similar influences as for the diasporic moving to a Western culture) but they differ in other respects. An unpacking of the aspects of culture that are captured in the cultural group variable that are significantly related to care provision awaits future research to clarify the cultural group variable. Nevertheless, these findings suggest important cultural differences in care provision beyond filial attitudes, intergenerational relationship, or background and health characteristics, all of which were controlled in these analyses.
Confirming earlier research by Keefe et al. (Reference Keefe, Rosenthal and Beland2000), both structural and cultural factors are related to care-giving behaviours, but structural factors, together with parental health, emerge among the strongest correlates of the provision of ADL and IADL support. Interestingly though, income is not a significant predictor of care-giving behaviours; further, care-givers with lower levels of formal education are more likely to provide assistance. In this respect, the findings challenge traditional structural explanations for care-giving (i.e. that those in lower socio-economic groups face greater barriers to being able to provide care). However, it is possible that the small sample size masked some significant associations. Employment, where significant, also shows a negative relationship whereby those who are unemployed are more likely to provide assistance. Two explanations are possible – some care-givers may have quit work in order to provide care; alternatively, it is also possible that being unemployed leaves more time available to provide care.
The findings also point to the importance of living arrangements for ADL and IADL assistance, supporting earlier research (Chappell Reference Chappell1992; Lee and Hong-Kin Reference Lee and Hong-Kin2005a). Ill health of the care recipient is a significant predictor of at least some of the care-giving behaviours, which is not surprising. Also not surprisingly, the quality of one's relationship with parents also emerges as important for care-giving behaviours – though more important for the reported provision of emotional support, companionship and financial support. These results suggest the need for a more careful look, in future research, at task-oriented care-giving behaviours in comparison with other types of care-giving behaviours that are less task-oriented and may have differing correlates. This includes care-giving behaviours that are more emotional/interactive (e.g. emotional support and companionship) as well as the provision of financial support. Indeed, the results indicate the desirability, when examining care-giving behaviours, of ensuring that several types of behaviours are included before drawing conclusions about the provision of help. These findings indicate different factors operate to result in different types of care from adult children.
It should be remembered that with the exception of the Hong Kong-Chinese group, these samples were non-random, and thus not necessarily representative – generalisation to larger groups is unwarranted except for the Hong Kong-Chinese sample. The small sample size, particularly for the cultural subgroup analyses, is another limitation. Further, the data are cross-sectional in nature, so conclusions about potential causality must be made with caution. Nevertheless, the findings suggest further research is warranted on gerontological assumptions about the role of societal norms and personal attitudes of care-giving in the provision of care to ageing parents.
Acknowledgements
The research reported here was funded by the Social Sciences and Humanities Research Council of Canada. We would also like to acknowledge the assistance of Dr Kee Lee Chou, Dr David Lai, Linda Outcalt, Kileasa Wong and Carren Dujela.
