• 11. Attitudes Toward Intergenerational Relations In Urban China And Taiwan
• Chapter
• University of Michigan Press
• pp. 277-301
• View Citation
• Additional Information

11

ATTITUDES TOWARD INTERGENERATIONAL RELATIONS IN URBAN CHINA AND TAIWAN

Jennifer C. Cornman, Jieming Chen, and Albert I. Hermalin

Introduction

This chapter examines attitudes and values related to familial support in China and Taiwan from the standpoint of two generations. We are in the fortunate position in that similar (but not always identical) questions were asked of elderly respondents and their children in both settings.

Three types of attitudes are examined: attitudes centering on the coresidence of older people with their children, the appropriateness of older widows remarrying, and the importance of older people retaining property to insure respect from children. There are “traditional” aspects of Chinese family structure associated with each of these attitudes, although it could be debated how ancient or Confucian the roots of each family norm are.¹ In any case, it is clear that strong expectations existed earlier that aging parents would coreside with one or more married son, that widows would not remarry, and that family property was at the core of the intergenerational bargain between aging parents and their grown children.

The availability of responses to similar questions from two generations in two countries permits a large number of analyses. In this chapter, two parallel analyses are performed, focusing first on parents’ attitudes and then on children’s attitudes. Each analysis is structured similarly: first, there is a comparison of the responses between Baoding and Taiwan; then the determinants of the attitudes (or level of agreement) are explored via a series of logistic regressions for each attitude in each country, in order to estimate the net associations between the independent variables and the outcomes of interest; finally, for questions common to both surveys, the data are pooled and logistic regressions are carried out that include a dummy variable for the country of residence to see if there is a net country effect beyond the other independent variables.

As mentioned in chapter 9, data for the present chapter come from two sources. Those on elderly parents and adult children in Taiwan come from the 1993 Survey of Health and Living Status of the Elderly in Taiwan. As in chapter 10, for analyses on elderly parents, we restrict the sample to those elderly respondents living in large cities, resulting in a sample of 809 elderly respondents. While sample characteristics for the 1989 elderly sample are described in Chapter 9, the primary difference between the 1989 and 1993 samples is that the 1993 sample of elderly is somewhat older (mean age in 1989 is sixty-nine while the mean age in 1993 is seventy-two).

Data for analyses on adult children in Taiwan also come from the same 1993 survey. However, several additional notes need to be made. First, unlike the Baoding sample of adult children, the adult children in Taiwan are the children of both urban and rural elderly and therefore they can reside in rural as well as urban areas (see Table 9.3). Due to the sampling procedure used to select the adult child sample in Taiwan, sample sizes would become too small if we restricted the analyses to only children living in large cities. To adjust for this difference in the adult child samples, we control for residence type (rural versus urban) in the analyses described below.

Second, because several of the attitudinal questions are asked only of coresident children in Taiwan, we utilized all cases in which coresident children were interviewed, randomly selecting one if there were two or more in a household. This allows us to maximize our sample size, which is 441 compared to 199 if we had used only the coresident children selected at random from all the children interviewed. The characteristics of this second sample of Taiwan children are somewhat different from those described in Table 9.3. The adult children in this sample are slightly younger, with a mean age of 38.3 compared to 40.0 for the 1989 Taiwan sample. In addition, more children are male (85 percent compared to 53 percent), and a slightly larger percent live in rural areas (79 percent compared to 71 percent).

The data for Baoding come from the 1994 Survey of Aging and Intergenerational Relations. As noted in chapter 9, we restrict the elderly Baoding sample to those respondents age 60 and over to better match the age profile of the elderly Taiwan sample. The two Baoding samples we use for analysis consist of 509 parents age 60 and over and 731 adult children.

Research Findings

Attitudes of the elderly: Agreement and structure

In this section we first examine the level of agreement between the Taiwan and Baoding elderly in three attitudinal realms, as shown in Table 11.1. Unfortunately, the questions on living arrangements were not identical in both countries. Appendix Table A shows the exact questions and response categories and how we combined them to achieve comparability. These summary categories have been brought forward to Table 11.1 and are shown with the percent distributions of the responses. In every case the traditional or conservative response category is given first to facilitate comparison. The responses to the two questions on living arrangements show that the Baoding elderly make a sharp distinction according to the specific situation. To the more general attitude on living arrangements, 70% agree that parents should live independently of children as long as their health permits. At the same time, when the question is asked in terms of widowed elderly, almost three-quarters agree that widowed elderly should not live alone.

The elderly in Taiwan, on the other hand, display a high level of support for both coresidence in general and for coresidence for the more vulnerable. Over 60% of the Taiwanese elderly feel that the elderly should live with a married child and, as in Baoding, the large majority (67%) supports the widowed elderly living with children. It seems, then, that the attitudes of Taiwanese elderly are more in line with traditional family norms than are those of Baoding parents. This traditional orientation is confirmed by the fact that the Taiwan elderly show significantly less support for coresidence if the choice is to live with a married daughter. Only 32% support coresidence in this case.²

In discussing attitudes toward remarriage as well as attitudes toward the role of property in ensuring respect for the elderly, we can make stronger comparisons between Baoding and Taiwan, because the questions used are identical or virtually so. As we found with attitudes toward coresidence, the Taiwanese elderly express more traditional attitudes toward these family norms, with the contrast in views about widow remarriage particularly striking.

Table 11.1 shows that over 60% of the elderly in Taiwan believe that it is inappropriate for older widows to remarry, while only 8% of Baoding elderly feel this way. The Taiwanese elderly also seem to hold more traditional opinions about the relationship between inheritance and respect. While 78% of Taiwan elderly believe that maintaining property is important for ensuring respect, only 42% of Baoding elderly feel this way.

These sharp differences in attitudes regarding widow remarriage and inheritance could be due, in part, to the nature of property ownership in the two countries. The ownership of private property is much more common in Taiwan than in Baoding. Compared to nearly 80% of the elderly in the Taiwan sample, only 6% of Baoding elderly live in homes owned by themselves or by their children. In addition, the Taiwan elderly are much more likely to be or have been self-employed, or to own their own business, while Baoding elderly are much more likely to work or have worked for state-owned enterprises and agencies.

Property ownership could also be an underlying factor explaining some of the differences in attitudes toward widow remarriage. Remarriage can significantly impact kinship relations, affecting not only children’s (particularly sons’) access to family property, but also their access to the patriline. Because the patrilineal kinship system appears to be stronger in urban Taiwan than in Baoding, the possible effects of widow remarriage on lineage and inheritance concerns are potentially greater in Taiwan.

The overall impression from the distributions of the four attitudes in Baoding and Taiwan is that the Baoding elderly reveal a less traditional stance on all but the question of living arrangements for widows, where the attitudes in the two settings are very similar. With this comparison in mind, we investigate which characteristics of the elderly are associated with these expressed attitudes, and whether the determinants are the same in each country. To this end, Table 11.2 presents a series of logistic regressions that display the effects of a number of basic social, demographic, and health variables on these attitudes.

In general, we find that the Taiwan elderly’s attitudes are more structured by individual characteristics, particularly for attitudes about living arrangements. For attitudes about general living arrangements in Taiwan, the elderly’s attitudes vary significantly by age, sex, coresidence status, education, health, number of sons, and ethnicity. Those who are older, male, Taiwanese, in poor health, who coreside with children, and have lower education levels are more likely to feel that the elderly should live with married children. Curiously, those elderly with more sons are likely to feel that the elderly should live on their own. One interpretation of this result is that the elderly who have many sons may have more sons living nearby from whom they can receive support. Coresiding with a married child, therefore, may not be as important. In Baoding, on the other hand, attitudes about general living arrangements vary significantly only by coresidence, with the number of sons having only a marginal effect. Those elderly in Baoding who coreside with children are more likely to feel that the elderly should not live on their own, even when they are healthy.

The effect of the number of sons is in the same direction as in Taiwan. We also find that more of the variation in the Taiwan elderly’s attitude is explained by the variables included in the model (pseudo R-square = 0.16 for the Taiwan model and only 0.03 for the Baoding model).

Similar results occur for the regressions concerning attitudes toward the living arrangements of widowed elderly. More of the variables in the model are significantly associated with the Taiwan elderly’s attitudes and more of the variation is explained (pseudo R-square = 0.13 for the Taiwan model and 0.05 for the Baoding model).³ In Taiwan, older Taiwanese respondents in poor health who currently live with a child are more likely to feel that widowed elderly should not live on their own. In Baoding, attitudes significantly vary only by sex and marital status, with men and married respondents more likely to feel that widowed elderly should not live alone.

Attitudes toward widow remarriage are not highly structured in either country. In both countries only sex is significant, with males being less likely than females to feel that widows should not remarry. It is notable, in other words, that in both countries men are more likely than women to express support for gender equality (in regard to remarriage rights). Similarly, attitudes toward the role of property in ensuring respect for the elderly are not highly differentiated by the characteristics examined. Attitudes about property and respect vary significantly by sex, coresidence status, and education in Taiwan and only by sex in Baoding, with a marginal health effect. In addition, as indicated by the low pseudo R-squares for models about widow remarriage and property and respect (all are less than 0.06), we do not gain much predictive power from the independent variables examined in either country.

The results of Table 11.2 indicate that these attitudes, for the most part, are not highly differentiated by the social and demographic characteristics included in the analysis. In addition, with the exception of some differences in significance, the structure of these attitudes appears to be similar in urban Taiwan and Baoding in that the effects of the variables examined are in the same direction. In pooling the data, we are interested in identifying significant differences in the levels of attitudes between the two countries, net of sample composition differences. To capture the country effects we created a three-category variable, which reflects country of residence as well as ethnicity. As we saw in the previous regressions, ethnicity was an important predictor of several of the elderly’s attitudes in Taiwan. To capture this effect, we constructed a combined variable. The categories for this variable are Baoding resident, Mainlander from Taiwan, and Taiwanese from Taiwan. Baoding resident is the reference category.

Turning to the results using the pooled data (Table 11.3), we find that there are significant country differences for attitudes toward a widow’s living arrangements, a widow’s remarriage, and the relationship between property and respect for the elderly.⁴ For widow’s living arrangements we find that, controlling for social and demographic differences, Mainlanders from Taiwan appear to be less traditional than Baoding residents in that they are more likely to feel that widowed elderly should live on their own. There is, however, not a significant difference between the Taiwanese and Baoding residents. Mainlanders, because of their history, have a much “thinner” kin network than the other two comparison groups, with many of them being retired soldiers who never married. Their experiences in Taiwan may well have contributed to their less traditional attitudes about the living arrangements of the elderly.⁵

For attitudes toward widow remarriage and the role of property in respect for the elderly, we also find strong country differences. Compared to Baoding residents, both Mainlanders from Taiwan and Taiwanese are significantly more likely to feel that older widows should not remarry and more likely to feel that it is important for the elderly to keep property in order to maintain respect from one’s family.⁶ These results suggest that the contrasts in attitudes between elderly respondents in Baoding and in urban Taiwan are “real” and not simply an artifact of ways in which the background characteristics of the two samples differ.

Adult children’s attitudes: Agreement and structure

The attitudes of the adult child generation show some sharp differences from those of their parents, often in unexpected directions, and present a complex combination of traditional and nontraditional views. As Table 11.4 reveals with regard to appropriate living arrangements for the elderly, a clear majority of the adult children in both Taiwan (65%) and Baoding (59%) support coresidence in general, and this proportion rises to almost nine out of ten when the more vulnerable population of widows is specified. What is striking about these results is the close agreement, on the one hand, between the adult children in Baoding and Taiwan, despite the considerable differences in the social and economic environments in which they were raised; and, on the other, the clear reversal from the attitudes expressed by their parents. In both Baoding and Taiwan, the proportions of children in favor of the more traditional living arrangement are higher than those expressed by the parents (shown in Table 11.1) and on some items the differences are substantial. For example, in Taiwan, while 67% of parents say an older widowed person should live with a married child, 88% of children express this view; in Baoding only 30%) of parents disagree with the view that an older person should live independently, while 59%) of the children disagree.7 (These findings for Baoding echo the discussion of parent and child attitudes toward coresidence in chapter 4.)

On attitudes toward widow remarriage, although the majority of the adult child generation in both Baoding and Taiwan supports the idea that an older woman should be able to remarry, Baoding children do so at a much higher level, 97%) compared to 61%). For both samples of children, support for this attitude is nontraditional compared to attitudes about coresidence. In other words, adult children in both Baoding and Taiwan are less supportive of the “traditional” ideal here (i.e., they favor the choice of remarriage), whereas in regard to coresidence they were more traditional (i.e., they favor coresidence for aging parents). Baoding and Taiwanese adult children are more likely to support remarriage for widows than are the corresponding parental generations. In Taiwan, 61%) of adult children support remarriage while only 39% of elderly parents favor remarriage. For Baoding, the difference is not as great, but 97% of the adult children support widow remarriage compared to 92% of the elderly parents. In short, on the issue of widow remarriage, children display more “modem” attitudes than their parents.

Children in Baoding show a higher level of disagreement than children in Taiwan with the notion that parents need to keep property to ensure respect (61%) of Baoding children compared to 46%) of Taiwan children disagree). In comparing the parental and adult child generations, both Baoding and Taiwan children are less supportive than their parents of the idea that the elderly need to keep property to ensure respect (54% of Taiwanese adult children agree compared to 78%) of elderly parents; 39%) of Baoding adult children agree compared to 42%) of elderly parents). It is also clear that the Baoding generations resemble each other more closely than do the Taiwan generations in regard to both widow remarriage and the importance of property for respect.

In sum, the adult children in Baoding and Taiwan resemble each other on several of the attitudes displayed in Table 11.4, but with the children in Baoding showing much higher support for widow remarriage than their counterparts in Taiwan and less support for the need to retain property. In comparison with their parental generation, the Taiwan children show sharp differences on each broad dimension, expressing a more traditional outlook with regard to coresidence, but less traditional viewpoints on the other attitudes. In Baoding, the only sharp generational difference occurs on the attitudes reflecting coresidence where, as in Taiwan, the children give more “traditional” responses.

Table 11.5 shows the extent to which the attitudes of the adult children are differentiated. It demonstrates this by employing logistic regression to examine the multivariate effects of demographic and social characteristics, as well as key parental characteristics, on the attitudes in question. In general, the attitudes of the adult children are not highly structured by the factors examined in Table 11.5. On coresidence, the results show a strong ethnicity effect in Taiwan, with the Taiwanese children more likely than the Mainlander children to support coresidence with a married child, but the Baoding children’s attitudes about coresidence are not related to any of the characteristics examined. For attitudes toward the living arrangements of widows, we find that Baoding male children are more likely than Baoding female children to feel that the widowed elderly should live on their own. In Taiwan, attitudes toward a widow’s living arrangements are related only to their parent’s attitude. If a Taiwan parent feels that a widowed elderly parent should not live on their own, their adult child is significantly more likely to feel the same way.

Adult children’s attitudes toward widow remarriage and the role of property in ensuring respect for the elderly, family norms which are not strictly dictated by norms of filial piety, reflect a strong continuity across generations, as parent’s attitudes have strong positive effects in predicting children’s attitudes. Net of the effects of children’s personal characteristics, if a parent feels that widows should not remarry, the adult child is significantly more likely to feel the same way. Similarly, if a parent feels that keeping property is important for ensuring respect, the child is more likely to report the same opinion. Although there are some differences in magnitudes, the relationship between parent’s attitude and child’s attitude is consistent across the two countries.

Of the other characteristics examined, none show consistent or strong effects across these two attitudes or across countries. In Taiwan, children’s attitudes about widow remarriage are significantly associated with education and marginally related to age. Adult children with a higher education are less likely to find remarriage inappropriate, and older children are somewhat more likely to feel that widows should not remarry. For Baoding, there are no significant effects of variables other than parent’s attitude toward widow remarriage. For attitudes about property and respect, we find that age and number of brothers are marginally associated with children’s attitudes in Taiwan, and for Baoding there is a marginal effect of coresidence with parents.

As noted at the beginning of this chapter, Taiwanese children included in these analyses live in both urban and rural areas. We included in the multivariate analyses of Taiwan children’s attitudes a control for this factor to determine whether this difference in the two samples might affect our results. As shown in Table 11.5, the urban-rural variable is marginally related to only one of the attitudes. Thus, the urban-rural factor is excluded from subsequent analyses.

The results in Table 11.5 indicate that it is difficult in both Taiwan and Baoding to account for the attitudes of the adult children with the independent variables employed, as was also the case in our earlier analysis of parental attitudes. Although there is substantial variance in the responses to many of the questions within each country, the likelihood of responding in a more or less traditional manner is not captured well by the variables at our disposal. The strong effects of parental attitudes on children’s attitudes in both countries on a number of dimensions is the most consistent effect and suggests that personal family history and dynamics may have more to do with shaping these attitudes than the socioeconomic and demographic characteristics of the respondents.

The general similarity in results for both settings makes it appropriate to pool the adult child data in order to see if significant differences exist between the two societies while controlling for the independent variables. Table 11.6 shows these results. We find that the country effect is significant only in the analysis of attitudes toward widow remarriage and property and respect. Mainlanders and Taiwanese are both significantly more likely than Baoding residents to feel that older widows should not remarry. We also find that the parent’s attitude about widow remarriage continues to have a large positive effect on the adult child’s attitude, indicating that, despite the country differences in this attitude, adult children’s beliefs appear to be influenced by their parent’s beliefs. For attitudes about property and respect, only the Taiwanese are significantly different from Baoding residents, and they are significantly more likely to feel that property is important for ensuring respect. Here again, parental attitudes about property and respect also exert an independent and significant effect on children’s attitudes.⁸

Conclusions

This chapter has examined a series of attitudes that address long-standing family arrangements in Chinese history to see whether the dramatic but distinctive changes over the last fifty years have produced notable differences in outlook in urban China and Taiwan and between parent and child generations. The elderly in Taiwan display more traditional values than their Baoding counterparts on almost all dimensions. They are more likely to favor coresidence of an elderly couple with a married child, oppose widow remarriage, and approve of maintaining property in order to ensure respect (Table 11.1). Only in the special case of a widowed older person is there a roughly similar level of agreement across the Taiwan Straits on coresidence as the best living arrangement. These country contrasts seem to be in accord with the social and economic realities in each country, with the Baoding elderly likely to have separate housing provided by work units, and with relatively little property to become entangled through remarriages or inheritance.

The multivariate analyses (Tables 11.2 and 11.3) show that individual attitudes of parents on these dimensions are difficult to predict from the sociodemographic characteristics available. Although there is variation within Taiwan and within Baoding in many of these values, these differences are more likely to arise from particular life experiences and circumstances, rather than from broader social statuses.⁹ Multivariate analyses do confirm that the Taiwanese elderly generally are more conservative than their Baoding counterparts (Table 11.3), and that with regard to living arrangements, the Mainlanders in Taiwan are less traditional than the Taiwanese (Tables 11.2 and 11.3), providing further evidence of how particular circumstances can color these attitudes.

With regard to living arrangements, grown children appear more traditional than their parents, and this reversal is particularly sharp in Baoding (Table 11.4). In the other attitudes (widow remarriage and property), however, children are less traditional than their parents. The net result of these contrasts is that adult children in Taiwan and Baoding have profiles more similar than do their parents on the range of attitudes examined here. As in the case of the older generation, the attitudes of the adult children are not highly structured by the sociodemographic characteristics employed in the multivariate analyses (Table 11.5), with parental attitude displaying the most consistent influence on children’s attitudes. The pooled regression results (Table 11.6) indicate that on widow remarriage and retention of property, the Taiwanese children remain somewhat more conservative than those in Baoding, and that some ethnicity differences persist within Taiwan.

This complex pattern of differences and similarities across countries and generations cautions against simple explanations of the relationship between attitudes and societal changes in socioeconomic structure. However, the contrasts in attitudes between urban Taiwan and Baoding for both generations are in accord with the types of changes and cultural influences that have occurred in each country, as described in chapter 9. To put the matter in a different way, and building on the results in the previous chapter, we are now in a position to respond to the question posed in chapter 9: Which is more important in explaining the degree of “traditionalism” of family attitudes and behavior—the level of economic development or the nature of the “microinstitutional structure” in which people live? With regard to the attitudes examined here, at least, the answer is clearly the latter. To put the contrast in an oversimplified form, although Taiwan is a much richer and more economically advanced society than the PRC, it is also a society in which the family remains a more central part of the social structure, in contrast to the more highly bureaucratized nature of the social order in urban China. And where family attitudes (as examined in this chapter) and behavior (as examined in chapter 10) differ between the two societies, it is generally Taiwan that looks more “traditional.” Specific institutional arrangements appear to be more important than level of development per se in producing change in these aspects of intergenerational relations.

At the same time the distinctive patterns of attitudes among the adult children show that unilinear trends away from family traditionalism and toward “modernity” cannot be taken for granted. As with many other facets of life, complex adjustments in family values appear underway, with movement toward less traditional views apparently underway in some realms, and persistence and even strengthening of traditional views and behavior in others.

NOTES

¹ Attitudes concerning coresidence are central to the historical Chinese value of filial piety (see the discussion of this central traditional value in chapter 1), while attitudes about widow remarriage and property are perhaps less so.

² Unfortunately, the Baoding survey did not have a comparable question about living with daughters.

³ This contrast brings to mind earlier research findings on changing family patterns in PRC cities which indicate that in the postrevolutionary period family behavior and attitudes were not closely related to the kinds of demographic and socioeconomic status traits commonly used by social scientists to explain variation. One possible conclusion is that in a society based upon bureaucratic allocation and assignment rather than markets and individual choice (such as the PRC in the Mao era and still to a considerable extent today), things such as bureaucratic position and personal connections are better predictors, and socioeconomic status a worse predictor, of social behavior and attitudes than is generally the case in market-based societies. See the discussion in Martin K. Whyte, “From Arranged Marriages to Love Matches in Urban China,” in Family Formation and Dissolution: Perspectives from East and West, ed. Chin-Chun Yi (Taipei: Academia Sinica, 1995).

⁴ We omit from this analysis the general questions about coresidence with grown children, since the differences in questionnaire wording (see Appendix A) make it doubtful that the pooling of Taiwan and Baoding data together is warranted. In other words, this pooled analysis is applied only where the question wording was identical or nearly so in the two countries.

⁵ Note also that in the analysis of attitudes toward the general living arrangements of the elderly outlined in Table 11.2, Taiwanese were more likely than the Mainlanders to feel that the elderly should live with children.

⁶ To further investigate structural differences in attitudes for each country, we also ran models that included interaction terms between a country dummy variable (coded 1 for Taiwan and 0 for China) and all of the other variables in the model. This allowed us to assess the degree to which the structure of the attitudes in each country significantly differs. The results of these regressions (not reported here) showed that almost all of the interaction terms were insignificant, indicating that the variables examined have similar effects on attitudes in the two countries. Thus, as indicated in the discussion of individual country results, the structure of attitudes in the two countries appears to be similar.

⁷ A generational difference also appears on the question posed in Taiwan on whether an older couple without sons should live with a married daughter. Almost 40 percent of children advocate living with a married daughter compared to 32 percent of parents. In this case, while the children are more likely to favor coresidence, they can also be viewed as less “traditional” than their parents since coresidence with married daughters is not customary among Chinese families.

⁸ We also conducted an analysis of the degree of agreement in attitudes between parents and children in each country. In general at least half of all parent-child dyads were in agreement on the various attitudes considered here, but the level of parent-child agreement was not consistently higher in one country than the other (results not shown here).

9 It is worth noting that although several potentially influential variables are omitted from Table 11.2, any hypothesized influence would have to be orthogonal to the variables included in order to greatly increase the explanatory power of the equation.

Appendix A: Wording for Attitudinal Questions on Familial Obligations and Norms

1. Who Should The Elderly Live With?

2. Who Should Widowed Elderly Live With?

3. Remarriage For Elderly Widows

4. Attitudes Toward Inheritance And Respect For The Elderly

Chapter 10, by Albert Hermalin, Mary Beth Ofstedal, and Shiauping Shih, then compares the pattern of exchanges between parents and grown children in Baoding and in urban Taiwan, in a manner somewhat parallel to the one presented in chapter 5. In chapter 11, by Jennifer Cornman, Jieming Chen, and Albert Hermalin, several categories of attitudes between parents and children of Baoding and urban Taiwan are compared, in a manner somewhat parallel to that presented in chapter 4. In general, the results presented in these chapters show that while strong filial obligations and extensive exchanges between generations are characteristic of both Chinese settings, at the same time there are a number of ways in which the results in urban Taiwan and in Baoding diverge. To oversimplify somewhat, we found that although Taiwan is clearly a more economically developed locale than the PRC or even Baoding specifically, in certain ways the patterns in the Taiwan data appear more “traditional” than their counterparts in Baoding.