Zhan,Montgomery /GENDER AND ELDER CARE IN CHINA 219 Analyses Procedures First,we conducted t-tests for differences between the means to locate gender differences in elder care practices and cultural values as well as differences between daughters and daughters-in-law.Then,we used multiple regression techniques to assess the impact of gender,cultural values (filial piety),and structural factors (family resources)on care practices. FINDINGS Gendered Care Tasks and Gendered Interpretations of Xiao Data on the provision of care by gender are shown in Table 2.Women caregivers spent more time than men in providing care-an average of 26.6 hours per week performing various caregiving duties compared to men's 15.6 hours.The care tasks of these adult children also were gendered in nature.Women spent more time on personal care and household chores,while men spent more time shopping and run- ning other errands.Notably,there were no significant differences between daugh- ters and daughters-in-law in the total hours spent providing care or in any of the three types of care.This absence of differences between daughters and daughters- in-law is inconsistent with traditional norms that would lead us to expect more care by daughters-in-law. When comparing the mean scores of caregivers'interpretations of xiao,men generally scored significantly higher than women in the measures of patrilocal norms (3.12 vs.2.68).Overall,the mean scores for social pressure were much higher for both men and women(3.76 and 4.06)than those of patrilocal norms.In addition,daughters-in-law reported higher levels of social pressure than daughters (4.39vs.3.95) Predictors of Gendered Care Tasks We used a two-step approach to assess the relative influence of cultural values and structural constraints on performance of care tasks.First,to reduce the set of independent variables to include in a multivariate analysis,we examined the zero- order correlations between each of the measures of care tasks and possible predic- tor variables.The final multivariate analyses included only those predictors that had significant zero-order correlations with one or more of the dependent variables. The remaining 11 independent variables were entered into the regression model in three steps as shown in Table 3.The initial model for each dependent variable included the gender of the caregiver,the four control variables for the gender and marital status of the care recipient,and measures of the elder's functioning levels (ADL and IADL).The two measures of cultural values(xiao)were entered into the Downloaded from gas.sagepub.com at UCSF LIBRARY CKM on January 27.2015Analyses Procedures First, we conducted t-tests for differences between the means to locate gender differences in elder care practices and cultural values as well as differences between daughters and daughters-in-law. Then, we used multiple regression techniques to assess the impact of gender, cultural values (filial piety), and structural factors (family resources) on care practices. FINDINGS Gendered Care Tasks and Gendered Interpretations of Xiao Data on the provision of care by gender are shown in Table 2. Women caregivers spent more time than men in providing care—an average of 26.6 hours per week performing various caregiving duties compared to men’s 15.6 hours. The care tasks of these adult children also were gendered in nature. Women spent more time on personal care and household chores, while men spent more time shopping and run￾ning other errands. Notably, there were no significant differences between daugh￾ters and daughters-in-law in the total hours spent providing care or in any of the three types of care. This absence of differences between daughters and daughters￾in-law is inconsistent with traditional norms that would lead us to expect more care by daughters-in-law. When comparing the mean scores of caregivers’ interpretations of xiao, men generally scored significantly higher than women in the measures of patrilocal norms (3.12 vs. 2.68). Overall, the mean scores for social pressure were much higher for both men and women (3.76 and 4.06) than those of patrilocal norms. In addition, daughters-in-law reported higher levels of social pressure than daughters (4.39 vs. 3.95). Predictors of Gendered Care Tasks We used a two-step approach to assess the relative influence of cultural values and structural constraints on performance of care tasks. First, to reduce the set of independent variables to include in a multivariate analysis, we examined the zero￾order correlations between each of the measures of care tasks and possible predic￾tor variables. The final multivariate analyses included only those predictors that had significant zero-order correlations with one or more of the dependent variables. The remaining 11 independent variables were entered into the regression model in three steps as shown in Table 3. The initial model for each dependent variable included the gender of the caregiver, the four control variables for the gender and marital status of the care recipient, and measures of the elder’s functioning levels (ADL and IADL). The two measures of cultural values (xiao) were entered into the Zhan, Montgomery / GENDER AND ELDER CARE IN CHINA 219 Downloaded from gas.sagepub.com at UCSF LIBRARY & CKM on January 27, 2015