Page 105 - IJPS-3-1

Zhang Z and Lin I-F

Widowed older adults’ needs: Based on previous literature, we included the
following characteristics of the widowed: education, pension, and the number of
Activities of Daily Living (ADL) difficulties. As the majority of the older adults
in the sample, the oldest old in particular, did not receive any formal education,
education was measured as a dichotomous variable (= 1, if received one or more
years of schooling, and = 0 otherwise). Those who had a pension were coded 1, and 0
otherwise. These two variables reflected the widowed older adults’ economic needs in
later life. We indexed ADL limitations by counting the number of ADLs that an older
adult could not perform independently, including bathing, dressing, eating, indoor
transferring, toileting, and continence, ranging from 0 to 6. This variable reflected the
widowed older adults’ personal care needs.
Child’s gender and number of children: We included the number of living sons, the
number of living daughters, and the number of children living nearby—living in the
same village, town, county, or city.
Widowed older adults’ marital history: Widowed older adults’ marital history was
captured using a dichotomous measure: 1 = married twice or more; 0 = married only
once.
Control covariates included older adults’ age (a continuous variable), gender
(1 = woman, 0 = man), ethnicity (1 = Han, 0 = Non-Han), and current residence (1 = rural,
0 = urban).

2.3 Analytic Strategy

This study comprises four analyses: First, we described the characteristics of the
widowed older Chinese in the sample. Next, we conducted multivariate analyses to
examine how widowed older parents’ needs, child’s gender and number of children,
and widowed older adults’ marital history are related to the likelihood of receiving
four types of support from their offspring. The receipts of net financial transfers, sick
care, and emotional support were examined using logistic regressions. For widowed
parents who received more financial assistance from their adult children than they
gave to the children, we also examined the determinants of the net amount of financial
transfers from adult children to widowed parents using Ordinary Least Squares
(OLS) regressions. As living arrangements include more than one category, we used
multinomial logistic regressions. In the third analysis, we examined who were the
major providers of financial assistance, sick care, emotional support, and coresidence,
separately by widows and widowers. Finally, for widowed older adults who had
difficulty performing any ADL, we examined the major care provider for each ADL
activity for widows and widowers.
Overall, there was very little missing data for all the variables examined, with
the exception of financial support. Roughly 13% of the respondents with at least
one child had missing values on the amount of money received or given to either
sons or daughters. To reduce the influence of missing items on our data analysis and
inferences, we used a multiple imputation procedure to fill in missing values (Allison,
2001). The results were based on 10 multiple-imputed replicates. We used sampling
weights in all descriptive statistics. As sampling weights were solely a function of
older adults’ age, gender, and current residence, and these variables were included in
the multivariate analysis, we presented the unweighted estimates for regression models
because they are unbiased and consistent (Winship and Radbill, 1994). All analyses
were performed using statistical software Stata version 14.
3 Results

3.1 Characteristics of Widowed Older Adults in China
Table 1 shows a high prevalence of transfers from adult children to widowed older
parents in China. Nearly 85% of widowed older adults received financial support
from their adult children. The average amount of money received was 1,225 yuan

International Journal of Population Studies 2017, Volume 3, Issue 1 99