Page 64 - IJPS-2-1

Correlates of parental satisfaction: a study of late life family relationships in a rural county in China

ranging from 1 (extremely dissatisfied) to 7 (extremely satisfied)?” Offspring’s socioeconomic status
(SES) was measured in the same manner. Like the independent variables previously mentioned, res-
ponses to each question were combined across multiple children and organized into three categories,
respectively: 1=all children met expectation, 2=at least one but not all met expectation, and 3=none
met expectation.
2.4 Analyses

We first calculated descriptive statistics and checked intercorrelations for measures employed. Next,
we ran logistic regression models to examine the influence of the hypothesized correlates on parental
satisfaction in late life, starting from relationship quality (Model 1), followed by emotional support
(Model 2), practical support (Model 3), financial support (Model 4), and then offspring’s filial piety
(Model 5). Parent and offspring characteristics such as age, gender, marital status, financial strain,
self-rated health, offspring’s marital status, and offspring’s socioeconomic status were controlled for
in each model. Model 6 represented the full model including all variables tested in previous models.
The analyses were performed using Stata/MP 13.1.

3. Results

3.1 Sample Description
Table 1 presents sample descriptive statistics. The sample was composed of 432 older adults with a
mean age of 66.24 years (range 60–79; SD=5.00). The majority of the respondents were women
(55.32%), married (77.08%), experiencing no financial strain (60.42%), satisfied with offspring’s
marital status (67.36%), and satisfied with being a parent (64.81%). Less than half of the sample was
satisfied (48.15%) with each of the grown children’s SES while near a third (27.08%) was satisfied
with none of the grown children’s SES, though.
3.2 Correlates of Parental Satisfaction

Table 2 summarizes bivariate correlations among all variables included in the analyses. All indepen-
dent variables correlated significantly to parental satisfaction in expected directions and ranged from
0.13 (practical support from offspring) to 0.32 (getting along with offspring). We assessed variance
inflation factors (VIFs) for multicollinearity diagnostics. All calculated VIFs (1.08–1.50) fell
well below 4.00, the suggested threshold of multicollinearity (Cohen, Cohen, West et al., 2003).
Table 3 displays the results from logistic regression models. In the first five models, the five tested
variables – getting along with offspring, offspring’s emotional support, offspring’s practical support,
offspring’s financial support, and offspring being filial — each exhibited a statistically significant and
positive association with parental satisfaction, respectively. Specifically, compared with respondents
who got along with all children, respondents who got along with none of his/her children were con-
siderably less likely to report being satisfied with the parental role (odds ratio=0.19, p<0.001). Simi-
larly, parents who reported being satisfied with none of his/her children’s emotional, practical, or
financial support, respectively, were markedly less likely to be satisfied with their parental role rela-
tive to those who reported all their children met expectations in terms of providing for them emotional,
practical, or financial support, respectively. In the fifth and the sixth model, a dose-response rela-
tion between offspring’s filial piety and parental satisfaction was revealed. Older parents who were
satisfied with none of their children’s filial piety were remarkably less likely to be satisfied with their
parental role compared with those who were satisfied with at least one child’s filial piety, while the
latter were less likely to be satisfied with their parental role relative to those who were satisfied with
all children’s filial piety. The sixth model represents the full model in which only relationship quality
with offspring and offspring’s filial piety remained significant. The overall model was statistically
2
significant (χ = 101.19, p<0.001).

58 International Journal of Population Studies | 2016, Volume 2, Issue 1