Skip to main content
. Author manuscript; available in PMC: 2022 Jun 1.
Published in final edited form as: J Cross Cult Gerontol. 2021 Mar 8;36(2):121–137. doi: 10.1007/s10823-021-09427-x

Table 5:

Predicted probabilities of primary sources of social support by childless status (N=14,575)

2005
2011
Dependent variables Childless With children Childless With children

Rely on for sick care
Children/children-in-law 0.14* 0.49 0.17 0.29
Spouse 0.53 0.46 0.29 0.19
Grandchildren or other relatives 0.08* 0.01 0.01 0.01
Nonrelatives 0.10* 0.01 0.00 0.01
Nobody 0.15* 0.02 0.00 0.01
Dead/lost to follow-up 0.52 0.49
Ask for help when having problems
Children/children-in-law 0.15* 0.53 0.15* 0.30
Spouse 0.46 0.43 0.22 0.18
Grandchildren or other relatives 0.09* 0.01 0.02 0.01
Nonrelatives 0.15* 0.02 0.05 0.01
Nobody 0.15 0.02 0.00 0.01
Dead/lost to follow-up 0.56 0.49
Talk frequently in daily life
Children/children-in-law 0.06* 0.30 0.04* 0.15
Spouse 0.61* 0.51 0.29 0.23
Grandchildren or other relatives 0.09* 0.03 0.05 0.02
Nonrelatives 0.23* 0.14 0.07 0.09
Nobody 0.01 0.02 0.00 0.02
Dead/lost to follow-up 0.55 0.49
Share thoughts
Children/children-in-law 0.09* 0.39 0.09* 0.21
Spouse 0.62* 0.53 0.29 0.23
Grandchildren or other relatives 0.07* 0.01 0.03 0.01
Nonrelatives 0.09 0.03 0.01* 0.03
Nobody 0.13* 0.03 0.05 0.02
Dead/lost to follow-up 0.55 0.49

Note:

*

refers to significant difference in predicted probabilities between childless elders and elders with children at least at the p<.05 significance level. Predicted probabilities are calculated based on multinomial logit regression models of primary support sources, assuming all the covariates at their means. The covariates include age, gender, years of education, marital status, difficulty in ADLs, and rural (vs. urban) residence in 2005.