Skip to main content
. 2011 Nov 22;2012:659806. doi: 10.1155/2012/659806

Table 2.

Marginal effect coefficients from ordered logistic regression models predicting older adult perceived social isolation.

(1) (2) (3) (4) (5)
Residence in HDB −.169** (.086) −.162* (.087) −.186** (.087) −.209** (.088) −.219** (.088)
Daily participation in HDB neighborhood events .008 (.006) −.114 (.070) −.100 (.070) −.164** (.071) −.160** (.071)
Living arrangements
Household size −.165*** (.022) −.140*** (.022) −.138*** (.022) −.139*** (.022) −.139*** (.022)
Widowed .137* (.071) .210** (.079) .200** (.079) .197** (.080) .199** (.080)
Residing alone .683*** (.134) .680*** (.134) .615*** (.135) .652*** (.135) .655*** (.135)
Coresiding with adult children .332*** (.082) .247** (.083) .248** (.083) .234** (.084) .230** (.084)
Social network
Relatives incl. non-coresiding children −.023*** (.004) −.010** (.005) −.012** (.005)
Friends −.035*** (.005) −.037*** (.005)
Neighbors .040 (.024)
Age .011** (.004) .003 (.004) .003 (.004) −.0007 (.0044) −.0006 (.004)
Covariates No Yes Yes Yes Yes
Observations 4,542 4,542 4,542 4,542 4,542

Notes: ***P < 0.01, **P < 0.05 and *P < 0.10. The covariates are physical health limitations—ADL and IADL, income, gender, and ethnicity. The likelihood ratio test for the restricted model without the covariates and the unrestricted model with covariates shows that there is no variation in the main coefficients of interest—HDB residence, daily participation in HDB neighborhood events, family living arrangements, and the social network. LR Chi2(4) = 336.02. Prob > Chi2 = 0.0000.