Skip to main content
. 2021 Feb 5;18(4):1526. doi: 10.3390/ijerph18041526

Table A2.

Logistic regression coefficients (and standard errors) of a hierarchical ordered logit model (i.e., model for ordinal dependent variables, 2nd order PQL-estimation) explaining self-rated health, using ‘very good self-rated health’ as the reference category (N = 2730).

Model Parallel to Model 4 in Table 3 B (S.E)
Level 2: Neighborhood-level characteristics (fixed effects)
Social trust (standardized) −0.102 * (0.049)
Informal social control (standardized) n.s.
Disorder (standardized) 0.110 * (0.050)
Level 1: Individual-level characteristics (fixed effects)
Gender (ref. cat. = women) −0.178 * (0.077)
Age (standardized) 0.459 *** (0.046)
Nationality (ref. cat. = Belgian) n.s.
Education (ref. cat. = high)
Low 0.745 *** (0.114)
Middle 0.350 *** (0.088)
Home ownership (ref. cat. = owner)
Tenants in social rented housing 0.339 * (0.137)
Tenants in private rented housing n.s.
Individual social support (standardized) −0.291 *** (0.043)
Network size (ref. cat. = ≥26)
≤10 0.303 ** (0.103)
11–25 n.s.
Intercept (ref. cat. = very good self-rated health)
Very bad self-rated health −5.683 *** (0.246)
Bad self-rated health or worse −3.884 *** (0.138)
Moderate self-rated health or worse −2.107 *** (0.107)
Good self-rated health or worse 0.694 *** (0.096)
Random effects
Neighborhood 0.005 n.s.

N (level 1) = 2730, N (level 2) = 142; *** p < 0.001, ** p < 0.01, * p < 0.05, n.s. = not significant. Estimated in MLwiN version 3.02.