Skip to main content

Table 3 Estimated hierarchical age-period-cohort-models of adjusted self-rated health, 1986 to 2013

From: Socioeconomic status and self-rated health of Japanese people, based on age, cohort, and period

Fixed effects Coefficient P value
 Intercept −2.533 <0.001
 First quartile of income (ref = fourth quartile of income) 0.253 <0.001
 Second/Third quartile of income (ref = fourth quartile of income) 0.036 0.044
 Age (grand mean centered) 0.094 <0.001
 Age2(grand mean centered) −0.0004 <0.001
 First quartile of income*age (grand mean centered) −0.011 <0.001
 Second or third quartile of income*age (grand mean centered) −0.002 0.030
 Male(ref = female) −0.105 <0.001
 Labor force participation (ref = non-participation) −0.410 <0.001
 Non-marriage (ref = marriage) 0.327 <0.001
 Divorce/Bereave (ref = marriage) 0.046 <0.001
 Hokkaido (ref = Kanto) 0.172 <0.001
 Tohoku (ref = Kanto) 0.057 <0.001
 Chubu (ref = Kanto) 0.015 0.252
 Kinki (ref = Kanto) 0.146 <0.001
 Shikoku (ref = Kanto) 0.109 <0.001
 Chugoku (ref = Kanto) 0.154 <0.001
 Kyushu (ref = Kanto) 0.018 0.212
 Mean of age by each cohort −0.0001 0.977
Random effects Variance P value
 Period   
 Intercept 0.061 0.024
 First quartile of income 0.002 0.066
 Second/Third quartile of incomea - -
Cohort   
 Intercept 0.066 0.003
 First quartile of incomea - -
 Second/Third quartile of income 0.0001 0.389
Model fit   
 -2 Res Log-Pseudo-Likelihood 3883618 (df = 652979.9)
  1. aThe results by Cross-Classified Random Effect Model indicated that G-matrix was not positive. As a result, we excluded these variables for the random effect model