The Association of Pension Income with the Incidence of Type I Obesity among Retired Israelis

Table 11

Random-Effect Regression 2015-2016: stratification by gender.

Variables	(1)	(2)	(3)	(4)
	Full	Stepwise	Full	Stepwise
	BMI30	BMI30	BMI30	BMI30
Constant	0.2245	0.2528	0.3330	0.2715
	(0.4940)	(<0.0001)	(0.3526)	(<0.0001)
INCPENS ÷ (10^4)	−0.0041	−0.0052	0.0015	—
	(0.0021)	(0.0001)	(0.4225)	—
Owner	0.0031	—	−0.0342	—
	(0.9375)	—	(0.3460)	—
Books	0.0806	—	0.0013	—
	(0.2047)	—	(0.9875)	—
Car	−0.0460	—	0.0610	0.0641
	(0.1774)	—	(0.0897)	(0.0469)
Age	−0.0014	—	0.0003	—
	(0.7353)	—	(0.9494)	—
Academic	−0.0623	—	−0.1014	−0.0993
	(0.0966)	—	(0.0118)	(0.0106)
HHSIZE	0.0239	—	0.0059	—
	(0.1549)	—	(0.7551)	—
Married	0.0356	—	−0.0442	—
	(0.4045)	—	(0.7685)	—
Divorced	0.0925	—	−0.1623	—
	(0.0319)	—	(0.3060)	—
Widow	0.1128	0.0658	−0.1004	—
	(0.0048)	(0.0419)	(0.5222)	—
IMM_EUROPE_AMERICA	0.0162	—	−0.0767	−0.0899
	(0.7047)	—	(0.0821)	(0.0398)
IMM_ASIA_AFRICA	−0.0544	—	−0.1163	−0.1199
	(0.2377)	—	(0.0245)	(0.0200)
OVERALL_HEALTH	−0.0655	−0.0747	−0.0345	—
	(0.0190)	(0.0036)	(0.2439)	—
Gender	Females	Females	Males	Males
Age	Above 67	Above 67	Above 67	Above 67
Observations	921	921	850	850
R-squared	0.0569	0.0370	0.0560	0.0370
Households	507	507	466	466
Wald chi^2: all coef. equals zero	47.25	28.59	25.09	19.67
Individual effect F-Statistics	6.79	6.98	4.58	4.69
Variance inflating factor (VIF)	2.34	1.07	2.78	1.25

Note. The table reports the estimation results of equation (3) for the group of retired women and men estimated via the random effect regression. The results of individual-effect F-statistics justify this procedure, which includes the individual effect dummies. The dependent variable is BMI30, a dummy variable, which equals one for Type I obesity (BMI ≥ 30) and zero otherwise. The stepwise procedure gradually omits variables with coefficients for whom . Variance inflation factor (VIF) is a measure for collinearity among independent variables. VIF < 10 implies low collinearity. Robust values are given in parentheses. . . .