Download PDF - Ivie

approximately 7,218 observations, a sample size consistent with those of previous
studies using PSID data to analyze obesity and the labor market (Cawley, Grabka,
Lillard, 2005; Conley and Glauber, 2007; Kano, 2008).
We run regressions of each spouse’s physical attributes, controlling for their
physical, demographic, and economic characteristics. We use three dependent variables
in our main analysis: BMI, weight, and height.
The other regressors are own age and the educational level (the latter defined as
the number of completed years of schooling, and top-coded at 17 for some completed
graduate work); the number of children in the household under 18 years of age; and the
earnings of each spouse. Non-working wives have zero earnings and are assigned 0 log
earnings, while working wives have positive earnings and are assigned the log of their
earnings. The health status originally recorded by the PSID is a 5-category variable
(from excellent to poor health); this is the basis of our health dummy variable: 1 if
excellent, very good, or good; 0 if fair or poor. State dummy variables are included to
capture constant differences in labor markets and marriage markets across geographical
areas in the US, such as the proportion of obese men and women and cultural attitudes
toward body size, and obesity in particular. As our analysis concerns several PSID
waves, year and time-state dummy variables are also used, along with clustering at the
head-of-household level. Finally, observations are weighed with the PSID-family
weights.
Because the PSID main files do not contain any direct question concerning the
duration of the marriages, we rely on the “Marital History File: 1985-2007” Supplement
of the PSID to obtain the year of marriage and number of marriages, to account for the
duration of the couples’ current marriage. We merge this information to our main
10