The regression results presented in table 1 show
that none of the price or availability contextual
factors are statistically significantly related
to men’s BMI in either the cross-sectional or
longitudinal models. For women, however, the results from the individual FE model reveal
that higher fruit and vegetable prices are statistically
significantly related to higher BMI: a
one dollar increase in price results in a 0.62 unit
increase in BMI, corresponding to a fruit and
vegetable price elasticity of BMI for women of
0.02. Greater availability of fast-food restaurants
is weakly statistically significantly related
to higher BMI, and increases in convenience
stores have a statistically significant positive
effect on weight, though the point estimates on
the outlet density effects are very small.
Turning to the individual- and householdlevel
covariates presented in table 1, there are several interesting differences in the estimates
by model specification. Whereas the
cross-sectional results for men suggest a positive
U-shaped relationship between income
and BMI, the individual FE results show men’s
BMI to be positively related to income. For
women, having near-high or high income is
associated with increasingly lower BMI in the
cross-sectional model,but the relationship is no
longer statistically significant once individual
FEs are controlled for. Also, in the cross sections,
higher-educated (college or more) men
and women are found to have statistically significantly
lower BMI, but the effects fall close to zero and are not statistically significant in
the individualFEmodels.Whereas in the crosssectional
model, separated women are heavier
than their married counterparts, the individual
FE results show that changes in status
from married to separated results in weight
loss (though not statistically significant), and
moving from never married to married statistically
significantly increases women’s BMI.
For men, the cross-sectional and within-person
variations of marital status effects are similar.