The dependent variable for the analysis of temporary
method choice is the log odds that couple i in village j
chooses Method k (pill, IUD, or injection) relative to Method
1 (no method). The dependent variable for the analysis of
sterilization is the log odds that a nonsterilized couple became
sterilized by 1988 or was lost to follow-up, relative to
remaining nonsterilized. Other than the formulation of the
dependent variable, the two analyses are structured identically.
C. is a vector of attributes associated with village j; X11
is a vec'tor of characteristics associated with couple i in village
j; CXv represents selected interactions between village
variables and the characteristics of couples. The village variables
included in our analyses are whether a subdistrict
health center is located in the village; remoteness, measured
as obstacles to traveling to Nang Rong town; the productiv
ity of rice (kg/rai of land); and involvement in jute production.
The individual, couple, and household variables are
wife's age, husband's age, whether the household is nonnuclear,
education, occupation, landownership, and the value
of agricultural assets. The J.J.
1 in Eq. (1) denotes unobserved
characteristics of the community affecting contraceptive
choice. The effects of the unobserved community variables
are presumed to be additive.