The present study design overcomes, in part, the difficulty
of establishing an adequate comparison sample for
WIC evaluations by using well defined individually matched
controls available from the State Birth Registry. Standardized
birth outcome information collected uniformly on all
births, independent of WIC participation status, should
eliminate any ascertainment bias between WIC and nonWIC
subjects. The inclusion of virtually the entire 1978
Massachusetts prenatal WIC population should make the
results more robust than studies based on smaller samples
and more limited numbers of WIC sites. The larger sample
size and the matched pair study design allows for analyses of
subpopulations and low frequency birth events, neither of
which were accessible to earlier studies.