The analysis is based on individual 1996 TDHS data combined with aggregate data from the 1988 census and the 1991/1992 TDHS. When various sources of spuriousness are taken into account, education is found to reduce fertility much less than suggested by univariate tabulations of total fertility rate.