In 1981 and 1992, identical questionnaires measuring lineal and lateral wealth flows and emotional nucleation were administered to comparable samples of male household heads in rural Kenya. The 1981 results, showing limited evidence of economic or emotional nucleation within the household, were consistent with prevailing high fertility. By 1992 desired and observed fertility had declined significantly, but nucleation levels remained the same. The question then arose as to whether our bivariate measurements had been too coarse to detect subtle shifts in nucleation and/or the simultaneous effect of other independent variables. To resolve these issues, we re-examined our 1981 and 1992 data sets using regression analysis. The findings suggest that the independent contribution of wealth flow and emotional nucleation variables to explained variance in the desired number of children by rural male heads of households was limited, being approximately seven to 10 percent in 1981 and 10 to 18 percent in 1992. However, we observed that the independent variables picking up the explanatory slack in the current exercise were largely ideational in nature, having to do with the expectation of future child support, education, form of marriage, and spousal perception, communication, and consensus on key family size and family planning variables.