Abstract
This study is one of the few multivariate analyses of the relations of several demographic and socioeconomic variables to fertility and the use of contraception. For fertility, 56 of the 92 hypothesized paths are found to be significant at the .05 level or better. The five variables having a significant and direct effect on fertility, as shown by their path coefficients (p), are: duration of marriage (p= .721), spouse’s cohort (p= −.093), spouse’s age at marriage (p = .052), caste (p = −.071), and number of siblings of husband (p = .050). p] The use of contraception is affected by, in order of importance, the spouse’s education (p = .267), the husband’s education (p = .099), the husband’s income (p= .089), and surplus children, i.e., number of living children exceeding ideal number of children (p = .059). Child mortality, which is linked to number of living children and thus indirectly to surplus children, is affected by, in order of importance, the number of children ever born (p = .723), the husband’s education (p = \t-.166), the spouse’s absence by death or separation (p = .084), and family structure (p = \t-.035). p] The advantages and disadvantages of path analysis for this type of research are briefly mentioned.
