Change of Address

Abstract

A sample of 1349 cousin pairs from 344 families was used to examine whether correlations between family characteristics and children's IQs represented causal effects. The family characteristics were: Age of the mother at the birth of first child, maternal self-esteem, highest grade completed by mother and father, family poverty status, mother-only household, and number of children in household. Causal effects imply that correlations between family characteristics in one family and IQs of children in the other family (Family–IQ cross correlations) are the product of the correlation between family characteristics and IQ times the correlation between family characteristics of cousins. Significance tests consistently showed that observed Family–IQ cross correlations were substantially larger than these predicted Family–IQ cross correlations assuming causal effects. This suggested that correlations between family characteristics and children's IQs did not represent pure causal effects and were for an important part confounded by genetic and environmental “third variables” shared by extended families.