Unobserved variables and marital status: the schooling connection

"This study focuses on the effect of schooling on marital status. A Hausman-type test shows that schooling cannot be legitimately treated as an exogenous determinant of marriage and divorce. It is shown that if schooling is treated as an exogenous variable, the negative effect of schooling on the odds of marriage is underestimated. Further, the results indicate that schooling has a significant negative effect on divorce if it is treated as an exogenous variable; the coefficient for schooling is positive if it is treated as an endogenous variable." The data, which concern the United States, are from the General Social Survey and cover the period 1986-1990.