Abstract
In his 2002 article in Mens en Maatschappij, Dronkers argues that data for the Netherlands (Enschede Panel Study 1964-1992) offer support to Herrnstein and Murray’s hypothesis that more intelligent persons have a lower divorce risk than less intelligent people. The replication of the data analyzed by Dronkers presented in this article shows that Dronkers’ conclusion is based on an incorrectly specified model in which educational attainment is not included. In a model that includes educational attainment the effect of intelligence is insignificant and educational attainment has a negative effect on divorce. This negative effect is conflicting with results from more representative data on the Netherlands. This may be due to the low number of divorces in the Enschede Panel Study (67), and to some flaws in the Enschede data. Models estimated on the Family Survey Dutch Population 1998 (123 divorces) and the survey Divorce in the Netherlands 1998 (1746 divorces) show that in the Netherlands the effect of educational attainment on divorce is positive.
