In recent years many studies have reported significant empirical associations between fertility and marital dissolution. Whether this is a causal effect or only a correlation is not clear. This issue is explored by using matching methods. First the effect of ''having children'' (binary treatment) on marital disruption is investigated. Then, the method is extended to the case of ``number of children in the household'' (multi-valued treatment). The main findings indicate that parents do not divorce less in the presence of children but they only postpone the decision to divorce until children get older.