In this paper, we study the evolution of US divorce rates across states, from 1956 to 2014, using a cluster algorithm. This methodology allows us to determine the existence of divorce convergence among the US states. Our findings indicate that there are four patterns of divorce behavior in the US. We explore whether the divorce convergence is due to the liberalization of the divorce laws. Supplementary analysis of the factors related to the club classification reveals that, in the pre‐reform period, geographical variables are important, but in the post‐reform period marital patterns appear to be associated with the club classification.