The hazards of predicting divorce without crossvalidation

Divorce prediction studies (e.g., Gottman, Coan, Carrere, & Swanson, 1998) suggest that couples eventual divorce can be very accurately predicted from a number of different variables. Recent attention to these studies has fai led to consider the need to crossvalidate prediction equations and to consi der the prevalence of divorce in the population. We analyze archival data t o demonstrate that accuracy and predictive value drops precipitously during crossvalidation. We conclude that results of studies without crossvalidati on analyses should be interpreted with extreme caution, no matter how impre ssive the initial results appear to be.