• divorce;
  • longitudinal;
  • marital distress;
  • marital interaction;
  • prediction

One hundred couples were followed for 13 years from the premarital period well through the primary risk period for divorce. Results of discriminant analysis indicated that couples who remain satisfied, become distressed, and divorce can be reliably classified on the basis of premarital data. Further, both previously identified demographic risk factors and couple interaction variables contributed to classification accuracy, suggesting that both types of variables play important roles in relationship outcomes. The method employed here addresses weakness in previous studies by: (a) following couples for an extended period after marriage, (b) using multiple validated self-report and observational measures, and (c) making predictions simultaneously for divorced, distressed, and satisfied couples.