We have had the impression that patients with spinal cord injury (SCI) experience fewer marriages and more divorces than their noninjured counterparts. To test this impression statistically, we examined the influence of SCI in association with other select variables on the marital status of 276 patients injured between 1973 and 1980 and treated at the University of Alabama in Birmingham Spinal Cord Injury Care System. The expected numbers of marriages and divorces in the study population were based on comparison with reported marriage and divorce rates for the general US population. Discriminant analysis was employed to identify variables associated significantly with a postinjury change in marital status. Substantially fewer marriages and more divorces occurred than were expected (p less than 0.05). No variables were associated significantly with marrying within three years of injury. However, divorcing patients were significantly more likely to be young black women who had been previously divorced, had no children, and had Barthel scores of less than 80. Using the most effective combination of these variables, 38.7% of the variance was explained, and the postinjury marital status of 81.5% of patients married at injury was predicted correctly. While other determinants of postinjury marital status undoubtedly exist, the likelihood of divorce can be assessed using a comparatively small set of predictor variables.