The relationship between persistent pain in spinal cord injury and medical-descriptive, demographic, psychological and familial-social data was studied. Multiple linear regression and discriminant analysis were used to predict (1) presence or absence of pain; (2) severity of pain; (3) time post-injury onset of pain; (4) whether or not pain interfered with activities of daily living. The best combinations of predictor variables accounted for only 15 and 19% of the dependent measures pain vs. no-pain and onset of pain, respectively. The best combinations of predictor variables accounted for 43 and 44%, respectively of the dependent measures severity of pain and whether or not pain interfered with activities of daily living. Higher levels of subjective pain were associated with greater age, higher verbal intelligence, higher levels of anxiety and a more negative psycho-social situation. Persons who reported pain interfering with activities of daily living were more likely to be older, of higher intelligence, more depressed, clinically rated as experiencing greater levels of distress and immersed in a more negative psycho-social environment. The importance of psycho-social variables in the understanding of persistent spinal cord injury pain and the need for prospective studies along these lines are demonstrated. © 1980.