McGill Pain Questionnaire profiles for 21 patients with chronic low back pain who underwent a relaxation training trial were recorded upon admission to the programme, at discharge from relaxation training and at follow-up. Significant differences were obtained on the following variables: present pain intensity, number of words chosen, pain rating index total, pain rating index sensory, pain rating index affective and pain rating index evaluative. Post-hoc testing revealed that, on four of the six variables, improvements were not maintained at follow-up, with significantly higher scores being recorded at follow-up than at discharge. Discussion centres on the possible way relaxation training influences the pain experience and the possibility of extraneous variables influencing these findings.