OBJECTIVE: Ratings of usual pain over a period of 1 week are commonly used to rate a patient's usual level of pain intensity. This study investigated the validity of weekly recall pain ratings and biasing effects of pain levels on these ratings. METHODS: Seventy-eight patients presenting to a chiropractic outpatient clinic with nonspecific neck pain completed a 7-day diary rating their pain 4 times each day on an 11-point numerical rating scale. From these 28 ratings, the patients' "actual average" pain was computed. On day 8, patients were asked to rate their current pain, as well as recall their pain "on average," at its "worst," and at its "least" over the previous week. RESULTS: Recall of average pain over the previous week was shown to be a valid measure using ratings of actual pain as the criterion standard (Pearson r = 0.95). The error between actual and recall readings was random and consistent across all levels of actual pain. Patients were also able to accurately recall their pain at its worst (r = 0.93) and at its least (r = 0.92) over the preceding week. In regression analyses, there were no appreciable biasing effects on the recall of average pain of either current pain at the time of recall, or of pain at its worst or at its least during the recall period. CONCLUSION: These results suggest that recall ratings of pain intensity may be valid for use in clinical research and practice on patients with nonspecific neck pain.