BACKGROUND AND PURPOSE: The efficacy of interventions for headache is often based on patient estimates of the headache parameters of frequency, intensity and duration, which can be ascertained from interview, by questionnaire or from a headache diary. The aim of the present study was to test the accuracy of retrospective patient reports of headache frequency, intensity and duration, with respect to measures derived from a daily headache diary. METHOD: Forty participants suffering from least one headache per month completed a daily headache diary for four weeks. At the end of the four-week period participants completed a questionnaire requiring them to estimate the frequency, usual duration and usual intensity of their headaches. Spearman rank correlation coefficients were calculated to determine the relationship between diary and questionnaire data. RESULTS: Correlation coefficients (Spearman's rho) for questionnaire and diary data for headache frequency and duration were 0.80 and 0.72, respectively, whereas that for intensity was 0.51. CONCLUSIONS: The findings suggest that when compared with the use of a headache diary, patient estimations of headache frequency and duration by questionnaire are reasonably accurate. Headache intensity appears to be more difficult to remember and report, possibly because of the multidimensional nature of pain, as opposed to the temporal characteristics of frequency and duration. It is recommended that frequency and duration be preferred over intensity as outcome measures if questionnaires or retrospective reports are used.