OBJECTIVE: We aimed to investigate (1) to which extent headache diary variables correlate among each other; (2) to which extent they correlate with generic measures of quality of life and disability; (3) and whether correlations vary over time. METHODS: We performed a secondary analysis of data from 301 patients suffering from migraine and 270 patients suffering from tension-type headache collected in 2 randomized trials of acupuncture. We calculated product-moment correlation coefficients. RESULTS: Correlations between diary parameters varied to a great extent. Attack frequency, a measure recommended by the International Headache Society as a primary outcome measure for migraine prophylaxis trials, correlated only moderately with other diary outcomes. At baseline the parameters measured in the headache diary correlated only weakly with the generic outcome measures. After 12 weeks these correlations had increased considerably. CONCLUSION: Our findings suggest that participants in headache trials "learn" to fill in diaries and generic instruments in a more consistent manner during the course of the study.