Psychometric properties of a brief headache questionnaire were evaluated in three samples of headache sufferers awaiting treatment. Although the test-retest reliability of the headache questionnaire was substantial over a two-week period, on several headache indices mean symptom levels were significantly different at the two administrations. Furthermore, when questionnaire reports were compared with daily headache recordings there was minimal correspondence between these two assessment methods. Although questionnaire methods of assessing headache symptoms appear reliable these findings suggest questionnaires may lack concurrent validity. It is suggested that headache questionnaires be supplemented with daily headache recordings whenever possible.