Criteria for the diagnosis of headache disorders have recently been developed but require empirical validation. We present an approach to the evaluation of headache criteria in a headache clinic: define the group of headache sufferers to be studied, obtain data used in establishing the diagnosis, apply the diagnostic criteria and assess their usefulness. In applying this method to a group of patients with chronic daily headache, we found that approximately one-third would not fit the current criteria for chronic tension-type headache. We conclude that empiric validation of headache criteria based on subjective data is a complex and iterative process. Operational criteria for headache diagnosis to be used by primary care physicians should be validated in that setting and ideally should be relatively simple.