The main methodological issues that should be considered in studies of premenstrual changes are discussed. They include: the selection of well-defined groups of subjects who reflect the diversity of subtypes of premenstrual changes (PMC); the confirmation of retrospective reports through daily monitoring of changes by ratings, or by objective procedures when possible; the need to consider the diversity of premenstrual biological changes instead of comparing average levels, since there is a likelihood that different pathophysiological changes are connected with diverse behavioral and mood changes; application of a multivariate, time-related approach to explore the pathophysiology of PMC; the need to exclude placebo responders prior to the active drug phase in treatment trials and the need for such trials to be double-blind, placebo-controlled and, if possible, of a cross-over design. Attention to such issues should lead to increased consistency of findings across studies and eventually to a better understanding of the pathophysiology of PMC and to a rational, effective treatment.