Investigation of the reliability and factor structure of a questionnaire for assessment of the premenstrual syndrome.

Data are presented from 696 women who completed a 35-item retrospective self-report PMS questionnaire. Item analysis resulted in the elimination of six items which predominantly reflected physical symptomatology. The resultant (29-item) questionnaire demonstrated a coefficient alpha value of 0.94 corresponding to a very high level of internal consistency and split-half reliability. It is argued that discrepancies in previous PMS research based on questionnaire data may, in part, be due to inadequate attention being devoted to achieving homogeneity of the PMS construct. A high level of internal consistency is a necessary condition for global PMS scores to be meaningful. Factor analysis of the revised questionnaire resulted in a two factor model corresponding to 'dysphoria' and 'dysfunction'. This leads to a simple four quadrant typology of PMS which may further understanding of this 'purer' PMS construct.