SummaryThe tension reduction hypothesis has been historically and cross‐culturally a predominant explanation of alcohol's effects. However, its validation has been inconsistent. One speculative reason for this inconsistency is a failure to control for preheld alcohol expectancies. Past research may have confounded results due to studying populations mixed regarding their expectations of alcohol's tension reducing effects. Only two studies could be found which have comprehensively attempted to control for tension reduction expectancies, both supporting a modified tension reduction hypothesis which does not offer tension reduction as the only explanation of drinking behaviour. The interaction of pharmacology, expectancy, gender role and situation suggest that tension reduction is of major importance in understanding the drinking of many, but not all, individuals. This is consistent with recent ‘metatheories’ of alcohol use which emphasize the complexity and multiplicity of alcohol's effects. Future theoretical and clinical implications of the revised tension reduction hypothesis and expectancies are briefly disc