INTRODUCTION AND AIMS: To investigate behaviours related to four alcohol policy variables (policy-relevant behaviours) and demographic variables in relation to typical quantities of alcohol consumed on-premise in six International Alcohol Control study countries. DESIGN AND METHODS: General population surveys with drinkers using a comparable survey instrument and data analysed using path analysis in an overall model and for each country. MEASURES: typical quantities per occasion consumed on-premise; gender, age; years of education, prices paid, time of purchase, time to access alcohol and liking for alcohol advertisements. RESULTS: In the overall model younger people, males and those with fewer years of education consumed larger typical quantities. Overall lower prices paid, later time of purchase and liking for alcohol ads predicted consuming larger typical quantities; this was found in the high-income countries, less consistently in the high-middle-income countries and not in the low middle-income country. Three policy-relevant behaviours (prices paid, time of purchase, liking for alcohol ads) mediated the relationships between age, gender, education and consumption in high-income countries. DISCUSSION AND CONCLUSIONS: International Alcohol Control survey data showed a relationship between policy-relevant behaviours and typical quantities consumed and support the likely effect of policy change (trading hours, price and restrictions on marketing) on heavier drinking. The path analysis also revealed policy-relevant behaviours were significant mediating variables between the effect of age, gender and educational status on consumption. However, this relationship is clearest in high-income countries. Further research is required to understand better how circumstances in low-middle-income countries impact effects of policies.