This thesis aims to examine culture as an explanatory variable of democracy. Culture is operationalised in terms of religiosity. Islamic culture is measured by subjective and objective religiosity. This thesis seeks to contribute to knowledge by developing a multi-causal theoretical framework to examine the propositions of sociological reductionism regarding democracy. The proposed framework depends on cost-benefit and risk assessment at the individual level linking structural phenomena like socio-economic factors to behaviour. The hypotheses derived from the theoretical framework were tested with original, high quality, representative survey data. These data are cross-cultural and the indicators we use have been rigorously tested for validity and reliability to control for culturally specific connotations in survey questions. Monocausal explanations focusing on religion alone are largely refuted. Multivariate analysis, which incorporates all relevant variables in the literature and control for their interactions, reveals that cultural variables (Islamic religiosity) are largely irrelevant to the explanation of variations in support for democracy and authoritarianism.