Arab public opinion research has grown significantly within the last decade thanks to the increasing number of public opinion surveys. While this trend has provided important insights for understanding the perceptions of Arab publics, numerous questions remain about the conceptual validity of items used in the Arab opinion polls. The lack of uniformity in survey methodologies, item incompatibility, and contextual differences are some of the issues casting doubt on the findings of social scientific research using Arab public opinion polls. In this paper, we aim to address the problems of contextual differences and conceptual validity of survey questions by introducing a novel conceptualization religiosity. Much of the existing literature examining public opinion surveys in the Middle East have addressed the question of how religiosity influences political attitudes. The country of origin for the participants have been taken into account in some studies as a dummy variable or through examining these variables within country. Although these studies have importantly contributed to an understanding of the complexity of religiosity and regime preferences, few studies have attempted to investigate patterns of preference across countries based on the domestic political context and the existing regime’s approach toward religious practices and identities in the public sphere. This problem is exacerbated by the assumption that there is little cross-country variability in the reception of the items measuring religiosity. We propose a conceptual model to assuage some concerns related to cross-country comparability of religiosity as a predictor variable across 12 Arab societies. Utilizing multiple waves of Arab public opinion surveys and using latent class analysis along with a more nuanced measurement of “religious classes” that might populate a domestic context, we intend to show that orientations toward politics by those identifying as religious are distributed in patterns that can be predicted by their existing regime type and its practices toward religious faithful. While regimes with a fairly liberal public sphere in regard to religious practice show a more normal distribution among religious classes and preferences, regimes with stricter control on the public sphere have a different pattern of distribution befitting the political context within which its devout citizens live. Thus, some of the variability in previous studies concerning the measurement religiosity can be attributed to the national context. This paper contributes to the emerging literature about the validity and cross-national comparability of public opinion surveys in Arab societies by introducing a novel conceptual model and context-dependent measurement strategies.