Political equality, local context and political participation in Sweden

Studies of political participation have traditionally focused on characteristics of the individual in order to explain why some citizens are more active than others. But as more advanced techniques of quantitative analysis have emerged, the question of context has gained an increasing interest in political science (see Huckfeldt 1970, 1979, 1980, 1986; Miller 1977; Books & Prysby 1988, 1991, 1995; Sprague 1982; Johnston et al. 2005, 2006; Aars et al. 2009). Contextual studies are sometimes criticized for the lack of plausible theoretical explanations as to why context should have an effect. However, political socialization models involving mechanisms such as personal observation and social interaction are portrayed in the literature. These theoretical models are however, generally better suited for explaining contextual effects in smaller, more intimate milieus of the individual. The lion part of contextual studies thus focuses on the neighbourhood (or equivalent) context. The paper therefore elaborates on the question of scale (see Johnston et al. 2001, Huckfeldt & Sprague 1990; Cho & Rudolph 2008), both theoretically and empirically. A theoretical model for analysis, which combines the Civic Voluntarism Model (Verba et al. 1995) and the contextual approach, is designed. The analysis is based on survey data from Sweden. The data set, which was collected in 2002 and 2003, covers more than 7000 respondents in 38 Swedish municipalities. Two local contexts are studied; Swedish municipalities (local environments) and city districts (local milieus). The empirical questions are whether context a) has an independent effect on the probability of the individual to involve in political action and b) whether the municipal-level and the city district-level produce different effects on the individual-level outcome. The dependent variable, political participation, is defined broadly; as voting in local/national elections, contacting local officials, taking part in manifestations and signing petitions. The data is analyzed using multilevel modeling in MlwiN.