Conference Paper American Political Science Association Annual Meeting, 1- 4 September 2005, Washington, DC
Does civic participation contribute to political engagement?: Using fixed effects panel analyses to investigate correlations and causation
Does civic participation contribute to political engagement and, if so, through which mechanisms? Data from the British Household Panel Survey (BHPS), which consists of thirteen recent waves of observations of over 10,000 individuals, is used to address this question. This study first tests whether observed commonly cross-sectional relationships between civic and political engagement are potentially causal, as opposed to both driven by an individual’s particular fixed personal characteristics. It finds that they often may not be, as they are often not robust to fixed effects panel analyses. This implies that commonly used cross-sectional analyses of the relationship between civic and political engagement may sometimes be misleading and should be interpreted with caution. Secondly, where the relationship between civic and political engagement is robust, this study attempts to determine the likely type of causal mechanism. ‘Situational effects’ arise from changes in self-interest, information flows, and other highly context-dependent factors, and they may lead to more immediate and temporary changes in political behavior. ‘Educative effects’ arise from long-term changes in civic-political skills, experience, confidence, social capital, civic-mindedness, and other internalized factors, and they may lead to more gradual and enduring changes. Voting propensity seems to be increased via situational effects associated with joining certain types of civic groups, such as religious groups. Political interest is also apparently affected by situational factors associated with membership in issue oriented, public service oriented, and private welfare oriented groups. External political efficacy seems to be raised through both situational and educative means, particularly through issue-oriented groups.