Abstract
Typically the relationship between civic engagement and religion has been investigated using only basic measures of religious attendance or denomination. We utilize more detailed indicators of various types of religious participation and a more valid concept of religious traditions to examine the influence on civic engagement, finding that typical measures of religious attendance or denomination camouflage much of religion's influence. We show that while several religious traditions are positively related to civic engagement, being an Evangelical Protestant or Black Protestant is negatively related. The results also indicate that religious attendance reduces overall civic engagement, while other types of religious participation increase civic engagement. At the specific levels of civic engagement such as belong, contribute, volunteer, and lead, we find similar variations in religious traditions and types of religious participation. Our findings suggest that scholars often overlooked the importance of religious participation beyond religious attendance and reached conclusions regarding civic engagement that do not accurately describe religion's effects. Thus, we contend that attendance and religious tradition, while they may be important and should be included in analysis, should not be the sole, or even primary, focus.
Notes
1Religious attendance is used as a generic term indicating participation in religious services for all faiths.
2A debate exist about the reliability of measures of religious attendance reported in national surveys (e.g., Demerath Citation2002; Hadaway et al. Citation1993; Hout and Greeley Citation1998). From the position of our research effort, such an exchange may be methodologically interesting, but based on the more nuanced measures of religious behaviors and beliefs now available, we agree with Iannaccone and Everton (2004, p. 204) that the narrow focus on religious attendance may be “unwarranted and unnecessary.”
3Previous studies were often limited by data on only specific churches (Becker and Dhingra Citation2001; Brewer et al. Citation2003; Greenberg Citation2000; Edgell and Docka Citation2007; Verba et al. Citation1995), quantitative studies with small samples (Guth et al. Citation2003; Williams Citation2002), single variables of religious attendance (Houghland and Christenson Citation1983), a single denomination (Uslaner Citation2002), a single city (Glanville Citation2004), or dichotomous variables of volunteerism (Park and Smith Citation2000). Nevertheless, these studies are informative and suggestive. The lack of comprehensive research with national samples of religious traditions that contain multiple types of civic engagement does not allow robust assessment of religion's effects on civic engagement.
4See Bader, Froese, and Mencken (2007) for a detailed description of this survey.
5The actual range extends to 156.
6We conducted a reliability test to empirically determine that the scale of religious participation holds together in a single measure. It is internally consistent with Cronbach's alpha score of .64.
7Pearson's coefficients of the religious variables are correlated at .62 for religious attendance and religious participation; at .60 for religious attendance and religious involvement; and at .48 for religious participation and religious involvement.
∗Significant at the .05 level, ∗∗significant at the .01 level, ∗∗∗significant at the .001 level.
∗Significant at the .05 level, ∗∗significant at the .01 level, ∗∗∗significant at the .001 level.
8Pearson's correlation coefficients of civic engagement (DV) and the religious variables (IV) indicate that religious attendance and civic engagement = .24; religious participation and civic engagement = .26; and religious involvement and civic engagement = .57.
9They typically do not include a national sample, but rather utilize a local, regional, or convenience sample (see Becker and Dhingra Citation2001; Edgell and Docka Citation2007); or do not employ Steensland and colleagues' (2000) religious traditions categories, but rather include only a single category or denomination (Uslaner Citation2002); or do not include measures of civic engagement, but rather include measures of volunteerism (Park and Smith Citation2000; Uslaner Citation2002).