Is There an Attendance Effect? Examining the Causal Link Between Religious Attendance and Political Participation

Abstract

Does the act of attending religious services “cause” individuals to participate in politics? There is no known literature that examines this question using longitudinal, individual-level data. Therefore, using the Youth Parent Socialization Panel Study, this analysis examines three theoretical possibilities: the indirect, direct, and null relationships. The results show that changes in religious attendance are primarily indirectly linked to political participation through civic activity, a factor highly correlated with political participation. There is also some limited evidence for a direct effect. As individuals increase their political participation over time, they are slightly more likely to participate in political activities and vote. But, the findings also imply that the previous literature has likely overstated the role of religious attendance in generating political participation. Once individuals start participating politically, they continue to do so regardless of changes in their attendance at places of worship.

Keywords

religious attendance political participation civic activity voting Youth Parent Study

Introduction

Does the act of attending religious services “cause” political participation? Over an individual’s lifecycle, are increases in religious attendance associated with increases in political participation and decreases in religious attendance associated with decreases in political participation? Regardless of religious tradition, scholars regularly refer to the positive political consequences from attending religious services, such as building civic skills, generating social capital, accumulating information, and creating social networks that connect individuals to the wider community (e.g., Brown & Brown, 2003; Calhoun-Brown, 1996; Harris, 1994; Jamal, 2005; Jones-Correa & Leal, 2001; Liu, Wright Austin, & Orey, 2009; Macaluso & Wanat, 1979; Milbrath & Goel, 1977; Peterson, 1990, 1992; Putnam, 1995, 2000; Rosenstone & Hansen, 1993; Verba, Schlozman, & Brady, 1995; Wielhouwer, 2009; Wolfinger & Rosenstone, 1980). In turn, these factors lead to a greater inclination for individuals to participate politically (see Wielhouwer, 2009, for a review of the literature). In sum, Wielhouwer (2009) writes “immersion in a congregation where politics is a topic discussed openly . . . appears to be baptism into political activism” (p. 415) because places of worship “can be a resource for information acquisition and serve as crucibles for developing civic skills useful in politics” (p. 409).

Although the association between religious attendance and political participation is well established by the literature, no known works directly explore the relationship using individual-level longitudinal data. Several scholars have explored the influence of religious denominational changes on changes in political orientations and behaviors (Green & Guth, 1993; Jennings & Stoker, 2007), but this literature is not specifically designed to understand the role of religion in shaping political participation. In addition, Gerber, Gruber, and Hungerman (2008) and Teixiera (1992) use aggregate-level data instead of individual-level data to imply a causal link between religious attendance and political participation.¹

As a consequence of the gap in the literature, this analysis concentrates on three theoretical possibilities. First, the mechanism linking religious attendance to political participation may be indirect. Changes in religious attendance may be positively related to changes in other variables that are highly correlated with political participation, like political interest, partisanship, and civic activity. Second, there may be a direct influence of changes in religious attendance on changes in political participation, even after controlling for indirect possibilities. Last, contrary to conventional wisdom, there may simply be no influence of declining religious attendance on political participation because individuals are likely to become regular participators after entering into the political arena (Gerber, Green, & Schachar, 2003; Plutzer, 2002). For each of these possibilities, two forms of political participation are examined: First, a change in political participation index, which summarizes changes in several electoral and campaign activities, and second, a change in presidential voting participation variable, which examines the continuation or stoppage of voting in presidential elections over time.

As a preview, using data from the Youth Parent Socialization Panel Study (YPS Study), the results provide evidence that religious attendance is related to civic activity, a variable that is highly correlated with political participation. In addition, there is also evidence that changes in religious attendance positively correlate with changes in political participation, implying the possibility of a direct “causal” process. However, the substantive influence of religious attendance on political participation is much smaller than previously thought.

The Indirect Relationship

Religious attendance may shape other factors that are highly correlated with political participation. The spillover effect theory proposed by Peterson (1990, 1992) posits that participation in one environment generates participation in other arenas. For Peterson, the spillover effect is an indirect process in which religious attendance translates into political activity through generating qualities and resources advantageous to political participation, such as increased civic engagement (Peterson, 1990). Following this line of thought, other works, namely the writings by Verba et al. (1995), suggest that places of worship help individuals build important civic skills necessary for political participation.² For example, Brady, Verba, and Schlozman (1995) write that “church provides opportunities to acquire the resources relevant to political activity” (p. 276; also see Verba et al., 1995). Others, like Putnam (1995, 2000), are also strong advocates of an indirect theoretical process, claiming that religious attendance helps generate social capital, which leads to positive democratic consequences like political participation (also see Brown & Brown, 2003; Liu et al., 2009).

Three indirect possibilities are considered here. Religious attendance may influence political interest, strength of partisan attachments, and/or civic activity. As religious services provide political and civic exposure to individuals, each of these variables is positively correlated with political participation (Brady et al., 1995; Peterson, 1990; Rosenstone & Hansen, 1993; Verba et al., 1995; Wolfinger & Rosenstone, 1980). First, places of worship are political, calling on members to participate in politics and to adopt the political orientations expected by other members of the congregation and the religious leadership (Wald, Owen, & Hill, 1988). For example, starting in the 1970s, the new Christian right’s growing connection between conservative, Republican politics, and Christian fundamentalism motivated political interest among religiously active individuals (Miller & Wattenberg, 1984).³ Generally, political churches in a wide variety of contexts have served as a force to generate political interest (Calhoun-Brown, 1996; Harris, 1994; Jamal, 2005).

Second, since the late 1970s, religious attendance has been clearly linked to measures of party identification (e.g., Layman, 1997, 2001). But religious attendance does not only drive Republicanism, but rather, it is likely related to partisan tendencies more generally. For example, Black Protestants are one of the most religiously devout and active groups; yet, they are strong Democrats. Similar patterns as those for Black Protestants hold for some Latino Christians as well. Therefore, for many Christian racial minorities, going to church does not generate Republican tendencies, but it likely strengthens their established partisan attachments (Calhoun-Brown, 1996). As a result, when individuals are invested in who wins elections, either through strong political party attachments or interest in politics; they are more likely to participate in politics.

Finally, religious attendance might also be related to civic engagement; places of worship likely connect individuals to the wider community (Peterson, 1990, 1992; Putnam, 1995, 2000; Zukin, Keeter, Anderson, Jenkins, & Delli Carpini, 2006). The social networks, connections, and capital built at places of worship may influence individuals’ ties to their community, leading individuals to participate civically, and in turn, politically.

From this brief discussion, three expectations can be expressed as follows:

Hypothesis 1: Individuals who increase their religious attendance over time should be more likely to be interested in politics than individuals who decrease or maintain the same level of religious attendance over time.

Hypothesis 2: Individuals who increase their religious attendance over time should be more likely to strengthen their partisan attachments than individuals who decrease or maintain the same level of religious attendance over time.

Hypothesis 3: Individuals who increase their religious attendance over time should be more likely to be civically active than individuals who decrease or maintain the same level of religious attendance over time.

Each of these relationships will be tested independently of the others.

The Direct Relationship

Emerging from the socio-economic, participation models of the 1970s and 1980s, several studies find a direct relationship between religious attendance and political participation (Houghland & Christenson, 1983; Macaluso & Wanat, 1979; Rosenstone & Hansen, 1993; Wolfinger & Rosenstone, 1980).⁴ For example, Rosenstone and Hansen (1993) provide a precise estimate of the effect, stating “people who attend church every week are 15.1 percent more likely to vote in presidential elections and 10.2 percent more likely to vote in midterm years than people who never attend religious services” (p. 158).

Therefore, although scholars have shown the direct link between religious attendance and some forms of political participation, especially voting, it remains unclear if there is something special about attending places of worship more frequently that drives political participation. Even though this section is called the direct test, it is also testing to see if all of the things that are intertwined with changing the rate in which an individual goes to a place of worship are related to changes in political participation. But, the key here is that this analysis is also controlling for the other factors most associated with changes in political participation. If a direct relationship exists, there should be evidence for the following expectation:

Hypothesis 4: Individuals who increase their religious attendance over time should be more likely to increase their political participation over time than individuals who decrease or maintain the same level of religious attendance over time, even after controlling for changes in other indirect possibilities.

“Habitual Participation” (The Null)

Legally speaking, in the United States, religious and political spheres are supposed to be separate. But it is not necessarily for this reason why changes in religious attendance might not influence changes in political participation. More likely, once individuals enter into political participation they become regular, or rather “habitual,” participators (Gerber et al., 2003; Plutzer, 2002). If individuals become habitual participators, changes in religious attendance should matter little to political participation. In this situation, the null relationship will be observed. The influence of changes in religious attendance on changes in political participation over time should not be statistically significant. From this theoretical perspective, the null pattern should also be the case for the influence of changes in other personal characteristics (e.g., income, education) on changes in political participation. Finding these patterns would be the result of too little variance in changes of political participation over time to lead to a discernible statistically significant pattern. Specifically, individuals who decrease their religious attendance should not also decrease their political participation.

Method

Data

This analysis uses the YPS Study.⁵ The original youth sample of the YPS Study, high school seniors in 1965, was contacted three additional times over three decades, in 1973, 1982, and 1997. In 1965, 1973, and 1982, one or both of the parents of the youth sample were interviewed. This design allows for the construction of youth–parent pairs, the unit of analysis for this research (see Jennings & Niemi, 1968, 1974, 1981 for additional details).

The models below examine the influence of changes in religious attendance on changes in political participation as well as other political orientations and behaviors from the 1973 to 1982 and 1982 to 1997 periods, or rather, when the youth are 25 to 34 years old and 34 to 49 years old, respectively. The analysis focuses on these time periods because the original youth were high school juniors, mostly 16 to 17 years old, in 1964, meaning they were not of voting age. Data from the 1965 wave are used in the construction of several important socialization controls. All of the variables are discussed in greater detail in the next section.

Variables

There are several dependent variables used here, including ΔPolitical Interest, ΔPartisan Strength, ΔCivic Activity, ΔPolitical Participation, and ΔVote.⁶ First, ΔPolitical Interest and ΔPartisan Strength are used as dependent variables and control variables. Because these measures are ordered and discrete, the estimation technique used here is ordered-logistic regression.⁷ Very few respondents make dramatic shifts to their levels of political interest or partisanship over time. For both the 1973-1982 and the 1982-1997 periods, just fewer than 60% of the respondents maintained the same level of political interest. Like political interest, partisanship is also highly stable over time. Roughly half of the respondents do not change their partisan strength over each of the change periods studied here.

Next, ΔCivic Activity is used as a dependent variable and a control variable. The YPS Study asks respondents about several organizations in which they were actively engaged, including church organizations, fraternal groups, neighborhood clubs, sports teams, informal clubs, business groups, service groups, racial organizations, veterans groups, political issue advocacy groups, and other civic and cultural organizations.⁸ Respondents could indicate that they were either not a member, not very active, fairly active, or very active in an organization.⁹ From this list, an additive index of the activity level for all the possible organizations, including if respondents indicated they were in an organization not specifically listed, was calculated and z-score standardized.¹⁰ Then, the civic activity indices from each wave of the study were used to generate the change variables. Distributional analysis reveals limited variance in civic activity over each of the change periods. Most individuals only make small changes to their civic activity over time. Models that estimate ΔCivic Activity are ordinary least squares (OLS) models because the measures are interval and continuous.

Third, ΔPolitical Participation is only used as a dependent variable.¹¹ The measures are constructed in a similar manner as ΔCivic Activity. The YPS Study includes several questions that gauge respondents’ political participation. These factors are combined into additive political participation indices for each wave. These additive variables combine several, conventional measures of electoral and participatory activities, including nine different political acts such as whether respondents (a) tried to persuade others, (b) went to a political event, (c) worked for a political campaign, (d) wore a button or sticker, (e) gave money, (f) wrote a political letter to a public official or (g) newspaper, (h) took part in a demonstration, and (i) worked to solve community problems.¹² Using z-score standardized versions of these additive participation indices, variables summarizing changes in political participation over time were constructed. Like for civic activity, many respondents ether maintain the same level of political participation over time or make only very slight changes to their level of political participation over time. Models that estimate ΔPolitical Participation are OLS models because the measures are interval and continuous.

Last, ΔVote is only used as a dependent variable. The variables are constructed by examining whether respondents voted in the presidential election immediately preceding the wave of the study. For the 1973 wave, this is the 1972 election. For the 1982 wave, this is the 1980 election. For the 1997 wave, this is the 1996 election.¹³ For both change periods, nearly 80% of the respondents did not change their presidential voting participation, supporting Plutzer’s (2002) perspective about habitual voting (and non-voting). Models that estimate ΔVote are ordered-logistic regression models because the measures are discrete and ordered.¹⁴

The primary independent variable is ΔReligious Attendance. It is measured on an ordinal scale and based on changes across four categories, “never/no religious preference,” “a few times a year,” “once or twice a month,” and “almost weekly or more.” Like many of the political measures, respondents do not change their religious attendance over their lifecycle. Just fewer than 60% of the respondents maintained the same level of religious attendance over each of the change periods studied here.¹⁵

The models also include several additional control variables. First, ΔPolitical Interest, ΔPartisan Strength, and ΔCivic Activity are included as controls when estimating the ΔPolitical Participation and ΔVote models. To determine the direct influence of ΔReligious Attendance on ΔPolitical Participation or ΔVoting, it is important to rule out these indirect possibilities. Second, other socio-economic variables typically related to political participation, such as ΔEducation, ΔIncome (z-score standardized), race (a dummy variable for Whites), and gender (a dummy variable for females) are also included in the models.¹⁶ In addition, two other life-event controls are included in the models, getting married and change in home ownership.¹⁷ These events are theorized to generate participation by attaching individuals to their community. Third, several socialization measures, the parent’s level of political activity as well as their educational attainment in 1965 and the youth’s number of civic courses as well as their political activity from high school in 1965 are also included in the models as controls. It is believed that these factors may influence respondents’ propensity to change their political participation overtime.¹⁸

Last, each of the initial wave values for all independent variables as well as the dependent variables of the models is included as controls.¹⁹ Individuals with very high rates of religious attendance, education, income, political interest, or any of the other variables in the model cannot increase their attendance, education, income, interest, or the like (a similar situation is true for individuals who have low values ability to decrease their attributes). This does not mean that high (or low) values on these variables at the start of the change period do not correlate with ΔPolitical Participation or ΔVoting over time.²⁰ For example, it could be the case that religious attendance might become more correlated with political participation over time, but changes in an individual’s religious attendance could matter little. This would be consistent with the perspective of several scholars who suggest that places of worship have become more effective institutions of political mobilization over time (see Wielhouwer, 2009).

Results and Discussion

This section proceeds by discussing the findings, providing an overarching discussion of the implications, and in closing, addressing some limitations of this research.²¹

Indirect Effects

Table 1 shows the relationships between ΔReligious Attendance and ΔPolitical Interest, ΔPartisan Strength, and ΔCivic Activity. The results clearly demonstrate that ΔReligious Attendance does not significantly predict ΔPolitical Interest or ΔPartisan Strength, providing no support for Hypothesis 1 or 2. However, there is a clear correspondence between ΔReligious Attendance and ΔCivic Activity, providing evidence for Hypothesis 3. Individuals who increased their religious attendance from 1973 to 1982 as well as from 1982 to 1997 were more likely to increase their civic activity over these same time periods than individuals who decreased or maintained the same level of religious attendance.

Table 1.

Testing the Indirect Effects—ΔPolitical Interest, ΔPartisanship, and ΔCivic Activity.

	Ordered-logit	Ordered-logit	OLS		Ordered-logit	Ordered-logit	OLS
	Model 1: ΔInterest, 1973-1982	Model 2: ΔPartisan, 1973-1982	Model 3: ΔCivic, 1973-1982		Model 4: ΔInterest, 1982-1997	Model 5: ΔPartisan, 1982-1997	Model 6: ΔCivic, 1982-1997
Political Interest 1973	−1.85***	—	—	Political Interest 1982	−1.82***	—	—
Political Interest 1973	(0.13)			Political Interest 1982	(0.12)
Partisanship 1973	—	−1.29***	—	Partisanship 1982	—	−1.34***	—
Partisanship 1973		(0.07)		Partisanship 1982		(0.08)
Civic Activity 1973	—	—	−0.70***	Civic Activity 1982	—	—	−0.59***
Civic Activity 1973			(0.04)	Civic Activity 1982			(0.04)
Religious Attendance 1973	0.10	0.05	0.20***	Religious Attendance 1982	0.18***	0.07	0.17***
Religious Attendance 1973	(0.06)	(0.06)	(0.03)	Religious Attendance 1982	(0.07)	(0.06)	(0.03)
ΔReligious Attendance 1973-1982	0.12	0.02	0.21***	ΔReligious Attendance 1982-1997	0.09	0.06	0.27***
ΔReligious Attendance 1973-1982	(0.08)	(0.07)	(0.03)	ΔReligious Attendance 1982-1997	(0.08)	(0.08)	(0.03)
Income 1973	0.10	−0.08	0.08*	Income 1982	0.07	0.02	0.01
Income 1973	(0.08)	(0.09)	(0.04)	Income 1982	(0.08)	(0.08)	(0.03)
ΔIncome 1973-1982	0.03	−0.02	0.10***	ΔIncome 1982-1997	−0.02	0.09	0.04
ΔIncome 1973-1982	(0.06)	(0.06)	(0.03)	ΔIncome 1982-1997	(0.07)	(0.07)	(0.03)
Education 1973	0.48***	0.11	0.13**	Education 1982	0.21*	0.04	0.13**
Education 1973	(0.11)	(0.13)	(0.06)	Education 1982	(0.12)	(0.11)	(0.05)
ΔEducation 1973-1982	0.27	−0.08	0.17**	ΔEducation 1982-1997	0.37	0.00	0.11
ΔEducation 1973-1982	(0.17)	(0.15)	(0.08)	ΔEducation 1982-1997	(0.27)	(0.26)	(0.11)
Married 1973	−0.25	−0.20	0.12	Married 1982	−0.04	0.42*	−0.19
Married 1973	(0.19)	(0.21)	(0.10)	Married 1982	(0.21)	(0.22)	(0.12)
ΔMarried 1973-1982	0.06	0.01	0.01	ΔMarried 1982-1997	0.40**	0.59***	−0.04
ΔMarried 1973-1982	(0.18)	(0.18)	(0.09)	ΔMarried 1982-1997	(0.19)	(0.19)	(0.09)
Own Home 1973	−0.26	0.27	0.05	Own Home 1982	0.02	−0.17	0.30**
Own Home 1973	(0.20)	(0.21)	(0.09)	Own Home 1982	(0.24)	(0.24)	(0.12)
ΔOwn Home 1973-1982	−0.21	0.04	0.11	ΔOwn Home 1982-1997	0.10	−0.01	0.21**
ΔOwn Home 1973-1982	(0.16)	(0.16)	(0.07)	ΔOwn Home 1982-1997	(0.20)	(0.20)	(0.11)
Parent’s Political Activity 1965	−0.02	−0.05	0.03	Parent’s Political Activity 1965	0.06	−0.08*	0.00
Parent’s Political Activity 1965	(0.05)	(0.05)	(0.02)	Parent’s Political Activity 1965	(0.05)	(0.05)	(0.02)
Parent’s Education 1965	−0.04	0.08	−0.02	Parent’s Education 1965	0.12	−0.01	0.02
Parent’s Education 1965	(0.07)	(0.07)	(0.03)	Parent’s Education 1965	(0.08)	(0.07)	(0.03)
Student’s Civics Courses 1965	0.13	−0.04	−0.01	Student Civics Courses 1965	−0.13	0.19*	−0.02
Student’s Civics Courses 1965	(0.10)	(0.10)	(0.05)	Student Civics Courses 1965	(0.10)	(0.11)	(0.05)
Student’s Political School Activity 1965	0.03	−0.01	0.06***	Student’s Political School Activity 1965	0.05	−0.02	0.07***
Student’s Political School Activity 1965	(0.04)	(0.04)	(0.02)	Student’s Political School Activity 1965	(0.04)	(0.04)	(0.02)
Female	−0.59***	0.32***	−0.15**	Female	−0.59***	0.04	−0.11
Female	(0.13)	(0.13)	(0.06)	Female	(0.14)	(0.14)	(0.06)
White	−0.25	−0.74***	−0.21	White	−0.35	−0.55**	−0.50***
White	(0.24)	(0.24)	(0.13)	White	(0.25)	(0.27)	(0.15)
Constant	—	—	−0.81***	Constant	—	—	−0.42**
Constant			(0.18)	Constant			(0.21)
n	1,130	1,085	947	n	1,107	1,067	911
χ² or F	243.07***	336.09***	28.60***	χ² or F	249.28***	361.18***	26.15***
Pseudo R² or R²	.16	.12	.41	Pseudo R² or R²	.15	.13	.35

Note. Data come from the YPS Study. Model estimation varies based on the dependent variables. All models use a sampling weight. YPS Study = Youth Parent Socialization Panel Study; OLS = ordinary least squares.

p < .10. **p < .05. ***p < .01.

In addition to the influence of ΔReligious Attendance on ΔCivic Activity, individuals who had higher levels of religious attendance at the beginning of each of these change periods were also more likely than their infrequently attending counterparts to become more civically inclined over time, potentially suggesting that places of worship became better catalysts of civic activity over the time periods studied here. The other possibility is that during these times in an individual’s lifecycle religious attendance may become more predictive of civic activity regardless of changes in the individual’s religious attendance.

How substantively important is the influence of ΔReligious Attendance on ΔCivic Activity? For every one unit increase in ΔReligious Attendance (e.g., such as going from never attending to attending a few times a year), there is a 0.21 and 0.27 (both statistically significant at p < .01) increase in ΔCivic Activity for the 1973-1982 and 1982-1997 periods, respectively. However, because ΔCivic Activity is a z-score standardized variable with a range of −4.14 to 3.52 for the 1973-1982 period and −3.18 to 3.66 for the 1982-1997 period, the direct interpretations have little value. Therefore, Figure 1 shows the estimated linear relationship between ΔReligious Attendance and ΔCivic Activity. Viewing the relationships graphically makes clear that these relationships are relatively weak. Increasing or decreasing the rate in which an individual attends religious services does not have a very large influence on changes in their civic activity. There are only small adjustments in civic activity across the range of changes in religious attendance. But, to be fair, if we think back to the distributions discussed above, given how few individuals change their civic activity and religious attendance, finding any relationship at all is impressive.

Figure 1.

The relationship between ΔReligious Attendance and ΔCivic Activity.

Taken together, there is limited evidence for the indirect explanation that suggests religious attendance is indirectly related to political participation through other variables associated with political participation. Nonetheless, the weak, statistically significant relationship between ΔReligious Attendance and ΔCivic Activity implies a causal process between religious attendance and civic activity in the United States.

Direct Effects

How do changes in religious attendance directly relate to changes in political participation? Tables 2 and 3 provide models that estimate the relationship between ΔReligious Attendance and ΔPolitical Participation and ΔVoting, respectively. Table 2 shows that when the civic activity variables are excluded from the models (Models 1 and 3 in Table 2) the ΔReligious Attendance variables are statistically significant predictors of the ΔPolitical Participation variables. However, as with the influence of ΔReligious Attendance on ΔCivic Activity, the effects are weak. In addition, when controlling for ΔCivic Activity, the influence of ΔReligious Attendance on ΔPolitical Participation is further diminished (Models 2 and 4 in Table 2). In Model 2, the influence of ΔReligious Attendance is no longer statistically significant. However, in Model 4, ΔReligious Attendance remains statistically significant, providing some support for Hypothesis 4.

Table 2.

Testing the Direct Effect of ΔReligious Attendance on ΔPolitical Participation (Index).

	Model 1: ΔPolitical Participation, 1973-1982	Model 2: ΔPolitical Participation, 1973-1982		Model 3: ΔPolitical Participation, 1982-1997	Model 4: ΔPolitical Participation, 1982-1997
Participation Index 1973	−0.63***	−0.62***	Participation Index 1982	−0.52***	−0.55***
Participation Index 1973	(0.03)	(0.03)	Participation Index 1982	(0.03)	(0.03)
Religious Attendance 1973	0.08***	0.01	Religious Attendance 1982	0.08***	0.05*
Religious Attendance 1973	(0.03)	(0.03)	Religious Attendance 1982	(0.02)	(0.03)
ΔReligious Attendance 1973-1982	0.10***	0.04	ΔReligious Attendance 1982-1997	0.12***	0.11***
ΔReligious Attendance 1973-1982	(0.03)	(0.03)	ΔReligious Attendance 1982-1997	(0.03)	(0.03)
Income 1973	−0.02	−0.02	Income 1982	0.02	0.04
Income 1973	(0.03)	(0.04)	Income 1982	(0.03)	(0.03)
ΔIncome 1973-1982	−0.01	−0.02	ΔIncome 1982-1997	0.02	0.02
ΔIncome 1973-1982	(0.02)	(0.03)	ΔIncome 1982-1997	(0.02)	(0.03)
Education 1973	0.17***	0.15***	Education 1982	0.02	−0.04
Education 1973	(0.04)	(0.05)	Education 1982	(0.04)	(0.05)
ΔEducation 1973-1982	0.16**	0.19**	ΔEducation 1982-1997	0.07	0.05
ΔEducation 1973-1982	(0.08)	(0.08)	ΔEducation 1982-1997	(0.09)	(0.10)
Married 1973	−0.03	−0.02	Married 1982	0.07	−0.01
Married 1973	(0.08)	(0.09)	Married 1982	(0.08)	(0.09)
ΔMarried 1973	−0.08	−0.11	ΔMarried 1982	0.02	−0.05
ΔMarried 1973	(0.07)	(0.08)	ΔMarried 1982	(0.07)	(0.08)
Own Home 1973	0.28***	0.19**	Own Home 1982	0.15	0.16
Own Home 1973	(0.07)	(0.09)	Own Home 1982	(0.09)	(0.11)
ΔOwn Home 1973-1982	0.28***	0.19***	ΔOwn Home 1982-1997	0.25***	0.25***
ΔOwn Home 1973-1982	(0.06)	(0.08)	ΔOwn Home 1982-1997	(0.08)	(0.09)
Civic Activity 1973	—	0.216***	Civic Activity 1982	—	0.15***
Civic Activity 1973		(0.04)	Civic Activity 1982		(0.04)
ΔCivic Activity 1973-1982	—	0.23***	ΔCivic Activity 1982-1997	—	0.15***
ΔCivic Activity 1973-1982		(0.03)	ΔCivic Activity 1982-1997		(0.03)
Political Interest 1973	0.34***	0.27***	Political Interest 1982	0.24***	0.25***
Political Interest 1973	(0.04)	(0.04)	Political Interest 1982	(0.04)	(0.04)
ΔPolitical Interest 1973-1982	0.21***	0.16***	ΔPolitical Interest 1982-1997	0.26***	0.26***
ΔPolitical Interest 1973-1982	(0.03)	(0.04)	ΔPolitical Interest 1982-1997	(0.04)	(0.04)
Partisan Strength 1973	0.09***	0.07**	Partisan Strength 1982	0.11***	0.06
Partisan Strength 1973	(0.03)	(0.04)	Partisan Strength 1982	(0.03)	(0.04)
ΔPartisan Strength 1973-1982	0.12***	0.07*	ΔPartisan Strength 1982-1997	0.15***	0.13***
ΔPartisan Strength 1973-1982	(0.03)	(0.04)	ΔPartisan Strength 1982-1997	(0.03)	(0.03)
Parent’s Political Activity 1965	0.04*	0.04*	Parent’s Political Activity 1965	0.10***	0.09***
Parent’s Political Activity 1965	(0.02)	(0.02)	Parent’s Political Activity 1965	(0.02)	(0.02)
Parent’s Education 1965	−0.02	−0.01	Parent’s Education 1965	0.01	0.03
Parent’s Education 1965	(0.03)	(0.03)	Parent’s Education 1965	(0.03)	(0.03)
Student’s Civic Courses 1965	0.08*	0.09**	Student Civic Courses 1965	−0.02	−0.04
Student’s Civic Courses 1965	(0.04)	(0.04)	Student Civic Courses 1965	(0.04)	(0.04)
Student’s Political School Activity 1965	0.05***	0.04**	Student’s Political School Activity 1965	0.05***	0.03
Student’s Political School Activity 1965	(0.01)	(0.02)	Student’s Political School Activity 1965	(0.02)	(0.02)
Female	−0.07	−0.01	Female	0.03	0.08
Female	(0.05)	(0.06)	Female	(0.05)	(0.06)
White	−0.02	0.02	White	−0.16	−0.09
White	(0.09)	(0.11)	White	(0.12)	(0.12)
Constant	−1.82***	−1.41***	Constant	−1.43***	−1.18***
Constant	(0.17)	(0.19)	Constant	(0.21)	(0.21)
n	1,083	903	n	1,067	878
F	27.98***	21.01***	F	22.75***	23.56***
R ²	.35	.38	R ²	.34	.357

Note. Data come from the YPS Study. Models are OLS linear regression. All models use a sampling weight. YPS Study = Youth Parent Socialization Panel Study; OLS = ordinary least squares.

p < .10. **p < .05. ***p < .01.

Table 3.

Testing the Direct Effects of ΔReligious Attendance on ΔVoting.

	Model 1: ΔVoting, 1973-1982		Model 2: ΔVoting, 1982-1997
Voted 1973	−22.39***	Voted 1982	−23.07***
Voted 1973	(0.54)		(0.51)
Religious Attendance 1973	0.32***	Religious Attendance 1982	0.27**
Religious Attendance 1973	(0.12)		(0.13)
ΔReligious Attendance 1973-1982	0.38***	ΔReligious Attendance 1982-1997	0.36**
ΔReligious Attendance 1973-1982	(0.14)		(0.16)
Income 1973	−0.30*	Income 1982	0.22
Income 1973	(0.16)		(0.16)
ΔIncome 1973-1982	0.05	ΔIncome 1982-1997	−0.03
ΔIncome 1973-1982	(0.02)		(0.14)
Education 1973	0.12	Education 1982	−0.08
Education 1973	(0.20)		(0.23)
ΔEducation 1973-1982	0.41	ΔEducation 1982-1997	0.44
	(0.31)		(0.79)
Married 1973	0.27	Married 1982	0.02
	0.39		(0.42)
ΔMarried 1973-1982	0.30	ΔMarried 1982-1997	0.14
	0.30		(0.39)
Own Home 1973	0.75*	Own Home 1982	0.93**
	0.40		(0.47)
ΔOwn Home 1973-1982	0.80**	ΔOwn Home 1982-1997	0.92**
ΔOwn Home 1973-1982	0.31		(0.39)
Civic Activity 1973	0.40**	Civic Activity 1982	0.91***
	(0.16)		(0.23)
ΔCivic Activity 1973-1982	0.38**	ΔCivic Activity 1982-1997	0.74***
ΔCivic Activity 1973-1982	(0.15)		(0.25)
Political Interest 1973	0.84***	Political Interest 1982	0.88***
	(0.19)		(0.18)
ΔPolitical Interest 1973-1982	0.65***	ΔPolitical Interest 1982-1997	0.80***
ΔPolitical Interest 1973-1982	(0.16)		(0.19)
Partisan Strength 1973	0.50***	Partisan Strength 1982	0.29
	(0.17)		(0.18)
ΔPartisan Strength 1973-1982	0.13	ΔPartisan Strength 1982-1997	0.60***
ΔPartisan Strength 1973-1982	(0.13)		(0.20)
Parent’s Political Activity 1965	0.16*	Parent’s Political Activity 1965	0.20*
Parent’s Political Activity 1965	(0.09)		(0.11)
Parent’s Education 1965	0.16	Parent’s Education 1965	0.14
Parent’s Education 1965	(0.14)		(0.15)
Student’s Civic Courses 1965	−0.05	Student Civic Courses 1965	−0.08
Student’s Civic Courses 1965	(0.19)		(0.22)
Student’s Political School Activity 1965	0.01	Student’s Political School Activity 1965	−0.10
Student’s Political School Activity 1965	(0.07)		(0.09)
Female	0.29	Female	0.92***
	(0.24)		(0.30)
White	−0.78	White	0.71
	(0.59)		(0.56)
n	898	n	874
χ²	2064.60***	χ²	See notes
Pseudo R²	.52	Pseudo R²	.57

Note. Data come from the YPS Study. Models are ordered-logistic regression. All models use a sampling weight. For Model 2, χ² was not generated because the inclusion of the Vote 1982 variable is highly correlated with changes in voting over the 1982 to 1997 period. YPS Study = Youth Parent Socialization Panel Study.

p < .10. **p < .05. ***p < .01.

Figure 2 shows the relationship between ΔReligious Attendance from 1982 to 1997 and ΔPolitical Participation from 1982 to 1997. The best linear prediction is only slightly positive. There are very small adjustments in political participation across the range of changes in religious attendance. However, as with the previous section, finding a statistically significant relationship at all given the distribution of both ΔReligious Attendance and ΔPolitical Participation is impressive.

Figure 2.

The relationship between ΔReligious Attendance and ΔPolitical Participation.

Although indices of political participation provide us a way to estimate changes in a general propensity of individuals to participate politically, it is also important to explore voting behavior more specifically. Therefore, Table 3 provides estimates for ΔVoting over time. Table 3 shows that individuals who already vote cannot start voting. Thus, controlling for previous voting behavior largely explains changes in voting behavior. That said, Table 3 also shows that ΔReligious Attendance significantly predicts ΔVoting in both models; however, Model 2 does not estimate properly (see the note in Table 3). Therefore, Figure 3 only shows the predicted probabilities of ΔVoting across the range of ΔReligious Attendance for the 1973-1982 period. While the change in probability of starting voting is very small across the range of ΔReligious Attendance, again, as previously suggested, it is impressive to find an influence at all given how few individuals change their voting participation over time. But, Figure 3 also shows the statistically significant differences observed are primarily for individuals who increase religious attendance, not those who decrease it.

Figure 3.

Predicted probabilities for ΔPresidential Voting across the range of ΔReligious Attendance, 1973-1982.

Discussion

First and foremost, the results suggest that we should not overstate the influence of religious attendance on political participation in the United States. While it is clear that changes in an individual’s rate of religious attendance are significantly related to changes in their political participation and presidential voting over time, it is also clear that the substantive effects of changes in religious attendance on changes in political participation and voting are very small. Previous literature has likely overstated the influence of religious attendance in generating political participation. This is largely because few people change their rate of religious attendance and political participation throughout their life. And decreases in attendance are not correlated with decreases in political participation, at least in terms of voting. Thus, the findings have strong implications for the perspective that individuals behave habitually.

However, these results should be considered with caution. First, these data are dated. It has been 17 years since the last wave of the YPS Study. This research is unable to account for the growing influence of places of worship in the American political system over the past several decades. In addition, these data do not include the growing number of religiously unaffiliated and secular individuals in America today. They also include a very few racial and religious minorities. All of these excluded Americans are important to understanding the role of religious attendance in stimulating political activity. Second, because this panel only includes four waves, a dynamic model that would account for a longer causal process between religious attendance and political participation cannot be estimated. Third, in an effort to be parsimonious, not all possible variables that influence religious attendance, political participation, and/or voting are accounted for by this analysis. It is certainly possible that future research could come to different conclusions by including a different set of variables. However, with these limitations in mind, the results remain a strong addition to the religion and politics as well as political participation literatures, primarily because they clarify the role of religious attendance in generating acts of political participation.

Footnotes

Acknowledgements

I would like to thank Kathleen Dolan, Thomas Holbrook, Michael A. Hansen, the editors of American Politics Research, as well as the anonymous reviewers for reading drafts and their valuable critiques and suggestions.

Author’s Note

An earlier version of this article was presented at the 2013 annual meeting of the American Political Science Association.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Sky L. Ammann is a recent PhD recipent from the University of Wisconsin-Milwaukee. His research focuses on the role of religion in shaping American political behavior and public opinion.

References

Brady

H. E.

Verba

Schlozman

K. L.

(1995). Beyond ses: A resource model of political participation. The American Political Science Review, 89, 271-294.

Brown

R. K.

Brown

R. E.

(2003). Faith and works: Church-based social capital resources and African American political activism. Social Forces, 82, 617-641.

Calhoun-Brown

(1996). African American churches and political mobilization: The psychological impact of organizational resources. The Journal of Politics, 58, 935-953.

Desmond

S. A.

Morgan

K. H.

Kikuchi

(2010). Religious development: How (and why) does religiosity change from adolescence to young adulthood? Sociological Perspectives, 53, 247-270.

Gerber

A. S.

Green

D. P.

Schachar

(2003). Voting may be habit-forming: Evidence from a randomized field experiment. American Journal of Political Science, 47, 540-550.

Gerber

A. S.

Gruber

Hungerman

D. M.

(2008). Does church attendance cause people to vote? Using blue laws’ repeal to estimate the effect of religiosity on voter turnout. Cambridge, MA: National Bureau of Economic Research. Retrieved from http://economics.mit.edu/files/4514

Green

J. C.

Guth

J. L.

(1993). From lambs to sheep: Denominational change and political behavior. In Kellstedt

Leege

(Eds.), Rediscovering the religious factor in American politics (pp. 100-117). Armonk, NY: Sharpe.

Harris

F. C.

(1994). Something within: Religion as a mobilizer of African-American political activism. The Journal of Politics, 56, 42-68.

Houghland

J. G.

Christenson

J. A.

(1983). Religion and politics: The relationship of religious participation to political efficacy and involvement. Sociology and Social Research, 67, 405-420.

10.

Jamal

(2005). The political participation and engagement of Muslim Americans: Mosque involvement and group consciousness. American Politics Research, 33, 521-544.

11.

Jennings

M. K.

Niemi

R. G.

(1968). The transmission of political values from parent to child. The American Political Science Review, 62, 169-184.

12.

Jennings

M. K.

Niemi

R. G.

(1974). The political character of adolescence: The influence of families and schools. Princeton, NJ: Princeton University Press.

13.

Jennings

M. K.

Niemi

R. G.

(1981). Generations and politics: A panel study of youth adults and their parents. Princeton, NJ: Princeton University Press.

14.

Jennings

M. K.

Stoker

(2007, July). Changing relationships between religion and politics: A longitudinal, multigenerational analysis. Proceedings from the annual meeting of the International Society of Political Psychology, Portland, OR.

15.

Jones-Correa

M. A.

Leal

D. L.

(2001). Political participation: Does religion matter? Political Research Quarterly, 54, 751-770.

16.

Layman

G. C.

(1997). Religion and political behavior in the United States: The impact of beliefs, affiliations, and commitment from 1980 to 1994. Public Opinion Quarterly, 61, 288-316.

17.

Layman

G. C.

(2001). The great divide: Religious and cultural conflict in the American political party system. New York, NY: Columbia University Press.

18.

Liu

Wright Austin

S. D.

Orey

B. D.

(2009). Church attendance, social capital, and Black voting participation. Social Science Quarterly, 90, 576-592.

19.

Macaluso

T. F.

Wanat

(1979). Voting turnout and religiosity. Polity, 12, 158-169.

20.

Milbrath

L. W.

Goel

M. L.

(1977). Political participation (2nd ed.). Chicago, IL: Rand-McNally.

21.

Miller

A. H.

Wattenberg

M. P.

(1984). Politics from the pulpit: Religiosity and the 1980 elections. Public Opinion Quarterly, 48, 301-317.

22.

Peterson

S. A.

(1990). Political behavior: Patterns in everyday life. Newbury Park, CA: Sage.

23.

Peterson

S. A.

(1992). Church participation and political participation: The spillover effect. American Politics Research, 20, 123-139.

24.

Plutzer

(2002). Becoming a habitual voter: Inertia, resources, and growth in young adulthood. The American Political Science Review, 96, 41-56.

25.

Putnam

R. D.

(1995). Bowling alone: America’s declining social capital. Journal of Democracy, 6, 65-78.

26.

Putnam

R. D.

(2000). Bowling alone: The collapse and revival of American community. New York, NY: Simon & Schuster.

27.

Rosenstone

S. J.

Hansen

J. M.

(1993). Mobilization, participation, and democracy in America. New York, NY: Macmillan.

28.

Teixiera

R. A.

(1992). The disappearing American voter. Washington, DC: Brookings Institution.

29.

Verba

Schlozman

K. L.

Brady

H. E.

(1995). Voice and equality: Civic voluntarism in American politics. Cambridge, MA: Harvard University Press.

30.

Wald

K. D.

Owen

D. E.

Hill

S. S.

Jr. (1988). Churches as political communities. The American Political Science Review, 82, 531-548.

31.

Wielhouwer

P. W.

(2009). Religion and American political participation. In Smidt

C. E.

Kellstedt

L. A.

Guth

J. L.

(Eds.), The Oxford handbook of religion and American politics (pp. 394-426). New York, NY: Oxford University Press.

32.

Wolfinger

R. E.

Rosenstone

S. J.

(1980). Who votes? New Haven, CT: Yale University Press.

33.

Zukin

Keeter

Anderson

Jenkins

Delli Carpini

(2006). A new engagement? Political participation, civic fife, and the changing American citizen. New York, NY: Oxford University Press.