Flexibility in American religious life: an exploration of loyalty and purity

For an example of the standard approach in a top-quality and widely used national survey, see: Schaffner and Ansolabehere (2015), detailing the questions in the common content of the CCES. The common content includes: a question about the frequency of church attendance, a question about the frequency of individual prayer, a general question about religious affiliation (“Jewish” “Muslim” “Hindu” “Roman Catholic” etc.), and a series of follow-up questions to identify the exact denomination of Protestants (“Lutheran Church, Missouri Synod” vs. “Lutheran Church, Wisconsin Synod”—for example), and similar follow-up questions for other groups (“Sunni” vs. “Shia” Muslim—for example). This is part of the list of standard control variables; researchers can add others (as control variables, or outcomes of interest), but numerous studies simply will make use of the broad denomination categories.

Jenkins-Smith et al. (2014, 491) raise a similar point about the way two-party elections force “multiple dimensions of conflict onto a single dimension of competition and choice.” Our observation here is primarily statistical: studies using this reduced dimension outcome as the key dependent variable are necessarily limited in their ability to explore future coalitions; particularly in regard to the “values voter” (as in Dillon 2014), this can miss much of the variation we explore.

An unusual alternative is the 2007 Faith Matters Study used in Putnam and Campbell (2010); for documentation, see: http://​www.​thearda.​com/​Archive/​Files/​Descriptions/​FTHMATT.​asp (last accessed 04/25/16).

For example, the national Republican Party platform for 2000 included the following on issues relating to sexual purity: “We renew our call for replacing ‘family planning’ programs for teens with increased funding for abstinence education, which teaches abstinence until marriage as the responsible and expected standard of behavior. Abstinence from sexual activity is the only protection that is 100% effective against out-of-wedlock pregnancies and sexually transmitted diseases, including HIV/AIDS, when transmitted sexually. We oppose school-based clinics that provide referrals, counseling, and related services for contraception and abortion. We urge the states to enforce laws against statutory rape, which accounts for an enormous portion of teen pregnancy. We support the establishment of Second Chance Maternity Homes, like the ones Governor Bush has proposed, to give young unwed mothers the opportunity to develop parenting skills, finish school, and enter the workforce. Because many youngsters fall into poverty as a result of divorce, we also encourage states to review their divorce laws and to support projects that strengthen marriage, promote successful parenting, bolster the stability of the home, and protect the economic rights of the innocent spouse and children. Finally, because so many social ills plaguing America are fueled by the absence of fathers, we support initiatives that strengthen marriage rates and promote committed fatherhood.” Archived at: http://​www.​presidency.​ucsb.​edu/​ws/​?​pid=​25849 (last accessed 04/25/16).

This sort of mass public opinion was of interest to Lasswell also, given his work on propaganda (Lasswell et al. 2003).

Support for political parties can also be treated, at least for individuals with some other over-riding identity, as a lower-tiered policy belief; Lugg and Robinson (2009) point out that many Evangelical Christians supported Democrat Jimmy Carter in 1976 only later to switch to Republican presidential candidates. Lugg and Robinson also note how changes in demographics can cause tactical shifts: “But beginning in the 1990s, some leaders of the Protestant Right began to engage in ‘racial reconciliation’ in hopes of finding common ground with African American Protestants with whom they had much in common theologically, if not politically” (2009, 260–261). Lugg and Robinson do not make an effort to formalize the concept of a hierarchy of beliefs in their paper, although this is certainly in line with what we develop.

“Neoclassical theory simply skips this step under the assumption that people know what they are doing. This may be true in evaluating opportunity costs at the supermarket; but it is wildly incorrect when it comes to making more complicated choices in a world of complex problems and incomplete information…” (North 1998, 19).

In much of formal political theory, individuals are given endowments (resources) and preferences (conceived as utility over available choices). Unlike in areas of economic policy in which preferences may be more readily assumed and stable (more money is usually better), dealing with the formation, ordering, and expression of preferences in social policy requires more investigation. Development of these preferences is not likely wholly random: human beings function in social settings and have evolved successful mechanisms to protect and sustain societies; the human mind is organized “…in advance of experience so that it is prepared to learn values, norms, and behaviors related to a diverse set of recurrent adaptive social problems” (Graham et al. 2012).

We exclude the church names and precise locations to protect the anonymity of the respondents (including, where cited, the ministers).

Our target population is the number of people who regularly attend the congregation. This makes calculating a response rate to be a complicated procedure, since the churches themselves may think about this number differently and may not have an exact number. See the Technical Appendix for more information. As a practical matter, the responses from the Catholic churches are more akin to traditional survey samples (ranging from 1 to 20% of the total parish—which is very difficult to estimate); for some of the Protestant churches the same holds, although for some we obtained near population-level data (in one case, 97% of the estimated regular attendance).

A graphical illustration of the changes in predicted probabilities for each response category is contained in supplemental Figure 2, displayed for the Wealthy Catholic congregation type only. The figure included here in the main section of the document conveys the important point that the movement only takes place with the Catholics.

Those who do not report looking to Scripture to solve conflicts: Wealthy Catholic, 76%; Diverse Catholic, 59%; Mainline Protestant, 71%; Black Protestant, 59%; Evangelical Protestant, 23%. While that is roughly what one might expect, even among Evangelicals, a quarter do not look to scripture to solve conflicts—and the Mainline Protestants are nearly as scripture-averse as the Wealthy Catholic congregation.

For additional information on the loyalty scale, see the Appendix—in particular the distribution of responses by type in Fig. 3a–e. The scale seeks to capture a variety of notions of loyalty; on the most direct question about loyalty to leader authority in case of conflict, very few respondents of any type responded affirmatively. Only 13% of the Wealthy Catholic congregation, 34% of the Diverse Catholic congregations, 3% of the Mainline Protestants, 18% of the black Protestants, and 12% of the Evangelical Christians indicated that in case of conflict they accept the authority of their leaders.

Mainline Protestants: 16%; Wealthy Catholic Church: 18%; Black Protestant Church: 31%; Diverse Catholic Church Respondents: 36%; and Evangelical Protestants: 61%.

The precise coefficients, levels, and significance for each of these vary somewhat from the results of the main regression reported in Table 2. There is enough variability in the question about providing condoms to high school students that the interaction effects for the two Catholic groups, although in the same direction (resulting in the predictions in the graph) as the others, do not reach conventional levels of significance. The interaction effects are significant at the 0.05 level in the homosexuality question. The point here is that the purity scale builds up all of these effects by adding the policy positions together and then the aggregate trend is a bit clearer.

To explain differences in levels: we have near uniform emphasis among the black Pentecostal Church and the Evangelical churches that homosexuality “seems unnatural” and is “against what God wants of human beings” over an alternative statement about equal rights. Indeed, for both groups, 95% + of the respondents focused on one answer. For the question about condoms in high school, only 60% of Evangelical Protestants agreed and only 22% of the black Pentecostal church agreed. But in both cases, the lines are in effect flat.

Personal email to author, November 4, 2016.