Over the past three decades there has been renewed interest by sociologists and political scientists in understanding the impact of religion on American politics and policies. Undoubtedly this interest was ignited by Jerry Falwell's founding of the Moral Majority and Pat Robertson's founding of the Christian Coalition, and the pundits' claims that the Religious Right played a decisive role in electing Ronald Reagan president in 1980. What added intellectual fuel to this scholarly endeavor was the argument that a new “Culture War” divides Americans according to whether their religious beliefs are “orthodox” or “progressive” (Hunter Reference Hunter1991; also see Wuthnow Reference Wuthnow1988). Although there is scholarly evidence that calls this thesis into question (Williams Reference Williams1997; Davis and Robinson Reference Davis and Robinson1996), the fact remains that religious values and cultural conflicts now affect both American political parties (Layman Reference Layman2001), with “conservative” Protestants shifting toward the Republican Party (Kellstedt and Noll Reference Kellstedt, Noll, Noll and Mark1990).

In 1972, the National Opinion Research Center (NORC) began its General Social Survey (GSS). The seminal work by then NORC co-director Tom Smith (Reference Smith1990) classified 154 religious denominations, with Protestants categorized into “liberal” or “moderate” or “fundamentalist” denominations. But researchers soon collapsed Smith's “moderate” and “liberal” categories into “mainline” Protestants to differentiate them from “evangelical” Protestants (see Leege and Kellstedt Reference Leege and Kellstedt1993). A substantive improvement on Smith's scheme came when Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000) further classified the GSS denominations as mainline or evangelical (white) Protestant but also as black Protestant, Roman Catholic, Jewish, and others. Today Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000) is considered the “gold standard” for denominational classifications.

Unlike survey research, students of morality policy utilize state-level or county-level aggregate data on religious denominations to assess their political and policy impacts. These studies take advantage of the denominational statistics that have been periodically estimated since the 1950s by the Glenmary Research Institute (e.g., see Jones et al. Reference Jones, Doty, Grammich, Horsch, Houseal, Lynn, Marcum, Sanchagrin and Taylor2002). Although use of the Glenmary database is common among policy analysts, rarely do they include the entire denominational population that belongs to “conservative” Protestantism, as did Camobreco and Barnello (Reference Camobreco and Barnello2008, 63) in tracking state abortion policies over three decades or as did Pierce and Miller (Reference Pierce, Miller, Mooney and Christopher2001) in showing their opposition to the adoption of state lotteries. More often narrower classifications are derived for purposes of analysis.

In his pioneering research on morality policy, Kenneth Meier included 13 “Protestant fundamentalist” religions and other non-fundamentalist but anti-drink churches to analyze state alcoholic beverage policies in 1985 (Meier and Johnson Reference Meier and Johnson1990, 426). Following Meier's lead, later researchers included the same 13 “Protestant fundamentalist” churches to study abortion policies in the 50 states (Mooney and Lee Reference Mooney and Lee1995, 623; Meier and McFarlane Reference Meier and McFarlane1993, 99), county-level votes on abortion referenda (Roh and Berry Reference Roh and Berry2008, 82; Roh and Haider-Markel Reference Roh and Haider-Markel2003, 22), county-level votes on gay rights referenda (Haider-Markel and Meier Reference Haider-Markel and Meier1996, 338), and hate crime enforcement in Unite States cities (Haider-Markel Reference Haider-Markel2002, 157).

This entire morality policy research agenda, based as it is on aggregate data, presumes that the size of conservative religious denominations, whether Protestant or Catholic depending upon the issue, is a vitally important constraint on politics and policymakers. Including only a limited number of denominations, as has been the norm in most morality policy research, raises the possibility that our understanding of the political and policy impacts of conservative Protestantism has been affected by this measurement decision.Footnote ¹

VALIDATING DENOMINATIONAL CLASSIFICATIONS

Religious Congregations and Membership in the United States, 2000 (RCMS), collected by the Association of Statisticians of American Religious Bodies (ASARB) and distributed by the Association of Religion Data Archives (www.theARDA.com), is essentially a decennial census of religious congregations that is published by the Glenmary Research Institute (these data can be found at http://www.thearda.com/mapsReports/RCMS_Notes.asp). ASARB categorized the 150 RCMS denominations as Evangelical Protestant, Mainline Protestant, Catholic, Orthodox, and Other. To differentiate between Evangelical and Mainline Protestant denominations, ASARB first accepted any of the RCMS denominations so classified by Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000). For other denominations, ASARB relied on the historical documentation in Melton (Reference Melton2009) as well as Mead and Hill (Reference Mead and Hill2001).

Whereas Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000) aggregated individual-level responses to the General Social Survey (GSS), the RCMS reports actual counts of religious adherents throughout the country. For that reason, RCMS contains more denominations than the GSS. For those denominations not listed in the GSS, RCMS utilized the same types of historical and doctrinal analysis employed by Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000) to classify Protestant denominations as Mainline, Evangelical, or Other. RCMS does not list Black Protestants because its data collection did not include the individual-level data on race available to Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000).

Smith (Reference Smith1990) originally coded denominations according to (1) prior classification systems, (2) membership in ecumenical associations, (3) surveys of denominational members, (4) surveys of denominational clergy, and (5) theological beliefs. The primary criticism by Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000, 292) was that Smith (Reference Smith1990) failed to “capture essential historical differences between American religious traditions” such as the distinctiveness of Black Protestantism. The Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000, 297) scheme was “more grounded in the development of American religious traditions” and, therefore, classified “religious groups by affiliation rather than by ideology” by looking at “denominational creeds” and “denominational membership” in national religious organizations. Both Steensland et al (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000) and RCMS emphasized denominational history and membership in religious associations as being more determinative than doctrinal statements. In sum, Smith (Reference Smith1990), Steensland et al. (Reference Steensland, Park, Regnerus, Robinson, Wilcox and Woodberry2000), and the RCMS relied on a variety of religious sources to derive their classification systems, but none relied exclusively on any measure of Biblical inerrancy. Therefore, our replication offers a true validation of the RCMS classification of Mainline and Evangelical Protestant denominations by applying a Biblical “inerrancy” Scale that was independently constructed by Mark Chaves (Reference Chaves1997). It ranges from 0 to 4 as follows:

We combined levels 3 and 4, and then derived the rate of adherence (numbers of adherents per 1000 population) in each county for level 3/4, level 2, level 1, and level 0 denominations. Although level 4 explicitly embraces inerrancy whereas level 3 mentions both inerrancy and infallibility, as a practical matter research shows that most people cannot differentiate between those two Biblical authorities (Jelen Reference Jelen1989). To derive our classification system, we used the Chaves codingFootnote ² for 42 of the denominations (labeled “Chaves” in Table 1), used Lindner (Reference Lindner2006) for another 32 and Mead and Hill (Reference Mead and Hill2001) for 11 more, and finally checked the websites for 52 denominations to ascertain their statements of belief. Those classifications derived from Lindner (Reference Lindner2006) or Mead and Hill (Reference Mead and Hill2001) typically were based upon one of two conditions: either the denomination was described as being fully in communion with another denomination for which we had a Chaves classification or the denomination was a precursor or a successor denomination to a denomination classified by Chaves. These sources are labeled in Table 1 as either “Lindner” or “Mead & Hill” classifications. In the 52 cases where we had to research their websites (“Website” in Table 1), we looked for a belief statement that indicated their stance on the Bible, information provided by virtually every denomination.Footnote ³ After these exhaustive coding procedures, only 11 of the 150 denominations remain unclassified (see Table 1).

Table 1.

Classification of Protestant, Catholic and Other Denominations and Churches

When we correlated the Chaves “inerrancy” rates of adherence against our Evangelical, Mainline Protestant, and Catholic adherents by county, these simple correlations were derived: 0.976 between level 3/4 and the RCMS Evangelical rate of adherence, 0.979 between level 1 and the RCMS Mainline Protestant variable, and 0.999 between level 0 and the RCMS Catholic variableFootnote ⁴ (when calculated for the 50 states, even stronger correlations were found). This is strong evidence for the validity of the RCMS classification system for Evangelical and Mainline Protestant denominations.

What about Ken Meier's 13-denominational measure of fundamentalist Protestants that had been employed in previous studies? It is also highly correlated with the Chaves “inerrancy” level 3/4 rates of adherence (r = 0.954), principally due to the high correlation at the county level between level 3/4 adherence and the rate of adherence of the largest Evangelical Protestant denomination, Southern Baptists (r = 0.926). (The correlation for 50 states is similar.) But we urge caution when doing county-level analysis within states, because the correlations between Meier's 13-denominations and the Chaves “inerrancy” Scale drop in most within-state county-level analyses, sometimes markedly so.Footnote ⁵

In conclusion, Evangelical Protestants are especially important to studies of morality politics, and this analysis indicates that the distinction between Evangelical and Mainline Protestants is empirically valid, whether based upon a close Biblical reading of doctrinal stance or on denominational family histories and membership in broad religious associations. The readily available RCMS classification is very highly correlated with the Chaves classification based on Biblical inerrancy. Comparison of these two classifications with a third commonly used measure, which is based on a smaller number of Evangelical Protestant denominations, shows that the Ken Meier classification performs much better in a national analysis than in single state analyses. Going forward, therefore, researchers seeking to use RCMS county or state level data on adherence rates can safely rely on these data in order to continue building a common and valid knowledge base in this research area.