1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References

Five competing explanations for why white evangelicals hold right-of-center political attitudes are examined using data from the 2000–2004 National Election Studies. Dependent variables include attitudes about abortion, homosexuality, immigration, national defense, and social spending. The five competing explanations accounting for conservative positions are: religiosity, moral standards traditionalism, gender and family ideology, class culture, and cultural geography. Moral standards traditionalism attenuated the evangelical effect on attitudes about abortion, homosexuality, and social spending. Religiosity and male-dominant gender ideology attenuated the effect on abortion and homosexuality only. In a second set of models, which include members of all major religious groups, these three variables, together with low levels of education, were significantly associated with conservative attitudes. Moral standards traditionalism demonstrated the most consistent, and generally the strongest, effects across dependent variables.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References

White evangelical Protestants are often described as “the base” of the Republican Party. While the link between evangelicals, conservative views on social issues, and Republican Party identifications is well established (Brooks and Manza 1997; Hillygus and Shields 2005; Langer and Cohen 2005; Manza and Brooks 1997; Wilcox and Larson 2006; Woodberry and Smith 1998), the cause remains in dispute. Other religious people have also been considered part of a conservative alliance of the “orthodox” (Hunter 1991) or the “traditionalists” (Green 2009a) both in the United States (see also Layman 2001; Wuthnow 1988) and cross-nationally (Norris and Inglehart 2004), but again the cause remains in dispute.

In this article, we use data from the 2000–2004 American National Election Studies (NES) to compare five competing explanations of the relationship between religion and conservative political views in the United States. These explanations focus, respectively, on: (1) religiosity, (2) moral standards traditionalism, (3) gender and family ideology, (4) class culture, and (5) cultural geography. We examine attitudes on two salient moral values issues—abortion and homosexuality—as well as attitudes about immigration, national defense, and social spending. We also examine self-identified ideology and Republican Party identification.

Our findings adjudicate the relative importance of these five explanations. We find that moral standards traditionalism is a particularly important explanation of conservatism on a wide range of issues, though religiosity and gender and family ideology are also important primarily because of their relationship to conservative views on abortion and homosexuality. These “values” issues have engaged religious people in recent years. Our findings indicate that many observers have overstated the direct influence of membership in evangelical denominations and others have overstated the influence of religious beliefs. Such mistaken assessments have brought membership in evangelical churches and strong religious beliefs under greater scrutiny than they deserve as foundational sources of the broader conservative movement in the United States.

Our analysis focuses on white evangelicals, but extends to members of other religious groups as well. We find that religious identity per se has little net impact on political attitudes. Instead, social circumstances and beliefs associated with membership in particular religious traditions are the underlying causes of political conservatism. The same characteristics associated with the conservatism of white evangelicals are also associated with conservatism among Catholics and mainline Protestants, but are found less often in these communities. Religious traditions do, however, have strong independent effects on Republican Party affiliation.

Competing Explanations

  1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References

Of the five competing explanations, three are particularly prominent in the literature: religiosity, adherence to traditional moral standards, and gender and family ideology. Two less prominent explanations focus on class and geographical cultures.


This explanation takes off from the observed “restructuring of American religion” in the 1960s and 1970s (Wuthnow 1988). Where Protestants, Catholics, and Jews were once separated by the doctrines of their religious communities and the “insider status” of Protestantism as compared to Catholicism and Judaism (Herberg [1955] 1983), the rise of secular liberalism created, according to Wuthnow, commonalities among strong believers in all major faith traditions.

Scholars who focused on religiosity as the arbiter of political views found a “great divide” developing between the religious and less religious and secular people on a variety of social issues, in their party identifications, and in their voting (Layman 2001; see also Davis and Robinson 1996; Green 2004; Kellstedt and Green 1993; Leege et al. 2002). Similar relationships were found throughout the developed world (Norris and Inglehart 2004), suggesting a robust connection between religiosity and political conservatism.

Social scientists have attributed the relationship between religiosity and conservative politics to both the beliefs and practices of believers. Beliefs in God, the afterlife, and the reality of evil in the world dispose Christian religious believers to hold themselves and others to high moral standards and to identify with politicians who are tough on perceived enemies of social and international order (Greeley and Hout 2006; Morone 2003). Frequent church attendance places believers in regular contact with those who dwell in the same cultural milieu, recreating ritual experiences that cement identities.

Religiosity is defined in these studies as beliefs and behaviors that express a high level of commitment to the church community and to the importance of religion over other aspects of life (Guth and Green 1993). Factor analysis is typically used to construct religiosity scales that include items on behaviors (such as frequency of church attendance and Bible reading) as well as items on beliefs (e.g., belief in the afterlife and biblical inerrancy).

Explanations for political conservatism based on religiosity suggest the following hypothesis:

  • H1: The political conservatism of white evangelical Protestants is explained by their high levels of religiosity.

Moral Standards Traditionalism

Moral standards traditionalism refers to a cognitive orientation that draws a strong contrast between right and wrong ways of living, is rooted in traditional standards, and can be threatened by social change. Social scientists have connected adherence to definitive moral standards with political conservatism since the post-World War II study The Authoritarian Personality (Adorno 1950; see also Stouffer 1955). Early studies were based on the Freudian idea that people who had been treated very strictly by their families and communities displace anger on members of outgroups. Contemporary approaches are more likely to take a cognitive-cultural approach, focusing on “moral vision” (Baker 2005) or “moral cosmology” (Hunter 1991), rather than personality traits formed in a repressive environment (see also Davis and Robinson 1996, 2001).

Baker (2005) operationalized “moral absolutism” with two items drawn from the World Values Survey about the existence of “absolutely clear guidelines about right and wrong” that “are applicable to everyone, regardless of circumstances” (2005:79). He found direct and indirect effects of moral absolutism on social attitudes related to family and life values. While moral absolutists were more religious than moral relativists, religiosity was not a highly distinguishing factor: “Moral visions are related to religious-cultural values … but these links are moderate to weak” (2005:104). Baker observed that moral codes can be based in political or social philosophies or even organizational loyalties, as much as in religion (see also Hunter 1991; Orru 1987).

Our concept of moral standards traditionalism captures an orientation that embraces traditional standards of right living, prescribes them as valid for all, and sees them as threatened by new values and new ways of life.1 Like Baker's (2005) moral absolutists, moral standards traditionalists are not relativists. They also do not believe standards should be adjusted, and they do not have an inclusive view of moral systems different than their own. However, unlike Baker, our conceptualization does not require that moral standards traditionalists base their judgments on abstract principles that provide absolutely clear guidance about right and wrong in all circumstances.

Using our conception of moral standards traditionalism, we posit the following:

  • H2: The political conservatism of white evangelical Protestants is explained by their propensity to uphold traditional moral standards and to see these standards as threatened by social change.

Gender and Family Ideology

Gender and family ideology refers to a set of beliefs and practices related to relations between spouses and between parents and children. Evangelical Protestants find support for male-dominant gender roles and strict child rearing in their reading of the Bible. The Bible refers to the husband as the spiritual “head” in marriage and refers to women as “submitting” to their husbands. These passages have been interpreted by contemporary evangelicals to maintain male-dominant gender roles, though typically without the harsh authoritarian overtones of biblical family ideology (Wilcox 2006).

Lakoff ([1996] 2002) argued that conservative and liberal worldviews “center on two opposing models of the family,” the “strict father” and “nurturing parent” models. The strict father model posits a traditional nuclear family with the father having primary responsibility for supporting and protecting the family as well as the authority to set overall policy, to define rules for the behavior of children, and to enforce rules (cf. Wilcox 2006). The mother, by contrast, has the day-to-day responsibility for the care of the household, raising children, and upholding the father's authority. Children are expected to respect and obey their parents and by doing so to build the capacity for self-discipline and self-reliance (see also Alwin 1986; Ellison and Sherkat 1993; Starks and Robinson 2007). By contrast, the nurturing parent model stresses love, empathy, and nurturance as primary. Relations between parents tend to be more egalitarian than hierarchical. In this model children become responsible, self-disciplined, and self-reliant through being cared for, respected, and by caring for others (Lakoff [1996] 2002:33–34).

These family models become translated into politics through the projection of the metaphor of the nation as family. According to Lakoff, conservatives see political leadership through the lens of the strict father model, while liberals see political leadership through the lens of the nurturing parent model. The “tough love” of conservative politicians puts a premium on controlling unnecessary spending on social welfare, but is unstinting on spending for national and domestic security. “Strict father” politicians demand law abiding citizens and encourage strict punishment of citizens who fail to comply with law. By contrast, the nurturing outlook of liberal politicians extends social rights of citizenship to all members of the community, seeks to provide respect and a minimal “social safety net” of governmental support to all citizens, and emphasizes pluralism and inclusiveness.

Two hypotheses capture the proposed relationship between gender and family ideology and political conservatism, one related to gender roles and the other to the strict disciplinary practices used in many evangelical families to socialize behavioral conformity in children.

  • H3: The political conservatism of white evangelical Protestants is explained by their tendency to support male-dominant gender roles.

  • H4: The political conservatism of white evangelical Protestants is explained by their tendency to support strict child-rearing practices and to expect children to be obedient.

Class Culture

The study of class influences on the beliefs of white Protestants goes back to the anti-immigrant and anti-minority perspectives of the 19th century. These interpretations emphasized the limited cultural horizons of lower-income and less-educated whites, who were thought to be the social base for reactionary movements in the United States. As economically vulnerable members of the dominant racial status group, lower-income and less-educated whites were depicted as hostile to change that might lead to greater opportunities for minorities as they were suspicious of cosmopolitan culture as a form of moral decay (see, e.g., Gusfield 1963; Hoftstadter 1955, 1963; Lipset 1955).

More recent treatments follow three major lines of thought. Similar to the mid-century liberal intellectuals, some social scientists have argued that the relative powerlessness of lower-status whites leads to greater insecurity and hence heightened distrust of unfamiliar others (see, e.g., Norris and Inglehart 2006; Verba, Schlozman, and Brady 1996). Some scholars argue that the restricted range and density of social networks among lower-status whites leads to greater insularity and hence less knowledge about and acceptance of the views of others (see, e.g., Campbell, Marsden, and Hurlbert 1986; Collins 1988:208–23). Scholars have also focused on the symbolic domain, arguing that conservative commentators effectively deploy symbols of cultural difference, such as images of latte-sipping liberals and minorities on welfare, as a means to stimulate resentments among lower-status whites (see, e.g., Leege et al. 2002:Chaps. 2–4).

In addition, organizational subordination of the interests of low-income and less-educated working-class whites may be a source of conservatism. The daily experience of order-taking, like that of economic deprivation, may create conditions of insecurity that foster distrust of unfamiliar others. Conversely, the daily experience of authority on the job is identified as creating psychological empowerment and greater trust of others (see, for example, Collins 1975:61–79, 1988:208–23; Kohn [1969] 1977).

The class circumstances of contemporary white evangelical Protestants are far more varied than they once were. Many middle and upper-middle class Americans have joined evangelical churches because they provide a more prescriptive source of moral guidance than mainline churches and greater support for economic individualism (Bartels 2008:Chap. 3; Hout and Greeley 2009). Many of those born into evangelical denominations have also experienced social mobility (Ammerman 2006). Socioeconomically, evangelical Protestants now look much more like the rest of the white population. At the same time, white evangelical Protestants remain less educated than other whites. Almost half of evangelicals in the National Election Surveys of 2000 and 2004 had a high school degree or less, and only 20 percent had baccalaureate or higher-level degrees, proportions that were nearly the reverse among mainline Protestants. Their vocabularies also lag behind other Americans. They are more likely to be blue-collar workers than other whites, and their family incomes fall below those of other whites by an average of $8,000 to $21,000 a year (Greeley and Hout 2006:98–100). Class culture may encourage political conservatism among white evangelical Protestants because of economic and organizational insecurities or culturally insular lifestyles.

Our approach to class culture highlights variables associated with conservatism, rather than the causal mechanisms producing relationships between class and conservatism. Three hypotheses capture the proposed relationship between class culture and conservatism:

  • H5: The political conservatism of white evangelical Protestants is explained by their lower incomes.

  • H6: The political conservatism of white evangelical Protestants is explained by their lower levels of organizational authority.

  • H7: The political conservatism of white evangelical Protestants is explained by their lower levels of education.

Cultural Geography

Since the turn of the 20th century (Simmel [1903] 1950), sociologists have theorized that rural areas and small towns are centers of traditionalism because of limited population diversity and opportunities for the expression of unconventional outlooks (see Fischer 1976:Chap. 6). By contrast, people who are the least likely to feel an affinity with conservative politics (e.g., the poor; people living in ethnic minority enclaves; young educated single people; and older professional people) are attracted to the cosmopolitan cultural environment of the city (Gans 1962). Compared to other settlement types, cities have much higher proportions of unmarried mothers; single professional women who are more likely to delay marriage and child-bearing; and cohabiting couples, including homosexual couples (Lesthaege and Neidert 2006). Such groups are less likely to feel an affinity with conservative politics and are more likely to take liberal positions on social issues.

Regional differences are rooted in histories of sectional opposition and in distinctive demographics and economies. Religion is an anchor for conservative outlooks in the Bible Belt. Receding racial politics opened the door to organizing southern conservative political energies along other lines. These energies were channeled by new leaders into protest against secularism and changing gender relations. “It is no surprise that almost as soon as the divisive issue of civil rights formally receded, the Religious Right emerged as a national movement with conspicuous Southern leadership” (Marsden 2005:237).

The geographical distribution of evangelicals has become more broadly based over time. Geographic mobility and the search for cultural stability in new suburban communities led to the formation of evangelical churches in every region of the country and in culturally homogeneous suburban and exurban communities (Mahler 2005). Recent studies show that more liberal and cosmopolitan groups continue to be clustered in cities (Lesthaege and Niedert 2006) and a majority of whites affiliated with evangelical denominations continue to reside in the South, while mainline Protestants, Catholics, and Jews are concentrated in the North and East (Greeley and Hout 2006:92–93). The conservative attitudes of white evangelical Protestants may be a function of their concentration in culturally homogeneous small town and rural settlements and in the more religious and predominantly Protestant South.

Two hypotheses are suggested:

  • H8: The political conservatism of white evangelical Protestants is explained by their concentration in exurban and rural areas and in small towns.

  • H9: The political conservatism of white evangelical Protestants is explained by their concentration in the South.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References

The data are from the American NES covering the years 2000–2004. Some measures were available in only one year, but where possible we pooled the data across the two presidential election survey years. These years represent a high point in the political influence of white evangelical Protestants. NES is the most useful data set for our purposes because adherence to traditional moral standards measures are included, as well as good measures of religious belief and practice, gender and family ideology, social class, settlement type, and region. NES also includes good indicators of partisan disagreement in the United States.2

Defining Evangelical Protestants

“Evangelical” has been defined in a number of ways. Following a long-standing convention (see, e.g., Warner 1988), we identify two major traditions in contemporary U.S. Protestantism. Although some significant overlap exists in the evangelical and mainline Protestant outlook, the mainline traditions are more likely to be critical of selfishness and understand religious duty as keeping the well-being of others in mind. They see the Bible as containing important truths, as well as ancient myths and legends, and rarely consider the Bible to be the literal word of God. By contrast, most evangelicals are less interested in helping the needy and more interested in saving souls. Their view of social reform focuses on the correction of individual moral failings. Moreover, evangelical denominations attribute religious authority to the Bible alone and accept it as the literal word of God. Evangelicals are, in the restricted sense, the descendants of fundamentalists (Marsden 2005:epilogue). However, where fundamentalists focused on in-group solidarity in opposition to the broader society, evangelical denominations encourage their congregants to participate fully in the larger society and to bear witness to their religious convictions. With these contrasts in mind, Kellstedt and Smidt (1991) defined four core beliefs of evangelicals: (1) the Bible is the literal word of God; (2) salvation is only possible through personal acceptance of Jesus as savior; (3) personal acceptance of Jesus as savior often occurs through the “born-again” experience, an intense event of spiritual renewal marking their life from that point on; and (4) the obligation to witness one's beliefs to others.

Evangelicals can also be defined by self-identification or by denominational affiliation.3 In this article, we will use denominational affiliation to identify evangelicals. Denominational definitions are appealing to social scientists because social scientists do not completely agree about the core beliefs of evangelicals (cf. Greeley and Hout 2006:Chap. 2; Wuthnow 2009), and most social surveys do not ask questions about all core beliefs of evangelicals. Moreover, a considerable overlap exists between holding the core beliefs of evangelicals and affiliating with an evangelical denomination or religious tradition. Members of evangelical denominations tend to be theologically conservative; they typically read the Bible literally; and emphasize the capacity of individuals to change their lives on the basis of accepting Jesus Christ as their personal savior.

Our classification relies on Steensland et al. (2000:Appendix A). This classification has become widely used in the literature (see, e.g., Greeley and Hout 2006; Green et al. 2007; Wilcox and Larson 2006). Steensland et al. (2000) divided Americans into six major categories based on theological criteria found in denominational creeds, as well as associational criteria such as membership in national religious organizations like the National Association of Evangelicals.4

We focus on white evangelicals. Many African-American Protestants hold the same core beliefs as white evangelicals (as do many Latino and Asian-American Protestants) and are almost as conservative as white evangelicals on some moral values issues, such as gay marriage (Loftus 2001). Differences are based on the divergent political paths taken by white and black evangelicals since the civil rights movement. The latter are firmly anchored in the Democratic Party and focus on equality and social justice. In this respect, African-American Protestants, whether evangelical or not, are closer to mainline Protestants in their attitudes about helping the poor and sharing abundance. Despite their social conservatism, African Americans have few ties to white evangelicals or other white religious conservatives. Moreover, African Americans tend to see the state as an ally because of its anti-discrimination laws and programs to aid the poor. The most conservative white evangelicals, by contrast, see government programs as a counterproductive and a wasteful substitute for individuals’ commitment to living a well-directed and self-disciplined life (Wald 2003:276–85).5

Dependent Variables

We examine attitudes about two salient moral values issues: abortion and homosexuality. The abortion measure (ABORTION) asked respondents to identify the conditions under which abortion should or should not be legal. The homosexuality variable (HOMOSEX) is a scale based on four items soliciting support/opposition to gay marriage, gay adoption, laws protecting homosexuals, and banning homosexuals from the military (see Table 1).

Table 1.  Descriptive statistics for dependent variables
 MeansStd. Dev.AlphaFactor Loading
ABORTION (1 = always permitted; 4 = never permitted)2.201.50  
 Oppose/support gay marriage   .65
 Oppose/support gay adoption   .69
 Support/oppose banning gays from military   .48
 Oppose/support laws protecting gays   .51
 Increase spending on military   .66
 Increase/decrease spending to fight terror   .56
 Using force is “best way” to ensure peace   .45
 Decrease/increase govt. $$:    
   Social services   .58
   Aid to poor   .53
   Welfare   .51
   Child care   .51
   Public schools   .49
   Min. income   .47
   Social Security   .41
 Policy goal: control illegal immig.   .70
 Inc/decrease govt. $$: border sec.   .63
 Immigration should decrease/increase   .54
REPUBLICAN ID (0 = non-Republican, 1 = Republican).44.50  
CONSERV (0 = extremely liberal, 6 = extremely conservative)4.381.57  

The immigrant, national defense, and social spending items assess polarizing and partisan-based attitudes along the conservative-liberal divide. The immigration scale (IMMIGRNT) is based on three items soliciting support/opposition to increased spending on border security and to increased immigration, and the importance of controlling illegal immigration as a foreign policy goal. The national defense scale (NATDEF) included three items soliciting support/opposition to increases in defense spending, increased spending on terrorism, and use of military force as “the best way” to ensure peace. The government social spending scale (SOCSPND) included seven items regarding support/opposition to spending for public schools, welfare, childcare, aid to the poor, Social Security, social services, and income maintenance programs to ensure a decent standard of living. We used principal factor analysis with Varimax rotation to create the scales for HOMOSEX, IMMIGRNT, NATDEF, and SOCSPEND.6

Finally, we investigate self-identified ideology as conservative (CONSERV) and political party identification (REPUB), the most important direct indicators of a conservative political identity. The self-identified ideology variable is based on a seven-point scale, ranging from very liberal to very conservative. REPUB is a dichotomous dependent variable, which separates Republican identifiers from Democratic identifiers and independents.7

Independent Variables

We dummy-coded variables for five ethnoreligious categories: (1) white evangelical Protestants, (2) white mainline Protestants, (3) black Protestants, (4) Catholics; and (5) secular and unaffiliated respondents. These are the largest religious and areligious groups in American society. We excluded other religious groups, including Mormons, Jews, Greek Orthodox, Hindus, and Muslims, due to their small numbers in the U.S. population.8 Eighty-two percent of African Americans in the NES sample identified as Protestants and were therefore classified as black Protestants. African Americans who did not identify as Protestants were combined with the religious groups with which they identified, or excluded from the analysis if they identified with one of the excluded groups.9

Our focal variables measure each of the five competing explanations for political conservatism among evangelicals and other white Christians. Religiosity is measured by six items (three related to belief and three related to practice) that loaded on a single factor: belief in the Bible as the literal word of God, views of the importance of religion in life, use of religion to guide action in everyday life, frequency of prayer, frequency of church attendance, and participation in church activities outside of services.

Moral standards traditionalism is measured by three items that loaded on the same factor: support/opposition to the view that moral views should adjust to a changing world, that people should be more tolerant of different moral standards, and that newer lifestyles are causing societal breakdown. We used principal factor analysis with Varimax rotation to create the religiosity and moral standards traditionalism scales.

Following our hypotheses, we disaggregated gender and family ideology into a measure of relations between spouses and a measure of relations between parents and children. We use one seven-point item to investigate male-dominant gender role attitudes. On this question, one end of the continuum is anchored by the statement that men and women should have equal roles; the other by the statement that a woman's place is in the home. We use a second item to investigate disciplinary strictness as an influence on child-rearing: the perceived importance of obedience in children.

The class culture variables are measured according to three bases of differentiation as hypothesized: income level (head of household's income), managerial authority (a categorical variable of jobs without managerial authority),10 and education (years of formal education completed). Cultural geography includes two dimensions: settlement type and region. We measure settlement type as separate categorical variables for rural/exurban and small town residence. We measure region as the dummy variable Southern residence.

We include two sociodemographic variables as controls. Age and sex are strongly associated political attitudes and party identification in the United States. In many attitude domains, older people and men tend to be more conservative than younger people and women. (Race, another important sociodemographic influence in American political life, is built into our ethnoreligious categories.) Finally, we include year as a control variable to capture the effects of time where data are pooled over two survey years (see Table 2).

Table 2.  Continuous and categorical independent variables
 MeanStd. Dev.AlphaFactor Loading
 Frequency of church attend.   .67
 Frequency of prayer   .58
 Degree to which rel. guides daily life   .55
 Importance of religion in life   .55
 Bible is literal word of God   .51
 People should not adjust their moral views to a changing world   .52
 Newer lifestyles are causing societal breakdown   .49
 Society should not be more tolerant of different moral stand.   .47
GENROLE (1 = Men and women should have equal roles; 7 = Women's place is in the home)1.501.20  
OBEY (0 = Self-reliance most important for children; 1 = Obedience most important for children).40.49  
EDUCATION (1 = 8th grade or less; 7 = advanced degree)4.281.63  
AGE 4617.20  
 White evangelical.20   
 Mainline Protestant.14   
 African-American Protestant.12   
 Small town.16   
SEX (Male).45   
YEAR (2004).48   
RACE (Black).13   

We found no reason for concern about multicollinearity and no reason to believe that the unique effects of the focal variables would be difficult to identify. The focal variables are not highly correlated in the NES 2000–2004 data. The strongest correlation (r = .26) was between religiosity and moral standards traditionalism (see Table 3).

Table 3.  Correlation matrix
GENROLE  1.00.07−.01−.12−.09−.01.00−.07.12
OBEY   1.00−.10−.11−.11.16.06−.01.17
MGR     1.00.14−.06−.03.16−.06
SMTOWN        1.00.04
RURAL         1.00.03
EVANG          1.00

Methods of Analysis

We use the term “explain” in two senses: to denote the extent to which each of the five competing explanations account for the conservative politics of white evangelical Protestants and to denote the extent to which our indicators predict the political views of the adult U.S. population, net the effect of religious affiliation. These two usages imply two distinct methods of analysis. One method looks for the attenuation of the relationship between membership in an evangelical denomination and conservative political attitudes as each of the indicators of the five explanations are introduced. To examine this issue, we first measure the “evangelical effect” in bivariate regressions. We then introduce indicators of the five competing explanations one at a time to determine which one(s) lead to the greatest attenuation of the evangelical effect. The other method tests the effect of the focal variables on the political attitudes of the U.S. population at large. We analyze the ordinal dependent variables (ABORTION and CONSERV) using ordered logistic regression, and we analyze the categorical dependent variable (REPUB) using binomial logistic regression. We analyze the four continuous dependent variables (HOMOSEX, IMMIGRNT, NATDEF, and SOCSPND) using ordinary least squares regression.

In this second analysis, we present results for two models for each dependent variable. Model 1 includes ethnoreligious categories and control variables only. Model 2 adds independent variables from each of the competing explanatory frameworks. If religiosity, moral standards traditionalism, family ideology, or the other explanatory frameworks in our study are equally relevant to nonevangelical Protestant populations, a case can be made that the “orthodox” (Hunter 1991) or “traditionalists” (Green 2009a) of all ethnoreligious types are the center of conservatism in the United States, rather than white evangelicals alone. Conversely, if associations between religious traditions and political attitudes hold up following the introduction of these variables, we have reason to conclude that religious traditions are themselves a distinct base of conservatism in the U.S. adult population.11


  1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References

Attenuation of the Evangelical Effect

Although sometimes characterized as hawkish on national defense issues and anti-immigrant, white evangelical Protestants showed no statistically significant tendency to adopt conservative attitudes on national defense or immigration, or to self-identify as conservatives. Results of the attenuation analysis are therefore reported for only four of the seven dependent variables (see Table 4). A highly significant evangelical effect was found for the models testing attenuation on abortion, homosexuality, social spending, and Republican Party identification.

Table 4.  Attenuation of the “evangelical effect”
βAdj. R2βAdj. R2βAdj. R2Odds RatioPseudo R2
  1. Note: ***p < .001, **p < .01, *p < .05, +p < .10.

Independent Variables
 White evangelicals.29***.08.17***.03.09*.012.09***.01
 White evangelicals.17***.23.12***.***.01
 Moral traditionalism.41*** .18*** .26*** .94 
 White evangelicals.23***.16.11***.16.09*.012.05***.02
 Religiosity.29*** .37*** −.03 1.06 
 White evangelicals.21***.14.13***.***.01
 Male-dominant gender roles.25*** .11*** .13** 1.02 
 White evangelicals.23***.14.18***.03.08+.012.41***.02
 Children: obey.25*** −.05 .05 .65** 
 White evangelicals.27***.10.15***.12.09*.012.22***.02
 Low education.15*** .30*** .07+ 1.16** 
 White evangelicals.24***.21.20***.12.09*.013.60***.19
 Manager−.37*** .31*** .06 11.60*** 
 White evangelicals.27***.12.17***.03.09*.042.40***.04
 High income−.19*** −.01 .18*** 1.09*** 
 White evangelicals.26***.09.19***.04.09*.012.56***.03
 South.12*** −.13*** −5.00 .47* 
 White evangelicals.26***.09.23***.17.08*.012.57***.04
 Rural−.11*** .38*** −.09* 2.34*** 
 White evangelicals.28***.08.19***.06.08*.012.22***.02
 Small town.05= −.18*** .04 .68** 
 N892 823 630 885 

The evangelical effect disappeared altogether on three of the four variables when the moral standards traditionalism scale was the only other variable introduced into the regression (Republican Party identification is the exception). The introduction of religiosity, male-dominant gender role attitudes, low education level, and managerial job attenuated the evangelical effect in the abortion and homosexuality models. These results suggest that moral standards traditionalism is the strongest and most consistent single source of conservatism among white evangelicals. It is a more important source of conservatism than religiosity or male-dominant gender role attitudes, even on an issue like abortion, which is often associated with religiosity and gender role attitudes. Class variables cannot be dismissed as a source of the conservative evangelical attitudes, but they are not a consistent source of conservatism when compared to moral standards traditionalism, religiosity, or male-dominant gender role attitudes.

None of the variables, including moral standards traditionalism, led to significant attenuation of the evangelical effect on Republican Party identification. These results indicate that membership in an evangelical denomination is associated with Republican Party identification, quite apart from the value commitments and social circumstances of congregants. Djupe (2000) argued that churchgoers are bound to political identities through (1) homogenous social ties that foster and reinforce political loyalties and (2) psychological identity standards that tend to encourage a habitual and commonsensical relationship between religious and political loyalties. For many white evangelicals, the Republican Party was by the early 2000s more or less automatically associated with “right living” in the world (see also Kaplan, Freedman, and Iacaboni 2007).

Bases of Political Conservatism in the U.S. Adult Population

The second set of analyses shifts the focus to a comparison of the five competing explanations using the full NES sample populations. The results of these analyses are reported in Table 5. The baseline models (Model 1) show that, controlling for age and gender, white evangelical Protestants were the most conservative ethnoreligious group on five of the seven dependent variables (abortion, homosexuality, national defense, self-identified ideology, and Republican Party identification). Catholics had marginally higher levels of conservatism on immigration, and mainline Protestants were more conservative on government social spending. The baseline models explained relatively little of the variance in our dependent variables, 15 percent or less in each case.

Table 5.  LS and logistic regressions testing competing explanations for conservative views of U.S. adults, 2000–2004 (standardized coefficients)
Model 1Model 2Model 1Model 2Model 1Model 2Model 1Model 2
 Age (+= older).06*−.05+.17***.03.16**.10+−.02.09**
 Sex (+= male).01.07*−.03.07*−.10+−.03.15***.10**
 Year (+= 2004).05.05
Ethnoreligious Groups
 Evangelical Protestants.24***.05.67***.32***.32**.08.19*.07
 Mainline Protestants−.06−.06.28***.14+.21+.10.24**.19*
 Black Protestants.04−.09*.42***.18**.14−.01−.02−.05
Moral Traditionalism.17***.24***.23***.26***
Gender & Family Ideology
 Male-dominant gender roles.15***.18***−.04.08
 Obedient children.08**−.07−.01.01
Class Culture
 Income (+= High)−.03.02−.01.04
 Education (+= Low).13***.23***.28***.01
Cultural Geography
 Small town.04.04.02−.04
Adj. R2.
Model 1Model 2Model 1Model 2Model 1Model 2
  1. Note: ***p < .001, **p < .01, *p < .05, +p < .10.

  2. aOdds ratios.

 Age (+= older).20***.15***.11***.07*1.00.99
 Sex (+= male).16***.18***.13***.11***1.73***1.52*
 Year (+= 2004)−.17***−.15***.961.01
Ethnoreligious Groups
 Evangelical Protestants.26*.03.10+.063.67***3.25**
 Mainline Protestants.07−.***3.38**
 Black Protestants.03−.***.02***
Moral Traditionalism.20***.11**2.05***
Gender & Family Ideology
 Male-dominant gender roles−.07.051.25***
 Obedient children.09+.06+1.76**
Class Culture
 Income (+= High).06.01.99
 Education (+= Low).25***.021.34***
Cultural Geography
 Small town.00−.041.02
Adj. R2.
Log-likelihood    −574.85−511.2
Nagelkerke pseudo-R2    .12.22

Variables measuring moral standards traditionalism, religiosity, and male-dominant gender role attitudes were more important than religious traditions in the explanation of conservative attitudes. Net of other significant covariates, the effect of being a white evangelical Protestant was not significant in five of the regression models: abortion, immigration, national defense, social spending, or self-identified conservative. Significant net effects were found only for homosexuality and Republican Party identification. The Catholic net effect was in a conservative direction only in the homosexuality model, while mainline Protestant (positive) and black Protestant (negative) net effects were found only for Republican Party identification.12

Moral standards traditionalism proved to be the most consistent predictor of conservative attitudes (Model 2). Net of other significant covariates, moral standards traditionalism showed relatively strong effects on each of the dependent variables. Both male-dominant gender role attitudes and religiosity were significantly associated with attitudes about abortion and homosexuality and male-dominant gender role attitudes significantly affected Republican Party identification. Low education was statistically significant on four of the seven of the dependent variables; less educated people were more conservative than highly educated people on abortion, homosexuality, immigration, and national defense. At the same time, less educated people were significantly less likely than highly educated people to identify as Republicans. Apart from educational level, neither class nor cultural geography variables showed consistent effects in Model 2, net of other significant covariates.

For each of the dependent variables, Model 2 demonstrated far more explanatory power than the baseline model. Model 2 was not, however, consistently superior to Model 1 in efficiency, as measured by BIC. In particular, BIC showed greater efficiency for Model 1 in the case of ABORTION.

Taken together, these findings provide support for the notion that the conservative attitudes of white Christians in the United States are not due to membership in particular religious traditions, but rather to social circumstances and value commitments that are found more often among them. Low education levels, moral traditionalism, religiosity, and male-dominant gender role attitudes are the proximate causes of conservative political attitudes. White evangelical Protestants tend more often to hold conservative views because they are more likely to share the social condition (low levels of education) and outlooks (moral standards traditionalism, religiosity, and male-dominant gender role attitudes) associated with these views than do white members of other Christian religious traditions.13

One plausible interpretation of the literature is that variables like religiosity, moral standards traditionalism, and gender and family ideology, which are associated with conservatism throughout the population, will have particularly strong effects within the subpopulation of white evangelical Protestants. In a separate analysis, we examined interactions between white evangelicals and other independent variables. Results of analyses incorporating interaction terms did not change the overall pattern of findings reported in Model 2.

They did, however, add nuance to these findings. They showed that more educated and less educated evangelicals were almost as likely to identify as Republicans. This similarity in partisanship stood in marked contrast to the rest of the population, where more educated people were much more likely to identify as Republicans. The results also showed that less educated evangelicals were quite a bit more conservative than less educated nonevangelicals on issues related to immigration. Some other popular conceptions of evangelicals were also supported in these analyses. Evangelicals who valued obedience in children were more conservative on homosexuality and immigration than nonevangelicals who valued obedience in children. Southern evangelicals were more conservative on defense issues than Southern nonevangelicals. However, other popular conceptions of evangelicals failed to find support in these analyses. Highly religious evangelicals were less conservative than highly religious nonevangelicals on national defense issues, and rural evangelicals were less conservative than rural nonevangelicals on the issues of abortion and homosexuality.

Based on the Bayesian information criterion (BIC), the models with interaction terms yielded improvements in efficiency for all dependent variables other than Republican Party identification, but accounted for little change in the proportion of explained variance. Moreover, these results were not supportive of the idea that the primary causes of conservatism have consistently stronger effects among evangelicals than they do among members of other Christian religious traditions. Instead, interaction terms showed no consistent pattern of effects across dependent variables.14


  1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References

Scoring the Five Explanations

Adherence to traditional moral standards accounts for the link between white evangelicals and conservative politics across a wide range of issues. Moral standards traditionalism emerged as an important explanation for conservative politics in the U.S. adult population generally, controlling for other covariates. Many evangelicals as well as members of other Christian religious traditions see the world as governed by moral standards and these standards lead them to oppose changes in society that threaten these standards. Commitments to traditional moral standards and fears about threats to these standards shape both their attitudes about government social provision and their attitudes about acceptable forms of social relations within families and society (see also Hetherington and Weiler 2009).

One reason for the strength of these relationships may be that traditional moral standards are supported by the value structure of many institutions in American society, including not only religious injunctions to obey God's commandments, but also business ideologies of leadership, military ideologies of command, and athletic ideologies of fortitude in competition. Republicans during this period recognized the importance of appeals to traditional moral standards by focusing on such qualities as “moral strength,”“moral clarity,” and “traditional values” as characteristics of their candidates and their party (and as failings of the opposition's candidates and party).

We found more limited support for two of the remaining four explanations. On the issues of abortion and homosexuality, religiosity and male-dominant gender role attitudes were closely linked to conservative views, but they were not as important in the other areas we investigated. We conclude that these sources of conservatism play a specialized role in contemporary American politics; they are important for binding together the conservative moral values coalition, led by evangelicals and supported by conservative members of other Christian religious traditions (Green 2009a), but they have little influence beyond this arena. This finding should not be interpreted as downplaying the importance of religiosity and gender and family ideology in American politics. Although their influence is more narrowly focused, the arena of moral values politics in which they play an important role has been an important source of mobilized energy for the Republican Party, and the moral values coalition represents a sizable part of the party's electoral base (see also Brint and Abrutyn 2009).

Nevertheless, once we go beyond the issues appealing to the moral values coalition, religious people and people who uphold male-dominant gender roles were not predictably conservative in their political views (see also Greeley and Hout 2006; Smith 1998; Woodberry and Smith 1998). This is particularly true of religious people. They were not, in the main, opposed to government social programs or to internationalism and diplomacy in foreign policy. Nor were they notably anti-immigrant or opposed to environment-friendly policies (on environmentalism, see, e.g., Breslau and Brant 2006; and Totten 2006). In this respect, commentators have frequently overstated the support of religious people for the broader conservative movement. They have failed to see how narrowly focused this support has been on issues like abortion, gay rights, and end-of-life care.

We find less evidence in favor of the final two explanations. In the context of salient divisions in American politics during the period of this study, class failed as a general explanation and, in particular, neither low income nor lack of managerial authority was consistently or strongly associated with conservative views.

However, educational level showed significant net associations on several issues (see also Davis 1982; Kingston et al. 2003). In so far as these patterns persist, it will be important to reformulate the class argument to focus on the divisions between highly educated and less educated people.15 Among whites, the conflict between less educated conservatives and more educated liberals has certainly been an important one in American politics over the last 40 years (see, e.g., Hodgson 1976:Chap. 14; Leege et al. 2002:Chaps. 2–4; Perlstein 2008; Williams 2009). As we have noted, several explanations for the conservatism of less educated people are now in contention. These explanations focus on the influence of powerlessness; restricted social networks; lack of information; and symbols of cultural difference, as deployed by conservative commentators. Each one of these explanations is interested in discerning the sources of inclusive as opposed to distrustful attitudes. The sources of liberal attitudes among the highly educated also merit further investigation.16

The fifth explanation, cultural geography, finds very little support in these analyses. At most, rural people show signs of distrust in the face of unfamiliar others, as indicated by their conservative views on homosexuality. But rural residence operates independently of religion. Indeed, rural evangelicals were significantly less conservative on abortion and homosexuality than rural nonevangelicals. The causes of this finding require further investigation, but the finding itself is, in our view, sufficient to lay the cultural geography argument to rest, as it applies to evangelicals.

The Search for Parsimony and the Search for Votes

We began by proposing to identify the most important source of conservatism's appeal to white evangelicals and members of other religious groups. Although our quest to discover the key to unlock the relationship between religion and politics has been at least moderately successful, reflection on our findings, as well as consideration of the developing literature on political parties, has led us to question the usefulness of looking for a single master key to explain this relationship. We can best show the limitations of such a quest by comparing the interests and methods of social scientists and those of political party organizations.

Social scientists often look for parsimonious explanations because social scientific reputations can be built upon elegant explanations that unlock or provide a new perspective. The search for qualities and processes that explain large amounts of variance in social outcomes, or provide new ways of looking at old problems, are consequently the lode stars of many social science careers. Those who succeed in discovering the underlying cause of a social pattern increase the likelihood that they will attract attention and followers (see, e.g., Lamont 2009:Chap. 3; for the case of philosophy, see Collins 1998).

Some social scientists argue that politics work similarly in so far as they are based on creating singular, powerful, and coherent views of the world. The sociolinguist George Lakoff writes, for example: “What conservative and liberal political leaders and ideologues do is to try to get voters to become coherent in their view to move to one pole or the other, that is, to be entirely liberal or entirely conservative over the full range of issues” (Lakoff [1996] 2002:15–16). However, a developing literature in political science and sociology (see, e.g., Brint and Abrutyn 2009; Leege et al. 2002; Monson and Oliphant 2007) shows that modern party politics work in a fundamentally different way than Lakoff indicates, and that this difference undermines the possibility of finding master keys to unlock the source of relationships such as that between religion and conservatism.

Political parties are responsible for aggregating interests in society to gain electoral power. In a highly differentiated electorate, leaders of political parties understand that no singular master key exists to bring voters into their electoral coalitions; instead, they operate to assemble electoral coalitions by explicitly appealing both to broad and narrow bases of affiliation. They must also keep in mind the mobilization potential of the different constituencies to which they appeal. Constituencies that contain many activists are valuable because activists can stir enthusiasm for candidates and help to deliver votes.17

Parties rely on modern campaign technology (including polling, focus groups, test marketing of advertisements, and sophisticated market segmentation) to detect what symbols and symbolic associations can be used to create support for the party or anxiety about the opposition (Leege et al. 2002:Chaps. 2–4). Our data suggest that messages about moral strength and moral perseverance attract conservatives, both those who are religious and those who are not. We can see the attractive power of these messages by examining individual NES items. Three out of five NES respondents believe that newer lifestyles are causing societal breakdown, and only two out of five say that people should adjust their moral views to a changing world. Messages about the importance of family ties also attract very widespread support; more than four out of five NES respondents said they favor more emphasis on traditional family ties. Many institutions in American society, not only churches, support these traditionalist views.

Views supporting the centrality of religion in public life or the validity of male-dominant gender ideology have narrower appeal because the number of people who are particularly responsive to these messages is smaller, and few institutions in American society, outside of theologically conservative churches, reinforce these messages. Again, we can see the limits of these messages by examining responses on relevant NES items. Thus, nearly three out of five NES respondents rejected the idea that the Bible is the literal word of God, and fewer than two out of five said that religion provided a great deal of guidance in their day-to-day life. Similarly, three out of four NES respondents said that men and women should have equal roles in society, and three out of five rejected the idea that obedience was one of the most important qualities in children. Conservative messages about the centrality of religion and the different roles of men and women are nevertheless important for drawing religious people and people favoring male-dominant gender roles into the Republican Party's moral values coalition. In this respect, they play a specialized role in the cultural system of electoral politics, but one that is important for rallying activists.

Broadly resonant messages and specialized messages attuned to activists are two instruments of political parties. More targeted messages are also used by parties to reach out to smaller segments of the electorate, such as people with children serving in the armed forces or swing voters worried about changes in their marginal tax rates. These micro-targeting efforts began in earnest a generation ago (Blumenthal 1980:Chap. 12) and have become ever more sophisticated since that time. A study of political communications of the Republican National Committee in the 2004 presidential election showed that the market segmentation of political messages has advanced to the point that small slivers of the electorate are targeted with messages appealing to their particular interests (Monson and Oliphant 2007). Similar micro targeting informs communications strategies of the Democratic Party as well (Penn and Zalesne 2007:Chap. 6).

The partisan search for votes, in short, deploys a wide variety of symbolic resources, some with relatively broad appeal, some with specialized appeal to mobilized constituencies, and many with very narrow and targeted appeal, to assemble winning electoral coalitions, piece by piece, across numerous organizational and demographic contexts. For Republicans, messages resonating with adherence to traditional moral standards represent one important symbolic resource, but not the only one.

The social scientist's search for parsimony and the partisan search for votes thus work on different principles. This is an important reason why no overarching social science explanation of modern conservatism's appeal can be possible. In Berlin's (1953) terms, social scientists have incentives to be hedgehogs, but political parties have incentives to be foxes.

  • 1

    Moral values traditionalism closely parallels Hetherington and Weiler's (2009) conceptualization of “authoritarianism.” However, we have reservations about the connotations of the term “authoritarianism,” given its past usage to describe individuals and regimes that deny basic civil liberties and civil rights to dissenters.

  • 2

    Two other widely used data sets, the General Social Survey (GSS) and the Pew Political Landscape Surveys (PEW), have significant weaknesses for purposes of comparing rival explanations for the conservatism of white Protestants. GSS includes only a few items on national defense/national security and no items that measure moral standards traditionalism. PEW does not consistently measure frequency of church attendance and does not include measures of moral standards traditionalism.

  • 3

    Definitions based on core beliefs yield estimates of more than 30 percent of the adult population in the United States; self-identification definitions yield estimates of fewer than 20 percent. Denomination definitions yield estimates of around 25 percent of the adult population (see Kellstedt and Smidt 1991; Wald 2003:162–13).

  • 4

    In Steensland et al. (2000), both neo-evangelicals and fundamentalists are included within the evangelical Protestant denominational measure. They use church attendance to classify nondenominational and “no denomination” respondents. Those who attended frequently (once a month or more) were placed in the “evangelical” category, while others were placed into the mainline Protestant category. They justified this coding based on the burgeoning nondenominational evangelical movement. In addition, frequent church attenders in nondenominational and no-denominational churches express attitudes and beliefs that closely paralleled those in theologically conservative denominations.

  • 5

    Some tensions between white evangelicals and black Protestants date from the days of racial segregation in the South; white evangelicals played either a complicit or active role in maintaining the institutions of Jim Crow.

  • 6

    Scales are standardized to produce a mean of 0 and a standard deviation of 1. Where possible, polychoric and polyserial correlations are preferred for any correlations that involve noncontinuous indicators. Several of our scales (HOMOSEX, IMMIGRNT, NATDEF, SOCSPEND, religiosity, and moral standards traditionalism) include indicators that are not continuous. We attempted to use polychoric factor analysis, but it produced high collinearity inefficiencies, likely the result of the number of cases (under 2,000) and the number of indicants (greater than 20). See Yung and Bentler (1994).

  • 7

    Support for unregulated market capitalism (Somers and Block 2005; cf. Felson and Kindell 2007) and intolerance for dissenters (Woodberry and Smith 1998) have been identified by some scholars as important bases for the conservatism of white evangelicals. NES includes no items that allow us to examine the views of evangelicals on these two attitude dimensions.

  • 8

    Mormons and Jews each represent approximately 2 percent of the NES sample. Small numbers produce parameter estimates that are unlikely to be statistically reliable. Hindus, Muslims, Buddhists, and a variety of other small religious communities make up 6 percent of the sample. We excluded these groups because the diversity of one category could lead to results that are difficult to interpret substantively. We note that Mormons are more Republican and conservative than white evangelicals, according to recent survey data (Pew Forum on Religion and Public Life 2009).

  • 9

    The NES sample contains a small number of African-American Catholics (5 percent of African Americans, or about one half of 1 percent of the total sample) and seculars (7 percent of African Americans). We classified these respondents with other Catholics and other secular/unaffiliated people. Approximately 6 percent of African Americans were excluded from the analysis because they identified with groups not large enough to be included in the analysis, such as Jews and Muslims.

  • 10

    We also examined a six-category occupational scale—including managers and executives, professionals and technical workers, clerical workers, skilled blue-collar workers, unskilled blue-collar workers, and farm workers—using dummy codes for each category. However, this scale failed to yield consistent empirical results.

  • 11

    Models were fit using case-wise deletion of missing values. All analyses are based on constant Ns so that coefficients and odds ratios can be compared.

  • 12

    Notably, white evangelical Protestants were no more likely to identify with the Republican Party than were white members of mainline Protestant denominations, and they were less conservative than mainline Protestants on government social spending. We find, as have others (see, e.g., Greeley and Hout 2006; Manza and Brooks 1999), that Protestants, in general, are conservative on government social spending and other economic issues.

  • 13

    We developed structural equation models to examine relationships among causal variables and two dependent variables, abortion and homosexuality. We posited a direct effect of membership in an evangelical denomination on conservative abortion and homosexuality attitudes and an indirect effect in so far as membership influenced higher levels of moral standards traditionalism, religiosity, and male-dominant gender role attitudes. The direct effects of membership were weak but significant for both abortion and homosexuality. Evangelical Protestantism was, however, a strong predictor of religiosity, moral standards traditionalism, male-dominant gender role attitudes, and being Republican. Evangelical membership was associated with less education, was a good predictor of conservative views on homosexuality, and, though less strongly, of attitudes on abortion. Results of the structural equation models are available from the authors upon request.

  • 14

    Results of the analyses with interaction terms are available from the authors upon request.

  • 15

    The other class variables (income and managerial authority) showed weak effects once class-linked demographic variables, such as age and gender, were controlled. This does not mean that class variables are irrelevant to the explanation of political attitudes. Older men, for example, are more likely to be in positions of authority and perhaps also to identify with those who are in positions of authority. Consequently, managerial authority may be an important structural source of conservatism, as Kohn ([1969] 1977), Collins (1975, 1988) and others have theorized, but its political importance may be obscured in statistical analyses that also include age and gender.

  • 16

    The effects of higher education may be due to feelings of efficacy and confidence related to higher education; the extensiveness and diversity of the personal networks of educated people; or the value climate of social tolerance and equality characteristic of contemporary U.S. colleges and universities, which tend to attract young people of more liberal views (see Kingston et al. 2003). Commentators on the Left have frequently deployed symbols emphasizing the cultural “otherness” of less educated and more religious people, and this, too, may help to explain the association of higher education with liberalism.

  • 17

    Political parties have been active over the last several decades in building networks of support in national and local church communities (see, e.g., Brint and Abrutyn 2009; Tipton 2007; Wald, Owen, and Hill 1988) in large part because they recognize that churches bring together people who associate and interact regularly with one another and therefore represent valuable sites for the reinforcement of partisan loyalties and the creation of support for party issues (Leege et al. 2002:Part 1; Wald 2003:Chaps 1 and 2). Within church communities, partisan activists associated with the dominant party will find receptive audiences among congregants, while those who identify with the opposition party will find fewer sympathizers. These in-group processes undoubtedly lead some who have divergent political and social views to leave their congregations in search of more congenial churches (see Pew Forum on Religion and Public Life 2008; Sherkat 2001), leading to greater partisan homogeneity within church communities.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Competing Explanations
  5. Methods
  6. Results
  7. Discussion
  8. References
  • Adorno, Theodor. 1950. The authoritarian personality. New York : Harper.
  • Alwin, Duane F. 1986. Religion and parental child-rearing orientations: Evidence of a Catholic-Protestant convergence. American Journal of Sociology 92(2):41240.
  • Ammerman, Nancy T. 2006. Deep and wide: The real American evangelicals. American Interest 2 (October). Retrieved December 9, 2008, from
  • Baker, Wayne E. 2005. America's crisis of values: Reality and perception. Princeton , NJ : Princeton University Press.
  • Bartels, Larry M. 2008. Unequal democracy: The political economy of the new gilded age. New York : Russell Sage Foundation Press.
  • Berlin, Isaiah. 1953. The hedgehog and the fox. New York : Simon & Schuster.
  • Blumenthal, Sidney. 1980. The permanent campaign. New York : Simon and Schuster.
  • Breslau, Karen and Martha Brant. 2006. God's green soldiers. Retrieved January 8, 2007, from
  • Brint, Steven and Seth Abrutyn. 2009. Moral-values politics: The emergence of an electoral system. In Evangelicals and democracy in America, vol. 2: Religion and politics, edited by StevenBrint and Jean ReithSchroedel, pp. 10540. New York : Russell Sage Foundation Press.
  • Brooks, Clem and Jeff Manza. 1997. The social and ideological bases of middle-class political realignment in the United States, 1972 to 1992. American Sociological Review 62(2):191208.
  • Campbell, Karen E., Peter V. Marsden, and Jeanne S. Hurlbert. 1986. Social resources and socioeconomic status. Social Networks 8(1):97117.
  • Collins, Randall. 1975. Conflict sociology. New York : Academic Press.
  • Collins, Randall. 1988. Theoretical sociology. New York : Harper and Row.
  • Collins, Randall. 1998. The sociology of philosophies: A global theory of intellectual change. Cambridge , MA : Harvard Belknap.
  • Davis, James A. 1982. Achievement variables and class cultures: Family, schooling, job and forty-nine dependent variables in the cumulative GSS. American Sociological Review 47(5):56986.
  • Davis, Nancy J. and Robert V. Robinson. 1996. Are the rumors of war exaggerated? Religious orthodoxy and moral progressivism in America. American Journal of Sociology 102(3):75687.
  • Davis, Nancy J. and Robert V. Robinson. 2001. Theological modernism, cultural libertarianism, and laissez-faire economics in contemporary European societies. Sociology of Religion 62(1):2350.
  • Djupe, Paul. 2000. Religious brand loyalty and political loyalties. Journal for the Scientific Study of Religion 39(1):7889.
  • Ellison, Christopher G. and Darren E. Sherkat. 1993. Obedience and autonomy: Religion and parental values reconsidered. Journal for the Scientific Study of Religion 32(4):31329.
  • Felson, Jacob and Heather Kindell. 2007. The elusive link between conservative Protestantism and conservative economics. Social Science Research 36(2):67387.
  • Fischer, Claude. 1976. The urban experience. New York : Harcourt Brace.
  • Gans, Herbert. 1962. Urbanism and suburbanism as ways of life: A re-evaluation of definitions. In Human behavior and social processes, edited by Arnold M.Rose, pp. 62548. Boston : Houghton-Mifflin.
  • Greeley, Andrew and Michael Hout. 2006. The truth about conservative Christians: What they think and what they believe. Chicago , IL : University of Chicago Press.
  • Green, John C. 2004. The American religious landscape and political attitudes: A baseline for 2004. Pew Forum on Religion and Public Life. Available at
  • Green, John C. 2009a. Evangelicals and the “Traditionalist Alliance.” In Evangelicals and American democracy, volume 1: Religion and society, edited by StevenBrint and Jean ReithSchroedel, pp. 11758. New York : Russell Sage Foundation Press.
  • Green, John C. 2009b. Personal communication based on analysis of the fifth national survey of religion and politics, post-election sample.
  • Green, John C., Lyman A. Kellstedt, Corwin E. Smidt, and James L. Guth. 2007. How the faithful voted: Religious communities and the presidential vote. In A matter of faith: Religion in the 2004 presidential election, edited by David E.Campbell, pp. 1536. Washington , DC : Brookings Institution Press.
  • Gusfield, Joseph. 1963. Symbolic crusade: Status politics and the American temperance movement. Urbana , IL : University of Illinois Press.
  • Guth, James L. and John C. Green. 1993. Religious salience: The core measure? In Rediscovering the impact of religion on political behavior, edited by David C.Leege and Lyman A.Kellstedt, pp. 15774. Amonk , NY : M.E. Sharpe.
  • Herberg, Will. [1955] 1983. Protestant, Christian, Jew. Chicago , IL : University of Chicago Press.
  • Hetherington, Marc J. and Jonathan D. Weiler. 2009. Authoritarianism and polarization in American politics. Cambridge , UK : Cambridge University Press.
  • Hillygus, D. Sunshine and Todd G. Shields. 2005. Moral issues and voter decision making in the 2004 presidential election. Political Science and Politics 38(2):20109.
  • Hodgson, Godfrey. 1976. America in our time. New York : Doubleday.
  • Hoftstadter, Richard. 1955. The age of reform: From Bryan to FDR. New York : Knopf.
  • Hoftstadter, Richard. 1963. The pseudo-conservative revolt. In The radical right, edited by DanielBell, pp. 7596. New York : Doubleday.
  • Hout, Michael and Andrew Greeley. 2009. Conservative Protestants and the increasing Republican Party identification of American voters, 1972–2006. In Evangelicals and American democracy, volume 2: Religion and politics, edited by StevenBrint and Jean ReithSchroedel, pp. 5782. New York : Russell Sage Foundation Press.
  • Hunter, James Davison. 1991. Culture wars: The struggle to define America. New York : Basic Books.
  • Kaplan, Jonas T., Joshua Freedman, and Marco Iacaboni. 2007. Us versus them: Political attitudes and party affiliation influence neural response to faces of presidential candidates. Neuropsychologia 45(1):5564.
  • Kellstedt, Lyman A. and John C. Green. 1993. Knowing God's many people: Denominational preference and political behavior. In Rediscovering the religious factor in American politics, edited by David C.Leege and Lyman A.Kellstedt, pp. 5371. New York : M.E. Sharpe.
  • Kellstedt, Lyman A. and Corwin Smidt. 1991. Measuring fundamentalism: An analysis of different operational strategies. Journal for the Scientific Study of Religion 30(3):25978.
  • Kingston, Paul W., Ryan Hubbard, Brent Lapp, Paul Schroeder, and John Wilson. 2003. Why education matters. Sociology of Education 76(10):5370.
  • Kohn, Melvin. [1969] 1977. Class and conformity. Chicago , IL : University of Chicago Press.
  • Lakoff, George. [1996] 2002. Moral politics: How liberals and conservatives think. Chicago , IL : University of Chicago Press.
  • Lamont, Michele. 2009. How professors think: Inside the curious world of academic judgment. Cambridge , MA : Harvard University Press.
  • Langer, Gary and Joel Cohen. 2005. Voters and values in the 2004 election. Public Opinion Quarterly 69(5):74459.
  • Layman, Geoffrey. 2001. The great divide: Religious and cultural conflict in American party politics. New York : Columbia University Press.
  • Leege, David C., Kenneth D. Wald, Paul D. Mueller, and Brian S. Krueger. 2002. The politics of cultural differences: Social change and voter mobilization strategies in the post-new deal period. Princeton , NJ : Princeton University Press.
  • Lesthaege, Ron J. and Lisa Neidert. 2006. The “second demographic transition” in the U.S.: Exception or textbook example Population and Development Review 32(4):66998.
  • Lipset, Seymour M. 1955. Three decades of the radical right: Coughlinites, McCarthyites, and Birchers. In The radical right, edited by DanielBell, pp. 373446. New York : Doubleday.
  • Loftus, Jeri. 2001. America's liberalization in attitudes toward homosexuality, 1973 to 1998. American Sociological Review 66(5):76282.
  • Mahler, Jonathan. 2005. The soul of the new exurb. New York Times Magazine (March 27). Available at
  • Manza, Jeff and Clem Brooks. 1997. The religious factor in U.S. presidential elections, 1960–1992. American Journal of Sociology 103(1):3881.
  • Manza, Jeff and Clem Brooks. 1999. Social cleavages and political change. Oxford , NY : Oxford University Press.
  • Marsden, George. 2005. Fundamentalism and American culture, 2nd ed. New York : Oxford University Press.
  • Monson, J. Quin and J. Baxter Oliphant. 2007. Microtargeting and the instrumental mobilization of religious conservatives. In A Matter of faith: Religion in the 2004 presidential election, edited by David E.Campbell, pp.95119. Washington , DC : Brookings Institution Press.
  • Morone, James A. 2003. Hellfire nation: The politics of sin in American history. New Haven , CT : Yale University Press.
  • Norris, Pippa and Ronald Inglehart. 2004. Sacred and secular: Religion and politics worldwide. Cambridge , UK : Cambridge University Press.
  • Orru, Marco. 1987. Anomie: History and meanings. Boston , MA : Allyn and Unwin.
  • Penn, Mark J. and E. Kinney Zalesne. 2007. Micro-trends: The small forces behind tomorrow's big changes. New York : Hachette Book Group.
  • Perlstein, Rick. 2008. Nixonland: The rise of a president and the fracturing of America. New York : Scribners.
  • Pew Forum on Religion and Public Life. 2008. The U.S. religious landscape —2008. Retrieved November 1, 2009, from
  • Pew Forum on Religion and Public Life. 2009. A portrait of Mormons in the U.S. Retrieved November 1, 2009, from
  • Sherkat, Darren E. 2001. Tracking the restructuring of American religion: Religious affiliation and patterns of religious mobility. Social Forces 79(4):145993.
  • Simmel, Georg. [1903] 1950. Metropolis and mental life. In The Sociology of Georg Simmel, edited by KurtWolff, pp. 40927. Glencoe , IL : Free Press.
  • Smith, Christian. 1998. American evangelicalism: Embattled and thriving. Chicago , IL : University of Chicago Press.
  • Somers, Margaret R. and Fred Block. 2005. From poverty to perversity: Ideas, markets, and institutions over 200 years of welfare debate. American Sociological Review 70(2):26087.
  • Starks, Brian and Robert V. Robinson. 2005. Who values the obedient child now? The religious factor in adult values for children, 1986–2002. Social Forces 84(1):34359.
  • Steensland, Brian, Jerry Z. Park, Mark D. Regnerus, Lynn D. Robinson, W. Bradford Wilcox, and Robert D. Woodberry. 2000. The measure of American religion: Toward improving the state of the art. Social Forces 79(1):291318.
  • Stouffer, Samuel A. 1955. Communism, conformity, and civil liberties: A cross section of the nation speaks its mind. Garden City , NY : Doubleday.
  • Tipton, Steven M. 2007. Public pulpits: Methodists and mainline churches in the moral argument of public life. Chicago , IL : University of Chicago Press.
  • Totten, Mark. 2006. A new agenda for U.S. evangelicals. Christian Science Monitor. Retrieved January 4, 2007, from
  • Verba, Sidney, Kay Lehman Schlozman, and Henry Brady. 1996. Voice and equality. Cambridge , MA : Harvard University Press.
  • Wald, Kenneth D. 2003. Religion and politics in the United States, 4th ed. Boulder , CO : Rowman and Littlefield.
  • Wald, Kenneth D., Dennis E. Owen, and Samuel S. Hill. 1988. Churches as political communities. American Political Science Review 82(2):53148.
  • Warner, R. Stephen. 1988. New wine in old wineskins: Evangelicals and liberals in a small-town church. Berkeley , CA : University of California Press.
  • Wilcox, Clyde and Carin Larson. 2006. Onward Christian soldiers? The religious right in American politics, 3rd ed. Boulder , CO : Westview Press.
  • Wilcox, W. Bradford. 2006. Soft patriarchs, new men: How Christianity shapes fathers and husbands. Chicago , IL : University of Chicago Press.
  • Williams, Rhys H. 2009. Politicized evangelicalism and secular elites: Creating a moral other. In Evangelicals and democracy in America, vol. 2: Religion and politics, edited by StevenBrint and Jean ReithSchroedel, pp. 14378. New York : Russell Sage Foundation Press.
  • Woodberry, Robert D. and Christian Smith. 1998. Fundamentalism et al.: Conservative Protestants in America. Annual Review of Sociology 24(1):2556.
  • Wuthnow, Robert. 1988. The restructuring of American religion. Princeton , NJ : Princeton University Press.
  • Wuthnow, Robert. 2009. The cultural capital of evangelicals. In Evangelicals and democracy in America, vol. 1: Religion and society, edited by StevenBrint and Jean ReithSchroedel, pp. 2743. New York : Russell Sage Foundation Press.
  • Yung, Yiu-Fai and Peter M. Bentler. 1994. Bootstrap-corrected ADF test statistics in covariance structure analysis. British Journal of Mathematical and Statistical Psychology 47(1):6384.