A Complex Religion Approach to the Differing Impact of Education on Black and White Religious Group Members’ Political Views

This paper examines the interaction of education for both Blacks and Whites in all major religious groups on four key political issues: Abortion, gay marriage, feelings toward redistribution, and political party identification. We find that for most Blacks, race is the most salient factor across all four political dimensions; whereas there is significant variation by religion and education for Whites, there is very little difference for Blacks. As previous research has noted, Blacks are generally more conservative on gay marriage and Blacks are generally positive about redistribution, much more so than most Whites regardless of education and religion. We find education is more liberating to Whites than Blacks. The only issue for which education has significant effects for Blacks is abortion, but even in this case, unlike for Whites, there are not large religious differences among Blacks. This study corroborates previous research that abortion and gay marriage are less politically central to Blacks, who at all education levels are more likely to be Democrat than the most Democrat identified Whites.


1
We use race as a variable to imply the role of racial group membership within a stratified society, not innate biological or cultural characteristics.

Race and Views on Gay Marriage
Black Americans are much more conservative on the issue of homosexuality than Whites (Loftus 2001;Lewis 2003;Perry et al. 2013;Dillon 2014). There is evidence that the gap in disapproval of gay marriage is widening as support is increasing at a faster rate for Whites than Blacks (Sherkat et al. 2010).
There is also evidence that these differences are heavily dependent upon religious views. Studies that control for denominational affiliation almost cut the difference in Whites being more supportive of same-sex marriage than Blacks in half (Egan and Sherrill 2009;Sherkat et al. 2010), primarily because some White religious groups have much more tolerant views towards homosexuality than others. Although some scholars have indicated Blacks are less supportive of homosexuality as a result of the Black Church (Schulte and Battle 2004;Ward 2005;Cohen 1999), others argue that this is connected to their more traditional beliefs in gender roles in general (Dalton 1989;Green 2007;Lemelle and Battle 2004;Ward 2005;Whitley et al. 2011). Others argue that Blacks' more conservative views about homosexuality are due to socialization in the Black community (Greene 2007;Ward 2005;Negy and Eisenman 2005), connected to public Black figures, such as Cleaver (1968), framing homosexuality as a White disease that was contradictory to Blackness (Rhodes 2017). Still, how this relates to variations in education or religious affiliation within the Black community has not received much academic attention.

Race and Views on Redistribution
The picture changes completely once we begin to examine feelings toward redistribution and political party identification. Blacks have a greater commitment to liberal economic policies, including faith-based welfare reform (Bartkowski and Regis 2003), and a much less individualistic and more structural understanding of poverty (Cobb et al. 2015) than Whites (Alesina and Giuliano 2009;Davis and Robinson 1996b;Emerson et al. 1999;Mayrl and Saperstein 2013;Pyle 1993;Wald and Calhoun-Brown 2014). Wealthy Blacks are more likely to identify with their race than class (Erikson and Tedin 2015), strongly believe that their economic progress is dependent on the success of Blacks as a group (Tate 1994), and are almost as supportive as poor Blacks of social welfare policies like redistribution (Dawson 1995;Erikson and Tedin 2015). Theorists attribute this to the fact that Blacks are structurally oppressed, so there is a racialized sense of the social structure that does not exist for Whites (Omi and Winant 2014).
As for potential religious effects, Davis and Robinson (1996a) found that the Orthodox (defined as those who hold beliefs in "biblical inerrancy and a God-directed world") are more favorable to economic redistribution than progressives (p. 758). This is of particular importance because "with strong Evangelical Protestant roots, [Blacks] are much more likely than [Whites] to hold orthodox theological beliefs" (McDaniel and Ellison 2008, p. 182). In general, White Protestants are less approving of economic redistribution compared to those of no religion, Catholics, and Jews, who are more supportive (Alesina and Giuliano 2009), while Black Protestants the most likely to have structural interpretations of poverty (Hunt 2002). Views toward redistribution are certainly mediated by education. Edgell and Tranby (2007) found that religious conservatives are more economically conservative on charity for Blacks as they become more educated (Edgell and Tranby 2007). However, this inverse relation may not hold true for other White religious affiliations or racial groups.

Disaggregating Diversity within and Outside the Black Church
The "Black Church" is a reference to the pluralism of Black Christian churches in the United States (Lincoln and Mamiya 1990). In response to racism and discrimination, Black churches and denominations were founded independently from their White counterparts (Wilmore 1983). Within a segregated and oppressive society, the Black Church developed as the medium for Black society (Wald and Calhoun-Brown 2014) and acted as a unifying force for Blacks that helped advocate for Black civil rights (McDaniel 2008).
Many scholars have argued that the experiences of Blacks living as a subjugated group in an unequal society have shaped the sociopolitical nature of the Black Church (Greenberg 2000;Frazier 1974;Lincoln 1974;Lincoln and Mamiya 1990;McDaniel 2008;McDaniel and Ellison 2008). Thus, while Black and White denominations may be theologically and organizationally alike, some scholars have found that they highlight disparate elements of Christian doctrine (Gilkes 1980;Paris 1985;Pattillo-McCoy 1998), suggesting there will be a difference between racial groups of the same denomination. Whereas many Black churches (see Spence 2015 for exceptions) emphasize structures of oppression (Gilkes 1980) and biblical resistance to oppression (Harris 1999), the equality of all people (Paris 1985), and the connection between the individual and the community (Pattillo-McCoy 1998), White churches often stress individualistic ideas (Hinojosa and Park 2004;Lincoln and Mamiya 1990). It is therefore likely that we will see a difference in support for economic redistribution.
In this paper, we do not dispute the institutional and ideological importance of religion for African Americans. However, we note that most analyses that examine Blacks' political views include only Black Protestants. About three-quarters of Blacks are Protestant, the vast majority of whom are theologically conservative and demonstrate traditionalism in beliefs, frequency of Bible reading, and church attendance (Greeley and Hout 2006). 8 However, 18% of Blacks report no religion today-comparable, albeit slightly lower, to the percentage of Whites who do so (Pew Religious Landscape Study 2015). While non-religious Whites have received scholarly attention, the literature on non-religious Blacks is sparse. In addition to the non-religious, 5% of Blacks are Roman Catholic (Feagin 1968;Pattillo-McCoy 1998) and 3% are "Other" religions-mostly Muslim. Thus, more than a quarter of Black voters are not Black Protestants and are therefore missing from the discussion of how the intersections of religion and race affect political opinions until now.

Complex Religion
Given the importance of the religion for Blacks socially and politically, Blacks' overall educational disadvantage, and how these two factors may be coupled with Blacks' support for liberal economic policies, this paper attempts to untangle the importance of race, religion, and education for Blacks' political views. In doing so, the analysis that follows is theoretically informed by the "complex religion" approach. The term "complex religion" builds on theories of complex inequality (Choo and Ferree 2010;McCall 2001). Researchers of complex inequality argue that social stratification is multidimensional. Different kinds of disadvantages lead to different outcomes and experiences. These researchers therefore urge others to examine how inequalities of gender, race, or socioeconomic status interact to create a unique impact on social experience. Complex religion extends these theories to include religious group membership among the social structures that matter for inequality (Wilde and Glassman 2016). This study seeks to disaggregate the impact of racial group membership on core political views in a way that is more conscious of internal variation, particularly socioeconomic status through examining education, across religious and racial groups.
In employing the complex religion approach, we acknowledge many possible causes for the socioeconomic differences that exist between American religious groups. While there is some, primarily older, evidence that some people may choose their religious affiliation because of their social class (Loveland 2003;Sherkat and Wilson 1995), we do not think that the vast majority of these differences result from such religious switching. Similarly, although there is some limited evidence that some religious subcultures discourage class mobility (Darnell and Sherkat 1997;Lehrer 2009), we also do not argue that they are primarily a result of religious groups encouraging or discouraging class mobility. While both of these are possible mechanisms behind some of the differences between American 8 Though we focus here on Black Christian Protestants, who are the focus of most research on race and religion in the United States, it is worth noting that scholars are also examining the intersection of race, class, gender, and religion in non-Christian faiths, especially Islam (Jamal 2005;Prickett 2015).
Religions 2020, 11, 477 6 of 37 religious groups, a theory of complex religion would posit that the majority of these differences are a result of the process of social reproduction (Bourdieu 1984) set in place long ago by variations in immigration, settlement, and mobility patterns over the course of American history.

Data and Methods
To explore whether and how educational differences between religious groups and racial groups influence political views, we interact education with religious group membership for both Blacks and Whites in our examination of four key political issues: Abortion, gay marriage, economic redistribution, and political identification. For our analysis, we use the Pew 2014 Religious Landscape Survey. The 2014 Religious Landscape Survey is a telephone survey of more than 350,000 respondents across every state in the United States, which probes issues of religious identification, social and political attitudes, and demographic measures (Pew Religious Landscape Study 2015). Although its response rate is low, Pew is preferable for our purposes because its massive sample size combined with its excellent information on religion and politics allows us to examine Blacks in comparison to Whites in a way that we cannot with smaller traditional surveys like the General Social Survey (GSS). We ran all of these analyses on the GSS and found no significant differences besides losses of statistical significance because of sample size limitations (analysis available upon request).

Dependent Variables
We look at party affiliation (Guth et al. 2006;Kohut et al. 2000;Layman 1997;Miller and Hoffman 1999;Woodberry and Smith 1998) and views on politically salient issues (Hutchings and Valentino 2004), including the issues most closely associated with the "culture wars"-namely abortion and homosexuality (Hunter 1991) and views toward redistribution. 9 Selecting these well-studied issues allows us to see the impact of taking a complex religion approach and disaggregating Black Protestants.
We assess a single ordinal indicator of support for abortion. Respondents were asked, "Do you think abortion should be legal in all cases (4), legal in most cases (3), illegal in most cases (2), or illegal in all cases (1)." The variable created for feelings toward gay marriage was also on a four-point scale. Respondents were asked, "Do you strongly favor (4), favor (3), oppose (2), or strongly oppose (1) allowing gays and lesbians to marry legally?" For political identification, the Pew question asks: "In politics today, do you consider yourself a Republican, Democrat, or Independent?" Respondents who answered that they were Independent were also asked "as of today, do you lean more to the Republican Party or more to the Democratic Party?" Of those who originally identified as independents, about a third said they leaned Democrat or Republican and the rest responded that they leaned neither way. We created a five-point scale with Republicans (1), Republican-leaning Independents (2), Independents (3), Democrat-leaning Independents (4), and Democrats (5). Since these ordinal variables have more than four categories and the sample size is large, there is little difference between ordinal regression and ordinary least square (OLS) regression. For the ease of interpretation and model parsimoniousness, we choose OLS regression. To adjust for heteroscedasticity, robust standard errors are used in all models. 10 Our question regarding attitudes toward redistribution is a bit different from the rest. Respondents were asked which statement on government aid to the poor 11 comes closest to their own views: 9 By necessity, we save for future scholars related topics such as the study of social movements (Wood 1999;Young 2002;Yukich 2013) or political cultures (Berezin 1997) and nationalism (Zubrzycki 2006) and the way they intersect with race, class, and religion. 10 We run all the models using ordered logistic models and logistic models (for attitudes toward the poor in the following section) to double-check our results. Results are almost the same in significance levels and relative effect size, and thus there are no fundamental differences between OLS models and logit models for our analyses. In addition, we add the results from logit models in the Appendix A so that readers could see and compare these results (Tables A3 and A4). 11 We use the term redistribution although the question is more closely aligned with creating equal opportunity and meeting the basic needs of the poor so they can get out of the poverty.
Religions 2020, 11, 477 7 of 37 "Government aid to the poor does more harm than good, by making people too dependent on government assistance (0)" OR "Government aid to the poor does more good than harm, because people can't get out of poverty until their basic needs are met (1)." To be consistent with the other models, we use linear probability model (LPM), which is an OLS model with a binary dependent variable. After the regression, the predicted values of respondents' views all fall within the range between 0 and 1, which is consistent with the assumption of LPM that the dependent variable should be between 0 and 1. Thus, the linear probability model works well. We used robust standard errors to adjust for heteroscedasticity.

Independent Variables
Our study examines the impact of religious affiliation and how it intersects with both race and education. We compare Mainline Protestants, Conservative Protestants, Catholics, those of other religions, and those of no religion using Pew's categorization scheme. 12 In order to examine both race and religion, we created a nine-category religion model that included racial identification. This new religion variable includes White Mainline Protestants, White Conservative Protestants, White Catholics, Whites of other religions, Whites of no religion, Black Protestants, Black Catholics, Blacks of other religions, and Blacks of no religion. In an additional set of analyses, we also explore differences among Black Protestants using a new coding scheme advanced by (Shelton and Cobb 2017). This coding scheme includes six new categories: Baptist, Methodist, Holiness/Pentecostal, Non-denominational Protestant, Historically White Evangelical Protestant, and Historically White Mainline Protestant.
We measured education in the Pew data set by recoding Pew's variable on the respondent's highest level of education ("Educ") into the corresponding years of education. We did this to make it easier to interact education with religion and race, rather than control for education as almost all previous studies have done (cf. Edgell and Tranby 2007).

Control Variables
In all of our models we control for variables known to influence political views: Age (Inglehart 1977;Nie et al. 1999;Scott 1998), religious attendance (Green et al. 1991;Inglehart 1977;Sherkat et al. 2010), 13 and gender (Green 2007;Klein 1985;Roth and Kroll 2007;Wirls 1986). We also control for rural and Southern residence because region has been found to figure into the relationship between religion and politics (Davis and Robinson 1996a;Ellison and Musick 1993;Wald et al. 1989;Woodberry and Smith 1998). We examined interactions for both of them and found that, although they are significant, their presence does not change the overall findings presented here. Descriptives for all of the variables in our model are presented on Table 1. 14 12 Because of Pew's large number of cases, the Pew religion categorization scheme is able to separate out distinct groups such as the Mormons and Jehovah's Witnesses, which sometimes get coded as Conservative Protestants, and sometimes as religious "others," depending on the religious categorization scheme used. The FUND scale developed by Tom Smith codes them as Conservative Protestants (Smith 1990), whereas Reltrad considers them "others" (Steensland et al. 2000). We initially examined both groups separately, but found that they act largely like Conservative Protestants, so we included them in that category in all analyses reported here. 13 We interacted attendance with our religion variable to see how the effect of attendance changes across different religious groups. The r-squared for each model were similar to the results of education. Graphs of predicted values available upon request. 14 All items and scales were recoded so that higher values reflect more progressive views. Note: Weight is used. Jews are not included. Frequencies of each religious tradition are rounded to integers. Age range starts from age 24 or below with 1 increase as an increment of 5 years. Table 2 presents the results of our regression analysis on all dependent variables. 15 We discuss each of these in turn below. 16 Table 2. Effects of education in ordinary least squares (OLS) models with the interaction of education and nine category religion on political views.

Abortion
To examine racial and religious differences in views toward abortion, we examine Pew's question, "Do you think abortion should be, legal in all cases, legal in most cases, illegal in most cases, illegal in all cases" for both Blacks and Whites in different religious groups by education in Figures 1 and 2. Table 2. Effects of education in ordinary least squares (OLS) models with the interaction of education and nine category religion on political views.  Note: Religious attendance, age, female, south, and urban are included as control variables. Robust standard errors in parentheses. * < 0.05, ** < 0.01, *** < 0.001.

Abortion
To examine racial and religious differences in views toward abortion, we examine Pew's question, "Do you think abortion should be, legal in all cases, legal in most cases, illegal in most cases, illegal in all cases" for both Blacks and Whites in different religious groups by education in Figures 1 and 2. Figures 1 and 2 present the findings of our regression analysis for both Blacks and Whites in different religious groups by education. 17 At lower levels of education, Blacks and Whites are not overall very different on abortion: For instance, at nine years of schooling there is little difference (0.29) between Whites of no religion (2.49) and Blacks of no religion (2.2). Previous literature has suggested that Blacks are more supportive of abortion than Whites (Strickler and Danigelis 2002), yet we find this is only true at higher levels of education. Moreover, studies have suggested this is a result of less religious (Evans 2002) Blacks; although this figure shows, Blacks of no religion have the most conservative positions, aside from Black Catholics with under 15 years of education.
Education is positively correlated to more liberal views on abortion, as is the accepted wisdom in the literature (Evans 2002;Jelen and Wilcox 2003;Kiecolt 1988;Petersen 2001), but it has much greater effects for some groups than for others-most notably little effect on White Conservative Protestants who are known to be, along with Catholics, most resistant to liberalization on this issue (Abramowitz and Saunders 2008;DiMaggio et al. 1996;Emerson 1996). Overall, the education effect is similar for Blacks and Whites on abortion, such that Black Protestants look closer to White Mainline Protestants than they do to their more theologically similar White Conservative Protestants. In fact, at the highest levels of education, Black Protestants show the highest acceptance (3.04), while White Conservative Protestants have the lowest (2.39).

Gay Marriage
We present the findings for the responses of Blacks and Whites in different religious groups by education to the question: "Do you strongly favor (4), favor (3), oppose (2), or strongly oppose (1) allowing gays and lesbians to marry legally?" in Figures 1 and 2 present the findings of our regression analysis for both Blacks and Whites in different religious groups by education. 17 At lower levels of education, Blacks and Whites are not overall very different on abortion: For instance, at nine years of schooling there is little difference (0.29) between Whites of no religion (2.49) and Blacks of no religion (2.2). Previous literature has suggested that Blacks are more supportive of abortion than Whites (Strickler and Danigelis 2002), yet we find this is only true at higher levels of education. Moreover, studies have suggested this is a result of less religious (Evans 2002) Blacks; although this figure shows, Blacks of no religion have the most conservative positions, aside from Black Catholics with under 15 years of education.
Education is positively correlated to more liberal views on abortion, as is the accepted wisdom in the literature (Evans 2002;Jelen and Wilcox 2003;Kiecolt 1988;Petersen 2001), but it has much greater effects for some groups than for others-most notably little effect on White Conservative Protestants who are known to be, along with Catholics, most resistant to liberalization on this issue (Abramowitz and Saunders 2008;DiMaggio et al. 1996;Emerson 1996). Overall, the education effect is similar for Blacks and Whites on abortion, such that Black Protestants look closer to White Mainline Protestants than they do to their more theologically similar White Conservative Protestants. In fact, at the highest levels of education, Black Protestants show the highest acceptance (3.04), while White Conservative Protestants have the lowest (2.39).

Gay Marriage
We present the findings for the responses of Blacks and Whites in different religious groups by education to the question: "Do you strongly favor (4), favor (3), oppose (2), or strongly oppose (1) allowing gays and lesbians to marry legally?" in Figures 3 and 4.   Figures 3 and 4 show, in contrast to their views on abortion on which education is correlated modestly with more progressive views, education has a much smaller effect on Black Protestants' views on homosexuality. On this issue, like White Conservative Protestants, they remain relatively conservative regardless of their educational backgrounds. In comparison, education is liberalizing for other Black religious groups.
Unlike abortion, Blacks of no religion are more progressive than most other black religious groups, after 11 years of education. Still, even the most progressive Blacks remain more conservative than all other Whites in groups except Conservative Protestants.    Figures 3 and 4 show, in contrast to their views on abortion on which education is correlated modestly with more progressive views, education has a much smaller effect on Black Protestants' views on homosexuality. On this issue, like White Conservative Protestants, they remain relatively conservative regardless of their educational backgrounds. In comparison, education is liberalizing for other Black religious groups.
Unlike abortion, Blacks of no religion are more progressive than most other black religious groups, after 11 years of education. Still, even the most progressive Blacks remain more conservative than all other Whites in groups except Conservative Protestants.   Figures 3 and 4 show, in contrast to their views on abortion on which education is correlated modestly with more progressive views, education has a much smaller effect on Black Protestants' views on homosexuality. On this issue, like White Conservative Protestants, they remain relatively conservative regardless of their educational backgrounds. In comparison, education is liberalizing for other Black religious groups.
Unlike abortion, Blacks of no religion are more progressive than most other black religious groups, after 11 years of education. Still, even the most progressive Blacks remain more conservative than all other Whites in groups except Conservative Protestants.
Other than the strong effect of education for Blacks of no religion, there is, by and large, much less variation among Blacks regarding gay marriage. 18 This means that at their most highly educated, Blacks vary from a high of 2.8 for Blacks of no religion to a low of 2.5 for Black Catholics on gay marriage (mid-way between favoring and opposing gay marriage). This is much less variation than compared to the variation in Whites, where Whites of no religion are a 3.2-far above the score that represents "favoring" legal gay marriage-but White Conservative Protestants are a 2.3-close to unilaterally "strongly opposing" legal gay marriage.
When looked at together, these findings demonstrate that when it comes to gay marriage, most Black religious groups are more conservative than most White religious groups, with the exception of White Conservative Protestants, who again are the most conservative group. While Conservative Protestants have the most negative views of gay marriage (Barringer et al. 2013;Evans 2002;Olson et al. 2006;Rayside and Wilcox 2012;Sherkat et al. 2010) and the most educated Americans are the most positive towards gay marriage (Bolzendahl and Brooks 2005;Brooks and Manza 1997b;Ellison and Musick 1993;Evans 2002;Fowler et al. 2013;Kiecolt 1988;Loftus 2001;Ohlander et al. 2005), research that examines religious affiliation and attitudes towards gay marriage universally controls for education and/or race. Thus, the extensive difference between the most and least educated Blacks of other religions, and the limited effect of education on Whites in Conservative Protestant denominations and all other Blacks in religious groups, has to our knowledge not been demonstrated.

Redistribution
Respondents were asked which statement on government aid to the poor comes closest to their own views: "Government aid to the poor does more harm than good, by making people too dependent on government assistance (0)" OR "Government aid to the poor does more good than harm, because people can't get out of poverty until their basic needs are met (1)." Our findings in Figures 5 and 6 are consistent with previous research.
Religions 2020, 11, x FOR PEER REVIEW 12 of 37 Other than the strong effect of education for Blacks of no religion, there is, by and large, much less variation among Blacks regarding gay marriage. 18 This means that at their most highly educated, Blacks vary from a high of 2.8 for Blacks of no religion to a low of 2.5 for Black Catholics on gay marriage (mid-way between favoring and opposing gay marriage). This is much less variation than compared to the variation in Whites, where Whites of no religion are a 3.2-far above the score that represents "favoring" legal gay marriage-but White Conservative Protestants are a 2.3-close to unilaterally "strongly opposing" legal gay marriage.
When looked at together, these findings demonstrate that when it comes to gay marriage, most Black religious groups are more conservative than most White religious groups, with the exception of White Conservative Protestants, who again are the most conservative group. While Conservative Protestants have the most negative views of gay marriage (Barringer et al. 2013;Evans 2002;Olson et al. 2006;Rayside and Wilcox 2012;Sherkat et al. 2010) and the most educated Americans are the most positive towards gay marriage (Bolzendahl and Brooks 2005;Brooks and Manza 1997b;Ellison and Musick 1993;Evans 2002;Fowler et al. 2013;Kiecolt 1988;Loftus 2001;Ohlander et al. 2005), research that examines religious affiliation and attitudes towards gay marriage universally controls for education and/or race. Thus, the extensive difference between the most and least educated Blacks of other religions, and the limited effect of education on Whites in Conservative Protestant denominations and all other Blacks in religious groups, has to our knowledge not been demonstrated.

Redistribution
Respondents were asked which statement on government aid to the poor comes closest to their own views: "Government aid to the poor does more harm than good, by making people too dependent on government assistance (0)" OR "Government aid to the poor does more good than harm, because people can't get out of poverty until their basic needs are met (1)." Our findings in Figures 5 and 6 are consistent with previous research. There is significant variation among all Whites in religious groups, and strong variation by education for Whites as Figure 5 shows. Like abortion and gay marriage, the gap in support widens with increased educational attainment. We can see there is little effect of education among Blacks. For Whites, increases in education are correlated to more progressive views for all religious groups outside of White Conservative Protestants. Whites in Conservative Protestant denominations had the lowest impact of education on more progressive positions for abortion and gay marriage within White religious groups, and are the only religious group who get more conservative significantly on redistribution with education. Figure 6 also shows there is little variation among Blacks in relation to the issue of redistribution-with all Blacks in religious groups being among the most progressive groups in America. There is a wide disparity between racial groups. Whites in any religious group at any level of education is more progressive on redistribution than members of the most progressive Black religious groups after 10 years of education (0.7). Only Whites of other religions (0.61) and those of no religion (0.69) at the highest level of education come close to or surpass the Blacks progressivism in relation to economic equality: Those or no religion with between nine years of education (0.61) and 13 years (0.68) or Black Catholics of all levels of education (0.66).

Political Party Identification
The picture regarding political identification for Blacks by religion and education is remarkably similar to that regarding redistribution. Our five-point scale with Republicans (1), Republican-leaning Independents (2), Independents (3), Democrat-leaning Independents (4), and Democrats (5) is presented in Figures 7 and 8. There is significant variation among all Whites in religious groups, and strong variation by education for Whites as Figure 5 shows. Like abortion and gay marriage, the gap in support widens with increased educational attainment. We can see there is little effect of education among Blacks. For Whites, increases in education are correlated to more progressive views for all religious groups outside of White Conservative Protestants. Whites in Conservative Protestant denominations had the lowest impact of education on more progressive positions for abortion and gay marriage within White religious groups, and are the only religious group who get more conservative significantly on redistribution with education. Figure 6 also shows there is little variation among Blacks in relation to the issue of redistribution-with all Blacks in religious groups being among the most progressive groups in America. There is a wide disparity between racial groups. Whites in any religious group at any level of education is more progressive on redistribution than members of the most progressive Black religious groups after 10 years of education (0.7). Only Whites of other religions (0.61) and those of no religion (0.69) at the highest level of education come close to or surpass the Blacks progressivism in relation to economic equality: Those or no religion with between nine years of education (0.61) and 13 years (0.68) or Black Catholics of all levels of education (0.66).

Political Party Identification
The picture regarding political identification for Blacks by religion and education is remarkably similar to that regarding redistribution. Our five-point scale with Republicans (1), Republican-leaning Independents (2), Independents (3), Democrat-leaning Independents (4), and Democrats (5) is presented in Figures 7 and 8.  The low variation among Blacks compared to Whites is visible within Figures 7 and 8. Blacks of all religious persuasions are more Democrat leaning (3.8-4.5) than all White groups. It also demonstrates that at all levels of education, all Black groups are more likely to identify with the Democratic party than any White group, even among Whites with other religions and Whites of no religion with the highest levels of education.
Black Americans with no religion are the only Black subgroup that are significantly impacted by education. Unlike their White counterparts of no religion who tend to be more progressive and Democratically identified than other White groups, Blacks of no religion are the least Democratically  The low variation among Blacks compared to Whites is visible within Figures 7 and 8. Blacks of all religious persuasions are more Democrat leaning (3.8-4.5) than all White groups. It also demonstrates that at all levels of education, all Black groups are more likely to identify with the Democratic party than any White group, even among Whites with other religions and Whites of no religion with the highest levels of education.
Black Americans with no religion are the only Black subgroup that are significantly impacted by education. Unlike their White counterparts of no religion who tend to be more progressive and Democratically identified than other White groups, Blacks of no religion are the least Democratically The low variation among Blacks compared to Whites is visible within Figures 7 and 8. Blacks of all religious persuasions are more Democrat leaning (3.8-4.5) than all White groups. It also demonstrates that at all levels of education, all Black groups are more likely to identify with the Democratic party than any White group, even among Whites with other religions and Whites of no religion with the highest levels of education.
Black Americans with no religion are the only Black subgroup that are significantly impacted by education. Unlike their White counterparts of no religion who tend to be more progressive and Democratically identified than other White groups, Blacks of no religion are the least Democratically identified Blacks. In fact, Blacks of no religion, below 13 years of education, are the only Black (non)religious group who report lower than Democrat-Leaning Independents (4). That said, they are still stronger Democrats at their lowest identification than any White religious groups at their highest education and Democratic affiliation.
Like abortion and gay marriage, education is significant for all White religious groups. However, similar to redistribution, White Conservative Protestants are unique to their White counterparts in the impact of education: Higher levels of educational attainment are correlated with more Republican identifications. This is true also for White Catholics, albeit to a lesser, but still significant, extent.

Differences among Black Protestants
The analysis presented above examines the key religious groups in the US for both Blacks and Whites. As the existing research to date almost exclusively examines Black Protestants on these issues, if it examines Blacks at all, the addition of Black Catholics, Blacks of other religions, and Blacks of no religion is a contribution, in and of itself. Our findings suggest that while many have attributed Blacks' progressive stances on redistribution and strong support for the democratic party to the role of the Black Church-it seems that race is a much bigger factor than religion on all four of the issues we examine here.
However, up until now, we have not examined any possible differences among the three-quarters of Blacks who identify as Black Protestants. When we do so, using the new coding scheme developed by (Shelton and Cobb 2017), we can see that there is substantial variation among Black Protestants. Table 3 presents the results of our regression analysis on all dependent variables. We discuss each of these in turn below.  Of all the political issues, education has the largest effect on views toward abortion for all Black Protestant groups, besides Methodists, as seen in Figure 9. This is in alignment with the findings from our nine-category race and religion model where education was not significant for Black Protestants' identification with a political party and had minimal effect on views toward gay marriage and redistribution.
Religions 2020, 11, x FOR PEER REVIEW 16 of 37 Protestants' identification with a political party and had minimal effect on views toward gay marriage and redistribution. Not including Methodists, at nine years of education, all groups are relatively unsupportive of abortion (with scores contained largely between 2.1 and 2.2). Although the influence of education was significant when looking at Black Protestants as a group, education is not significant for Methodists, who hold the most progressive views of all Black Protestant groups from nine to 15 years of education, and those in Historically White Evangelical denominations, who remain conservative like their White Conservative Protestant counterparts.
At the highest levels of education, Blacks in Historically White Mainline denominations, Baptists, and Methodists indicate abortion should be legal in most cases. This more closely resembles Not including Methodists, at nine years of education, all groups are relatively unsupportive of abortion (with scores contained largely between 2.1 and 2.2). Although the influence of education was significant when looking at Black Protestants as a group, education is not significant for Methodists, who hold the most progressive views of all Black Protestant groups from nine to 15 years of education, and those in Historically White Evangelical denominations, who remain conservative like their White Conservative Protestant counterparts.
At the highest levels of education, Blacks in Historically White Mainline denominations, Baptists, and Methodists indicate abortion should be legal in most cases. This more closely resembles the acceptance we saw of Black Protestants as a group than the more conservative beliefs of those in Non-denominational, Historically White Evangelical, and Holiness/Pentecostal denominations. These findings corroborate the importance of distinguishing denominational affiliation for Black Protestants.
Compared to the range in support for abortion, Figure 10 19 shows even more variation in views toward gay marriage at the highest levels at education.
Religions 2020, 11, x FOR PEER REVIEW 17 of 37 the acceptance we saw of Black Protestants as a group than the more conservative beliefs of those in Non-denominational, Historically White Evangelical, and Holiness/Pentecostal denominations. These findings corroborate the importance of distinguishing denominational affiliation for Black Protestants.
Compared to the range in support for abortion, Figure 10 19 shows even more variation in views toward gay marriage at the highest levels at education. To increase readability, we begin at 1.75 on the X-axis. All figures in the Appendix A begin at the same point of the X-axis. Please see Figure A5. At the highest levels of education, Holiness/Pentecostals are again the most conservative and oppose gay marriage with a score of 1.9. This is a full point lower on the Likert scale of support than the most progressive group, Historically White Mainline Denominations who come close to favoring gay marriage (2.9). Under 12 years of education, Historically White Evangelicals are the most likely to oppose gay marriage; however, at the lowest level of education, this is less than half a point lower on the scale of support than Methodists (the most progressive group). Like abortion, the gap in support widens with educational attainment.
When looking at Black Protestants as a whole, education has a significant but small effect (0.044) on views toward gay marriage. However, when we take a closer examination at different denominations, education is significantly more likely to be connected with more progressive views for Baptists, Methodists, and those in Historically White Mainline Protestant denominations. Unlike any other Blacks in religious groups, education increases demonstrate more conservative viewpoints for Holiness/Pentecostals, although this effect is not significant. These findings reiterate the need to disaggregate the Black Protestant category. Figure 11 shows that all Black religious groups are for redistribution. Education does not significantly influence any Black Protestant denomination's views toward redistribution.
In terms of views toward redistribution, those in Historically White Mainline denominations are the most progressive at all years of education. Still, the variation amongst groups decreases throughout additional years of education at the highest years of education, and the difference between the most progressive (Historically White Mainline at 0.8) and least progressive (Historically White Evangelicals at 0.7) is marginal. This is a change from views on abortion and gay marriage, where Holiness/Pentecostals are the least progressive group at higher levels of education; however, they still hold the most conservative under 14 years of education.
As would be predicted from our nine-category race and religion model, the results of political identification for Black Protestants notably resemble the picture presented regarding redistribution. Figure 12 shows the lack of education effect we saw in Black Protestants as a group on political identification is mirrored closely by all denominations but Baptists.
Baptists are the only Black Protestant group who are significantly influenced by education in their political party identification. While increased education demonstrates more Democrat identification for Baptists, it is somewhat (but not significantly) related to more Republican affiliations for those in Non-denominational, Historically White Evangelical, and Historically White Mainline denominations.
Here we see that Historically White Evangelicals are again among the less progressive religious groups.
Only three groups surpass the threshold of 4.5 (mid-way between Democrat and Democrat-leaning Independent). Methodists at all years of education, Baptists after 14 years of education, and those in Historically White Mainline denominations before 15 years of education. Still, at all levels of education, every Black Protestant group is above Democrat-leaning Independent (4).

Conclusions
As this research shows, taking a complex religion approach does not mean that one will always find a greater religious effect. In this case, for Blacks, the intersection of religion, race, and education, suggests that beyond a "general cultural" effect that is very difficult to measure, religion and education have very little effects on Blacks' political views. Blacks of all religious affiliations (including and especially those of no religion) and of all levels of education are politically liberal.

Conclusions
As this research shows, taking a complex religion approach does not mean that one will always find a greater religious effect. In this case, for Blacks, the intersection of religion, race, and education, suggests that beyond a "general cultural" effect that is very difficult to measure, religion and education have very little effects on Blacks' political views. Blacks of all religious affiliations (including and especially those of no religion) and of all levels of education are politically liberal. The differences among Blacks of different religious views are generally quite small, suggesting that Blacks' political views are much more about race than about religion or education. This is the case except, as we have already noted, for Blacks who attend Historically White Evangelical Churches. They look much more like White Evangelicals than Blacks on issues of economic redistribution and political party identification.
Strikingly similar to moderate (non-Catholic or Conservative Protestant) Whites on abortion, Blacks in all religious groups become more accepting of abortion as they become more educated. The same is true of gay marriage, although Black Protestants' views are less impacted by educational attainment on this issue, and Blacks overall are more conservative than all Whites except White Conservative Protestants. Taking a very broad view of politics, we can see that Blacks' political identities seem disconnected from the social issues of abortion and gay marriage, which are quite divisive, and politically central, among Whites. Instead, we see that Blacks' political identities are more connected to views of economic redistribution, and that race continues to hold primary significance in the political opinions of Blacks in the United States.
Even so, however, we revealed meaningful denominational variation in Black Protestants, a religious category that is rarely disaggregated. At the highest level of education, Blacks in Historically White Mainline denominations were the most progressive group on views toward abortion, support of gay marriage, and feelings toward redistribution; while those in Historically White Evangelical denominations held either the most conservative or near the most conservative beliefs on these issues of all Black Protestant groups. While education was correlated with more progressive views for Blacks in Historically White Mainline denominations on views toward abortion and gay marriage, education had no effect for those in Historically White Evangelical denominations for any political issues. This demonstrates the extent that Blacks' political views and interaction with education vary by denominational affiliation and the flaws with treating Black Protestants as a monolith.
Our findings corroborate Shelton and Cobb (2017) findings that Holiness/Pentecostals are the most religiously conservative Black Protestants, with those in Non-denominational and Historically White Evangelical denominations falling closely behind. While Baptists and Methodists were consistently among the most progressive for all four political issues, the difference with more conservative Non-denominational and Holiness/Pentecostal groups was most apparent in the social issues of abortion and gay marriage. This supports the need for future research to examine Black Protestant denomination groups separately, particularly for beliefs on sexual morality.
Future research should look at why Blacks with no religion are the most conservative Blacks. Future research should also explore why education seems have much less of an effect on Blacks' political views than Whites', perhaps by examining whether education has a different impact on the material lives of Blacks and Whites and how that influences political views.
This paper points to the need for employing complex religion in empirical analysis. Consistent with existing research, we find that while Blacks, regardless of religious identification and education levels, lean to the left on aid to the poor and are more likely to be Democrats, they are more conservative than Whites on abortion and gay marriage. However, we also find that Blacks who attend Historically White evangelical churches have political views that more closely resemble Whites who attend such churches, especially among those who have more years in full-time education. We also found Blacks of no religion to be the most conservative Blacks. Thus, this study demonstrates the importance of further research on Black religiosity and politics to examine in greater detail non-protestant Blacks and non-religious Blacks.