Introduction
Over the course of his two terms in office, President Barack Obama faced an increasingly extremist Republican Party that wanted nothing more than to see his policy initiatives—and him personally—fail at every turn. In an illustrative episode from January 2009, Republican House members at a policy retreat
were convinced that even if Obama kept winning policy battles, they could win the broader messaging war simply by remaining unified and fighting him on everything. Their conference chairman, … Mike Pence, underscored the point with a clip from Patton, showing the general rallying his troops …: ‘We’re going to kick the hell out of him all the time! We’re going to go through him like crap through a goose!’, [exemplifying their] strategy of kicking the hell out of Obama all the time, treating him not just as a president from the opposing party but an extreme threat to the American way of life (Grunwald Reference Grunwald2016).
This hostile obstructionism was communicated directly to the electorate as elite portrayals of Obama as “a socialist tyrant intent on destroying America” (Grunwald Reference Grunwald2016) took hold. The Tea Party movement quickly came into being after Obama’s 2009 inauguration, and conservative media outlets remained laser-focused on anti-Obama messaging. Surveys indicated that Republicans who disapproved of Obama’s job performance were inclined to disapprove of it strongly (Heimlich Reference Heimlich2011). By 2016, Obama’s approval ratings were more polarized along partisan lines than they had been for any other president in the modern era.Footnote 1
A significant subtext in the story of anti-Obama sentiment among elite- and mass-level Republicans concerns the roles that religion may have played in reinforcing assessments of the president and his policy initiatives (Guth Reference Guth and Steven2011; Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017). Religion is rarely a central aspect of the Democratic “brand” for strategic political reasons (Tam Reference Tam2020), but Obama stands out for how germane religious themes were in his campaign and to his overall image. The personal religious backgrounds of presidential candidates have always been a matter of political relevance and intense interest (e.g., Smith Reference Smith2015). As he was the first Black American elected to the presidency, though, Obama’s religious identity drew even more attention from voters than usual (Block Reference Block2011).Footnote 2 As befitted someone who cut his political teeth in faith-based community organizing, candidate Obama worked hard to make inroads with various religious constituencies and did so to some success (Guth Reference Guth and Steven2011). Nevertheless, his religious commitments became a major controversy during his first run for the presidency (Walker and Smithers Reference Walker and Smithers2009). His infamous “they cling to guns or religion” comment about working-class Americans on the campaign trail raised questions—and ire among conservative Christians—about what he really thought about people of faith.Footnote 3 Also, Obama’s middle name (Hussein) and exposure to Islam as a child made him susceptible to the conspiracy theory that he is secretly a Muslim (Hollander Reference Hollander2010; Layman et al. Reference Layman, Kalkan and Green2014; Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017). To correct misperceptions about his religion and respond to criticism of his pastor at the time, Rev. Jeremiah Wright, then-Senator Obama spent a great deal of time sharing the details of his religious beliefs with the public. Even so, he failed to generate much support among the most conservative people of faith in the United States, especially evangelical Protestants.Footnote 4
In this paper, we examine the effects of religious variables on Obama’s policy-related job performance by modeling approval of his handling of seven different policy areas. We emphasize policy approval as opposed to general job approval because as president, Obama was keenly focused on public policy, almost to the exclusion of politics (Zelizer Reference Zelizer2018). Specifically, we ask: How do individual-level religious differences influence Americans’ evaluations of Obama’s policy-specific performance? Is approval affected by different dimensions of religion (i.e., religious affiliation versus religiosity)? Does religion’s influence vary across policy areas? And how does religion’s effect on policy approval intersect with the effects of race and ethnicity?
We expect that religious affiliation plays a central role in shaping Americans’ attitudes about Obama’s policy-specific job performance. Even when the public lacks details about a specific policy, they are capable of judging the results of that policy by drawing upon heuristics (Mondak Reference Mondak1993; Popkin Reference Popkin1991; Sniderman et al. Reference Sniderman, Brody and Tetlock1991; Sniderman et al. Reference Sniderman, Hagen, Tetlock and Brady1986; Zaller Reference Zaller1992). In this context, religion, especially religious affiliation, is a powerful and often-used heuristic (e.g., Madrid et al. Reference Madrid, Merolla, Ruiz and Schroedel2022; Westfall and Russell Reference Westfall and Çelik Russell2019). The specific content and effects of religion-based heuristics should differ by religious tradition. We expect adherents of two religious groups—White evangelical Protestants and Black Protestants—to stand out in their evaluations of Obama’s handling of public policy. Due to their moral and social traditionalism as well as their overwhelmingly Republican partisanship, evangelicals are likely to have embraced a particularly critical view of Obama’s policies. Black Protestants, on the other hand, should have been especially supportive of Obama’s handling of policy across the board due to their preference for an activist government and the traditional significance of race in US politics.
We test these contentions using data from the 2013 “America’s Place in the World” General Public Survey fielded by the Pew Research Center and the Council on Foreign Relations. Instead of limiting our analysis to Obama’s handling of a broad public policy domain (domestic or international), we examine the determinants of support concerning seven specific policy areas spanning both policy domains: healthcare, the economy, foreign policy in general, terrorism, immigration, international trade, and climate change.
Our findings demonstrate that religious affiliation (“belonging”) plays a crucial role in shaping public evaluations of the president’s policy performance on matters that extend well beyond socio‑moral “culture wars” issues. Compared to Black Protestants, evangelicals, non-Hispanic Catholics, and (to a lesser extent) mainline (i.e., more progressive) Protestants and the religiously unaffiliated are less likely to approve of Obama’s domestic and international policy performance, even in the face of significant partisan, demographic, and socioeconomic controls. We conclude that religious affiliation as a basic social identity has its own distinct impact (separate from partisanship, even in this polarized era) on presidential policy approval and that the intersection of race/ethnicity and religion also shapes support for Obama’s handling of policy. In addition, religious participation (measured as worship attendance) is significant in four of our seven models, suggesting that social exposure to religious cues (whether from clergy or fellow practitioners) can play an important role in shaping public attitudes about presidential policy performance.
Theory and Hypotheses
Religion’s political influence in the United States is evident on the campaign trail, in elite policy positions, and in public opinion. This influence is driven by the fact that in general, religiously committed Americans perceive politics in a systematically different—and, at least in the present era, more conservative—way than do less religious and secular Americans (Green Reference Green2007; Layman Reference Layman2001; Putnam and Campbell Reference Putnam and Campbell2010). Religious affiliation is politically significant because it is both a membership-based social network (Kertzer and Zeitzoff Reference Kertzer and Zeitzoff2017; Sinclair Reference Sinclair2012) and a core social identity (Castle and Stepp Reference Castle and Stepp2021). Even in the face of rapidly declining rates of religious affiliation, a sizable majority of Americans continue to see themselves as religious people (Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025).
Organized religion shapes believers’ moral values, which in turn shape their assessments of politics (Goren Reference Goren2005; Kertzer et al. Reference Kertzer, Powers, Rathbun and Iyer2014). A voluminous literature documents strong relationships between religious variables and attitudes about economic, international, moral, social, and cultural issues (Arbuckle and Konisky Reference Arbuckle and Konisky2015; Baumgartner et al. Reference Baumgartner, Francia and Morris2008; Castle and Stepp Reference Castle and Stepp2021; Guth et al. Reference Guth, Green, Kellstedt and Smidt1995; Guth et al. Reference Guth, Green, Kellstedt and Smidt2005; Olson et al. Reference Olson, Cadge and Harrison2006; Taydas and Olson Reference Taydas and Olson2022; Taydas et al Reference Taydas, Kentmen and Olson2012; Whitehead and Perry Reference Whitehead and Perry2020). Many Americans also depend on their religious identities when evaluating candidates and political parties; political elites encourage this tendency by including religious references in their public messaging (Claassen Reference Claassen2015; Claassen et al. Reference Claassen, Djupe, Lewis and Neiheisel2021; Domke and Coe Reference Domke and Coe2008; Madrid et al. Reference Madrid, Merolla, Ruiz and Schroedel2022; McLaughlin and Wise Reference McLaughlin and Wise2014; Westfall and Russell Reference Westfall and Çelik Russell2019). Against this backdrop, it is natural also to expect religion to influence how people judge political elites’ handling of various policies once they are in office.
Another considerable body of scholarship analyzes factors that shape general and policy-specific presidential approval (Cohen Reference Cohen2002a, Reference Cohen2002b; Donovan et al. Reference Donovan, Kellstedt, Key and Lebo2020; Druckman and Holmes Reference Druckman and Holmes2004; Edwards et al. Reference Edwards, Mitchell and Welch1995; McAvoy Reference McAvoy2006; Newman and Otto Reference Newman and Otto2022; Rogowski Reference Rogowski2023), but little work has been done on how religion might contribute to public evaluations of how well the occupant of the Oval Office is doing their job (but see Olson and Warber Reference Olson and Warber2008). Even less is known about whether and how religion might affect policy-specific presidential approval. Yet, given how tightly religious identity is woven into the fabric of political polarization in the United States (Castle and Stepp Reference Castle and Stepp2021; Perry Reference Perry2022), we should expect it to play a systematic role in evaluations of the president’s handling of public policy. Moreover, the ways in which religion shapes political outlooks are hardly uniform (Claassen Reference Claassen2015; Green Reference Green2007; Layman Reference Layman2001), so we should expect variation in how members of different religious traditions evaluate the president’s handling of public policy.
Religion as a Political Heuristic
A touchstone finding in the literature on public opinion is that citizens lack political knowledge (Carpini et al., Reference Carpini, Michael and Keeter1996). Despite not knowing much about politics, citizens—even those who pay minimal attention to politics—navigate its complexities to develop consistent points of view and structured policy preferences (Popkin Reference Popkin1991; Sniderman et al. Reference Sniderman, Brody and Tetlock1991; Sniderman et al. Reference Sniderman, Hagen, Tetlock and Brady1986; Zaller Reference Zaller1992). While this fact is particularly relevant to elections (e.g., Lau and Redlawsk Reference Lau and Redlawsk2001), it also holds true in other contexts. For example, the public regularly renders judgments about complex foreign policy matters even when they have neither sufficient information nor strong interest (Kertzer and Zeitzoff Reference Kertzer and Zeitzoff2017).
Individuals form political attitudes absent large quantities of information by using heuristics to guide their decisions. Heuristics are means of simplification that allow citizens to “figure out what they think about political issues without necessarily knowing very much about them” (Sniderman et al. Reference Sniderman, Hagen, Tetlock and Brady1986: 406). Party identification is a common heuristic; party labels help people make decisions in the voting booth without knowing much about the candidates and their policy positions (Conover & Feldman Reference Conover and Feldman1984; Rahn Reference Rahn1993). Another important cognitive shortcut citizens use when forming political opinions is alignment between their own personal background and those of candidates (McDermott Reference McDermott2007, Reference McDermott2009; Westfall & Çelik Russell, Reference Westfall and Çelik Russell2019). Characteristics including race/ethnicity (Matsubayashi and Ueda Reference Matsubayashi and Ueda2011), gender (McDermott Reference McDermott1997), and even accent (Amira et al. Reference Amira, Cooper, Gibbs Knotts and Wofford2018) can serve as effective political heuristics.
Drawing on heuristic models of belief formation, we join others (Madrid et al. Reference Madrid, Merolla, Ruiz and Schroedel2022; McDermott Reference McDermott2007, Reference McDermott2009; McLaughlin and Wise Reference McLaughlin and Wise2014; Westfall and Russell Reference Westfall and Çelik Russell2019) in arguing that religion—especially religious affiliation—can ease the process of making political judgments, including evaluations of the president’s policy performance. People of faith rely on religious heuristics when they perceive connections between their religious identity and political matters (Layman and Green Reference Layman and Green2005). In the broader context of deficient political knowledge, a large portion of the public may be assumed to lack detailed information about the president’s policy priorities and initiatives. Evaluating presidential performance on specific policy areas is a Herculean task under such circumstances. To simplify information processing and avoid engaging in time-consuming analysis of the president’s policies, people may resort to the value-laden heuristics that arise from their religious commitments.
How Race/Ethnicity Shapes Religion’s Heuristic Influence
For many Americans, the impact of religious heuristics is dependent in one way or another on race and/or ethnicity. In fact, some scholars argue that religion’s effects on attitudes and actions must not be studied outside the intersectional context(s) of race and ethnicity (Wilde Reference Wilde2018; Wilde and Glassman Reference Wilde and Glassman2016). From this standpoint, research on religion and politics must take “religion as a starting point and [acknowledge] that one cannot study religion independently of race … in the United States” (Wilde and Glassman Reference Wilde and Glassman2016: 409). Indeed, there is ample empirical evidence that religious identities intersect with race and class to shape political attitudes (Bruce Reference Bruce2020; Lockerbie Reference Lockerbie2013; O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020; Perry and Schleifer Reference Perry and Schleifer2022; Schnabel Reference Schnabel2020; Zingher Reference Zingher2021).
We suspect that this intersection would have been especially relevant regarding attitudes about the US first Black president. As Williams observes, “what many Americans [saw] in Obama’s racial and religious status was a disruption of a … connection among religion, race, and national identity…. Obama’s election encapsulated for many people their fears regarding social changes that threaten those connections” (Reference Williams2013: 243, emphasis in original). Previous studies illustrate the complicated ways in which Black and White people reacted to Obama’s race (Abrajano and Burnett Reference Abrajano and Burnett2012; Block Reference Block2011; Piston Reference Piston2010; Redlawsk et al. Reference Redlawsk, Tolbert and Franko2010; Walters Reference Walters2007; Williams Reference Williams2013). Consider the fact that Black Protestants were much more likely than other Americans to think “God chose Obama to become president because God approved of Obama’s policies,” while evangelical Protestants—70 percent of whom today are White (Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025)—were much more likely than others to say the same about Donald Trump’s first election (Pew Research Center 2020).
At least two in three Black Americans are Protestant (Diamant and Besheer Reference Diamant and Mohamed2021),Footnote 5 and the vast majority of Black Protestants voted for Obama in both 2008 (94 percent) and 2016 (95 percent) (Pew Research Center 2012). Black support for Obama was widespread despite some observers’ contention that as president, he did little to advance the specific policy concerns of Black Americans (Harris Reference Harris2012). Black Americans are, of course, heavily Democratic in their party identification (Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025). This partisanship reflects the long-standing political significance of the fact that Black Protestants “have adopted the worldview that individual freedom can be realized only within the context of collective freedom” (Dawson Reference Dawson1994: 100; see also Harris Reference Harris2000; Lincoln and Mamiya 1990; Owens Reference Owens2007).Footnote 6 While Black congregations are hardly all the same, religion and politics traditionally have been tightly interwoven in a way that points in a progressive direction (e.g., O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020). This prophetic, explicitly political impulse has long set Black Protestantism apart from the rest of American religion, and progressive preferences persist among Black Americans despite their theological conservatism (O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020).
Black Protestantism has additional political influence on its adherents due to its powerful history as one of the only social institutions over which Black Americans consistently have wielded control (Lincoln & Mamiya 1990). Black churches have acted as centers of both political organization and political power; the most significant manifestation of this fact was when Black clergy led the civil rights movement from strength to strength (Harris Reference Harris2000; Morris Reference Morris1984; Owens Reference Owens2007; Smith et al., Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025). People who attend Black churches are often mobilized for civic participation to fight injustice and improve the circumstances of their everyday lives (Brown and Brown Reference Brown and Brown2003; McClerking and McDaniel Reference McClerking and McDaniel2005; Smith et al., Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025).
HispanicFootnote 7 ethnicity also intersects with religion in politically consequential ways (Bartkowski et al. Reference Bartkowski, Ramos-Wada, Ellison and Acevedo2012; O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020). Catholics are one of the largest religious groups in the United States, and more than a third of American Catholics are Hispanic (Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025). Hispanic Catholics are substantially more progressive and Democratic in their political orientations than are White Catholics (Gayte et al. Reference Gayte, Chelini-Pont and Rozell2018; Hoover Reference Hoover2014; Matovina Reference Matovina2012; Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025). The Pew Research Center (2012) found that more than seven in ten Hispanic Catholics—compared to less than half of White Catholics—voted for Obama in 2008 and 2012. The differences between Hispanic and non-Hispanic Catholics are sufficiently large that some scholars imply that the two traditions are distinct (e.g., Calvillo and Bailey Reference Calvillo and Bailey2015; O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020).Footnote 8
Though Hispanic Americans differ politically and along many other lines (Corral and Leal Reference Corral and Leal2020; Martínez and Gonzalez Reference Martínez and Gonzalez2020; Pérez and Cobian Reference Pérez and Cobian2024), including country of origin, proximity to the immigrant experience, and educational attainment, a significant political divide separates Hispanic Catholics from Hispanic Protestants (Corral and Leal Reference Corral and Leal2020; Espinosa et al. Reference Espinosa, Elizondo and Miranda2005; Kelly & Morgan Kelly, Reference Kelly and Morgan Kelly2005; Mulder et al. Reference Mulder, Ramos and Martí2017; Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025; Valenzuela Reference Valenzuela2014). Hispanic Americans are much more likely to be Catholic than Protestant, but the share of Catholics is declining while the percentage of Protestants is staying steady (Krogstad et al. Reference Krogstad, Alvarado and Mohamed2023; Mulder et al. Reference Mulder, Ramos and Martí2017 contend that the percentage of Hispanic Protestants is growing). Hispanic Protestants are markedly more conservative and Republican than their Catholic counterparts, reflecting the fact that most are evangelical (Kelly & Morgan Kelly, Reference Kelly and Morgan Kelly2005; Krogstad et al. Reference Krogstad, Alvarado and Mohamed2023; Mulder et al. Reference Mulder, Ramos and Martí2017; Valenzuela Reference Valenzuela2014). In short, Hispanic Catholics constitute an ethno-religious group that is both sizable and politically distinctive (Calvillo and Bailey Reference Calvillo and Bailey2015; O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020).
From the standpoint of religious affiliation, Obama’s greatest detractors were White evangelical Protestants, a highly mobilized group of Americans who identify strongly with the Republican Party and prefer public policy that reinforces a traditional social order and protects the homeland (e.g., Claassen and Povtak Reference Claassen and Povtak2010; Putnam and Campbell Reference Putnam and Campbell2010; Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025). Today, evangelicals comprise the core of the White Christian nationalist movement that swept Trump into power (Everton Reference Everton2024; Whitehead and Perry Reference Whitehead and Perry2020; Zhou Reference Zhou2023). Not all White evangelicals are Christian nationalists (Whitehead and Perry Reference Whitehead and Perry2020), but it is well established that those who holds negative views toward racial minorities (Davis and Perry Reference Davis and Perry2021; McDaniel et al. Reference McDaniel, Nooruddin and Shortle2022; Perry et al. Reference Perry, Whitehead and Davis2019).
Even before he won the 2008 presidential election, Obama was a target of suspicion among White evangelicals, who questioned his background and aims; they remain wary of him to this day (Benson et al. Reference Benson, Geer and Merolla2012; Guth Reference Guth and Steven2011; Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017; Williams Reference Williams2013). When the Tea Party emerged during the early days of Obama’s first term, its leaders and adherents included large numbers of evangelicals (Bonikowski and DiMaggio Reference Bonikowski and DiMaggio2016; Braunstein and Taylor Reference Braunstein and Taylor2017; Deckman Reference Deckman2016; Parker and Barreto Reference Parker and Barreto2014). In the aggregate, White evangelicals were even more likely than Republicans to believe the false account that Obama is a Muslim: during the early years of his presidency, 17 percent of Republicans in two Pew Research Center polls said they believed Obama was a Muslim, as did 19 percent of evangelical Protestants (Rosentiel Reference Rosentiel2009; see also Layman et al. Reference Layman, Kalkan and Green2014). In short, many White evangelicals “viewed Obama as the embodiment of their broader concerns about social change” (Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017: 52; see also Guth Reference Guth and Steven2011; Williams Reference Williams2013). As traditionalist conservatives, evangelicals are skeptical of social change, and Obama represented just that both in his overt personal characteristics and his campaign theme, “change we can believe in.”
The intersectional role of race/ethnicity cannot be overlooked when considering how religion affected approval of Obama’s policies. Abrajano and Burnett (Reference Abrajano and Burnett2012) document substantial racial differences in Obama’s general approval ratings, with White Americans judging him more harshly than they had judged Bill Clinton, and Black Americans doing the opposite. Likewise, Hispanic Catholics were largely supportive of Obama (Matovina Reference Matovina2012; Pew Research Center 2012). Several studies specifically document White evangelical Protestants’ tendencies toward racial resentment and disapproval of policies designed to help ethnic minorities (Layman Reference Layman2001; McDaniel et al. Reference McDaniel, Nooruddin and Shortle2022; Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017; Whitehead and Perry Reference Whitehead and Perry2020). In addition, racial resentment directly predicted negative evaluations of Obama (Pyszczynski et al. Reference Pyszczynski, Henthorn, Motyl and Gerow2010), and racial prejudice hurt him substantially at the polls (Piston Reference Piston2010).
Based on the foregoing discussion, we propose three hypotheses about how religion intersects with race/ethnicity to influence attitudes about Obama’s handling of various policies:
H1: Black Protestants will be most likely to approve of Obama’s handling of public policy.
H2: Hispanic Catholics will be likely, but not as likely as Black Protestants, to approve of Obama’s handling of public policy.
H3: Evangelical Protestants will be least likely to approve Obama’s handling of public policy.
Data and Methods
To evaluate these hypotheses, we analyze survey data gathered by the Pew Research Center in conjunction with the Council on Foreign Relations. Their survey titled “America’s Place in the World” was in the field from late October until early November 2013; the resulting dataset includes responses from 2,003 US adults who were asked to assess Barack Obama’s handling of a variety of policy areas.Footnote 9 We selected this dataset because in addition to a rich variety of dependent variables concerning approval of Obama’s policies, the survey includes measures of respondents’ religious commitments, partisanship, and demographic characteristics. We use the weighted sample recommended by Pew to ensure that the demographics of the sample closely approximate those of the national population.
Dependent Variables
We analyze the effect of religion on seven dependent variables measuring public support for Obama’s policies. Recall that we are not modeling general approval of Obama’s job performance; we focus instead on policy-specific approval. Our first two dependent variables (healthcare and the economy) measure how respondents view Obama’s domestic policy performance.Footnote 10 The remaining dependent variables assess respondents’ attitudes about Obama’s handling of several matters pertaining to foreign policy: overall foreign policy, the threat of terrorism, immigration, international trade, and climate change. Each dependent variable is a dichotomous measure of respondents’ approval of Obama’s handling of that policy area (approve = 1; disapprove = 0).Footnote 11 Because our dependent variables are all dichotomous, we estimate logistic regression models.
Independent Variables
We include two of the four major dimensions of religion (Saroglou Reference Saroglou2011) in our analyses, both of which have been shown to have powerful effects on political attitudes (e.g., Smidt et al. Reference Smidt, Kellstedt, Guth and Corwin2009).Footnote 12 One of these measures is a categorical indicator of religious affiliation (religious “belonging”). The other is an ordinal measure of how often the respondent attends worship services (a common form of religious “behaving”). To measure religious “belonging,” we generate a series of variables indicating the organized religious tradition with which the respondent identifies: evangelical Protestant, mainline Protestant, Black Protestant, Hispanic Catholic, non-Hispanic Catholic, other Christian, any other religion, or no religion (unaffiliated). Evangelical Protestants are those who self-identify as Protestant or “just Christian,” describe themselves as evangelical or “born again,”Footnote 13 and are not Black. Mainline Protestants identify as Protestant or “just Christian,” do not describe themselves as evangelical or “born again,” and are not Black. Meanwhile, Black Protestants are just that: Protestant (or “just Christian”) and Black.Footnote 14 Because of the salience of ethnicity within US Catholicism today (Calvillo and Bailey Reference Calvillo and Bailey2015; O’Brien and Abdelhadi Reference O’Brien and Abdelhadi2020) and the political differences between White and Hispanic Catholics (Gayte et al. Reference Gayte, Chelini-Pont and Rozell2018; Hoover Reference Hoover2014; Matovina Reference Matovina2012), we follow the Pew Research Center’s practice of splitting Catholic respondents into separate “Hispanic” and “non-Hispanic” categories.Footnote 15 “Other Christian” respondents include those who identify with a Christian tradition apart from Protestantism and the Catholic Church, including Eastern Orthodoxy, the Church of Jesus Christ of Latter-day Saints, and a few more groups. Respondents are classified as “other religion” if they said they are Buddhist, Hindu, Jewish, Muslim, or Unitarian Universalist. We classify respondents having no religion if they identify as being atheist, agnostic, or “nothing in particular.” All categories are mutually exclusive.Footnote 16
We exclude Black Protestant respondents from our models as the reference group, so results for each religious tradition should be interpreted as the difference between that group and Black Protestants. Comparing all other groups to Black Protestants allows us to directly test the theoretical intersection of religion and race on presidential policy approval. We decided to make Black Protestants the contrast group for two reasons.Footnote 17 First, H1 contends that this group of Americans will be most likely to support Obama’s policies. As discussed above, Black Protestants, in the aggregate, are strong Democrats who voted overwhelmingly for Obama in both 2008 and 2012. Second, general presidential approval ratings of Obama were systematically higher among Black Americans than other groups (Abrajano and Burnett Reference Abrajano and Burnett2012). It is therefore sensible to compare other US religious groups to Black Protestants given the historic significance of Obama being the first Black president. Most Black Protestants are also accustomed to a uniquely progressive connection between religion and politics.
To capture religious “behaving,” we include an ordinal measure of how often respondents report attending religious services. Worship attendance is one of the best indicators of overall religiosity in a majority-Christian country like the United States,Footnote 18 and the political cues clergy provide during worship services do not fall on deaf ears (Djupe and Gilbert Reference Djupe and Gilbert2009; Smith Reference Smith2008; Smith and Harris Reference Smith and Harris2005). Regardless of affiliation, Americans who attend religious services more frequently are systematically more conservative and more politically participatory than those who attend less faithfully or not at all (Olson and Green Reference Olson and Green2006; Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025; Smith & Demetrius Walker, Reference Smith and Demetrius Walker2012). Survey respondents were asked how often they attend religious services (aside from weddings or funerals) on a scale ranging from never (coded as a 1) to more than once a week (coded as a 6). We include this ordinal measure in each of our models to account for the possibility that greater religiosity might affect evaluations of Obama’s policy performance.
Control Variables
We control for several attitudinal and demographic characteristics that are known to affect general presidential approval and should therefore also shape assessments of a president’s handling of various policy areas (Donovan et al. Reference Donovan, Kellstedt, Key and Lebo2020). These controls include partisanship, ideology, age, gender, education, and income. Previous research demonstrates that in recent times, US presidential approval has been dominated by partisanship (Donovan et al. Reference Donovan, Kellstedt, Key and Lebo2020; Lebo and Cassino Reference Lebo and Cassino2007). We use separate dummy variables that classify respondents’ party identification as being a Democrat, an Independent, or a Republican. Following current scholarship on Independents (Klar and Krupnikov Reference Klar and Krupnikov2016), we code individuals who first identified as Independent but then reported a partisan leaning as affiliates of the party toward which they lean. Thus, Independents are those who chose the “Independent” label and did not report leaning toward either party. We also classify respondents as Independent if they reported no preference between the parties or if they selected “other party.” We measure ideology using a question asking respondents to describe their political views on a five-point scale ranging from very liberal (1) to very conservative (5).
We include key demographic controls as well. Our age variable captures respondents’ age in years. To measure gender, we include a dichotomous male variable. To measure individuals’ education, we use responses to a question asking respondents to indicate the highest level of school they have attended or the highest degree they have earned. This ordinal variable ranges from “less than high school” (coded as 1) to “postgraduate or professional degree” (coded as 8). Finally, we include an ordinal control for total family income in the previous year before taxes; this question’s response set has nine categories ranging from 1 (less than $10,000) to 9 ($150,000 or more).
Results and Discussion
Table 1 presents the results of logistic regression models testing the effects of the independent variables on each of our dependent variables measuring attitudes about Barack Obama’s handling of seven policy areas: healthcare, the economy, foreign policy in general, terrorism, immigration, international trade, and climate change. Footnote 19 The table reports logistic regression coefficients; significant results for religion variables are indicated in boldface. In every model except for the one concerning terrorism, at least two of the religion variables reach statistical significance at the .05 level, revealing the important role religion plays in influencing attitudes toward Obama’s handling of public policy.
Table 1. Approval of Barack Obama’s handling of seven policy areas
Note: Logistic regression models, reporting coefficients with standard errors are in parentheses. Reference group: Black Protestant.
* = p < .05, ** = p < .01, *** = p < .001.
Data source: Pew Research Center and Council on Foreign Relations, “America’s Place in the World,” October 30–November 6, 2013.
As expected, religious variables—especially “belonging”—have significant, independent effects on approval of Obama’s handling of both domestic and foreign policy. The significance of the religious variables is especially noteworthy given the presence of other strong predictors of approval in our models: partisanship, ideology, and demographic controls. We find ample support for H1 in the 14 instances where a religious group is significantly less likely than Black Protestants to approve Obama’s handling of a policy area. These results reflect both the strong support Obama enjoyed among Black Americans in general (Abrajano and Burnett Reference Abrajano and Burnett2012) and a lack of support for Obama and his policies among White Christians, especially evangelical Protestants and non-Hispanic Catholics (Benson et al. Reference Benson, Geer and Merolla2012; Guth Reference Guth and Steven2011; Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017; Williams Reference Williams2013). In only one instance is a religious group significantly more supportive than Black Protestants of Obama’s handling of a policy area: Hispanic Catholics are more supportive of Obama’s handling of international trade. Perhaps this is so because while Obama’s record on Latin America received mixed reviews, he was a strong supporter of free trade with countries in the region (Reid Reference Reid2015). This result supports H2, as does the fact that non-Hispanic Catholics are more likely than their Hispanic counterparts to support Obama’s handling of policy (in comparison with Black Protestants). In four of our seven models (healthcare, the economy, foreign policy in general, and immigration), non-Hispanic Catholics—but never Hispanic Catholics—are significantly less approving of Obama’s handling of policy compared to Black Protestants. Evidence in support of H3 appears in Table 1 as well, with evangelical Protestants significantly less likely than Black Protestants to approve of Obama’s handling of five policy areas (terrorism and international trade are the exceptions). We find it surprising that the evangelical variable is not significant in two of our models: terrorism and international trade, as other studies show that evangelicals tend to be concerned about terrorism and skeptical of globalization (e.g., Taydas and Olson Reference Taydas and Olson2022).
While we did not pose a hypothesis concerning domestic versus foreign policy approval, it is noteworthy that the effects of religious affiliation are much more pronounced in the domestic policy models. For both healthcare and the economy, Black Protestants are significantly more supportive of Obama than are evangelical Protestants, mainline Protestants, and non-Hispanic Catholics, with the largest difference existing between Black Protestants and evangelicals. These findings support all three of our hypotheses. The model statistics also reflect the relative strength of the domestic policy models. Perhaps, our domestic policy results reflect previous research establishing that domestic policy issues are more immediately “accessible” in voters’ minds (Lavine et al. Reference Lavine, Sullivan, Borgida and Thomsen1996) and that publicly salient issues are most likely to shape presidential approval (Edwards et al. Reference Edwards, Mitchell and Welch1995).
Differences in religion’s effect on approval of policy performance are not universal. It is worth considering the religious groups about which we did not pose hypotheses, but which nevertheless stand out for their disapproval of Obama’s handling of one or more policy areas. Mainline Protestants are less approving than Black Protestants are of Obama’s handling of both domestic policy issues, but the two groups do not differ on any of the foreign policy matters.Footnote 20 In the aggregate, mainline Protestant laity sit to the right of Black Protestants both theologically and politically, but there is plenty of diversity, even division, among mainline rank and file around domestic policy matters. Smith and colleagues (Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025) find that while a bare majority of mainline Protestants are Republicans, relatively few identify as conservative. Small majorities of mainline Protestants voted against Obama in both 2008 and 2012 (Pew Research Center 2012). Although mainline clergy are largely liberal, they have great difficulty conveying progressive political priorities to their congregations (Djupe and Gilbert Reference Djupe and Gilbert2009; Guth et al. Reference Guth, Green, Smidt, Kellstedt and Poloma1997; Lichterman and Williams Reference Lichterman, Williams, Braunstein, Fuist and Williams2017). Our results here suggest that mainline Protestant conservatism arises in the domain of domestic, but not foreign policy, which is consistent with previous work establishing that mainline Protestants are relatively progressive concerning foreign policy (Taydas and Olson Reference Taydas and Olson2022).
The similarity evident in Table 1 between evangelicals and non-Hispanic Catholics is not unexpected. White Catholics have become more conservative and Republican over time,Footnote 21 even finding some degree of common cause with the religious right (Gayte et al. Reference Gayte, Chelini-Pont and Rozell2018). Here, we find non-Hispanic Catholics significantly less likely than Black Protestants to approve of Obama’s handling of healthcare, the economy, foreign policy in general, and immigration. These results are evidence of the distance between rank-and-file non-Hispanic American Catholics and core Catholic doctrine. Catholic Social Teaching emphasizes human dignity, the common good, and the “special option for the poor and vulnerable” (US Conference of Catholic Bishops n.d.). Catholic leaders did not agree with Obama on abortion, but on the surface his efforts to expand access to healthcare and create a path to citizenship for undocumented immigrants (the Deferred Action for Childhood Arrival program) should have been appealing to Catholic clergy and laity alike. The reality, however, is much more complicated. US Catholic bishops can be hesitant to send direct signals on political issues, and even when they do, most parishioners ignore them (Gayte et al. Reference Gayte, Chelini-Pont and Rozell2018; Smith Reference Smith2008). Recent research shows that Catholic Americans tend to align their issue positions with those of their preferred political party, not the Church (Smith Reference Smith2020). In addition, Catholic leaders strongly objected to the so-called “birth control mandate” in the Affordable Care Act, criticizing Obama for a perceived affront to religious freedom (Griffin Reference Griffin2015). These complex dynamics combine to explain some of the skepticism we find among non-Hispanic Catholics for Obama’s handling of policy.
Non-Christians (here designated as “other religion”), a majority of them are Jewish, are significantly less supportive of Obama’s handling of the economy than the Black Protestants. Religiously unaffiliated respondents are also significantly less supportive of Obama’s handling of the economy, as well as foreign policy overall, compared to the Black Protestants. These results have a basis for support in existing literature because we know that non-Christians and religiously unaffiliated Americans are markedly more liberal than their Christian counterparts (e.g., Campbell et al. Reference Campbell, Layman and Green2021; Wald Reference Wald2015). Large shares of both groups voted for Obama in 2008 and 2012 (Pew Research Center 2012). Since we control for respondents’ ideology, we can argue that the progressivism inherent in not being part of the Christian majority in the United States has its own independent effect regarding Obama’s handling of the economy and foreign policy.
Finally, we find a complete absence of significant results for “other Christians.” It is surprising that there are no apparent differences in assessments of Obama’s handling of policy between the Black Protestants and Christian respondents who are neither Protestant nor Catholic. As noted earlier, this group consists mainly of Eastern Orthodox and Mormon respondents (in our sample, the majority are Mormons). Most Mormons are conservative and Republican; their political uniformity is like that of evangelical Christians (Campbell et al. Reference Campbell, Green and Quin Monson2014). There is some suggestion that the relationship between the LDS Church and the Obama administration was never hostile (Walch and Taylor Reference Walch and Taylor2017), and while “Mormons generally follow their church over their party” (Campbell et al. Reference Campbell, Green and Quin Monson2014: 23), the whole Mormons were not fans of Obama or his policies. Meanwhile, a plurality of Orthodox Christians voted for Obama in 2008, but not 2012 (Burge Reference Burge2021), and Eastern Orthodoxy is growing in the United States thanks to a trend of conversion by political conservatives (Namee et al. Reference Namee, Metrakos, Irwin, Morgan and Hensersky2024). Thus, there are few reasons to suspect that “other Christians” would have been supportive enough of Obama’s policies to not differ significantly from Black Protestants. We attribute the absence of significant findings in these instances to small sample size.
Another noteworthy finding pertains to worship attendance, an ordinal measure of how often respondents reported attending religious services. Worship attendance is significant in four of seven models: healthcare, terrorism, international trade, and climate change. In each instance, more frequent attendance contributes to greater approval. Substantively, these findings are surprising, especially because other studies show that increased frequency of worship attendance is associated with greater political conservatism (e.g., Olson and Green Reference Olson and Green2006). However, since we control for ideology and partisanship, these significant and positive effects of worship attendance are independent of conservatism. We suspect that our findings for worship attendance might be, at least in part, an artifact of the powerful negative effect of evangelical Protestant affiliation, as evangelicals are twice as likely as Americans at large to attend services on a weekly basis (Smith et al. Reference Smith, Cooperman, Alper, Mohamed, Rotolo, Tevington, Nortey, Kallo, Diamant and Fahmy2025).Footnote 22 Sermons contain a fair amount of political content, and by no means does all of it reflect conservative political orientations (Boussalis et al. Reference Boussalis, Coan and Holman2021; Guth et al. Reference Guth, Green, Smidt, Kellstedt and Poloma1997; Olson and Quezada-Llanes Reference Olson, Quezada-Llanes, Anand and Paul2024). One large-scale study of sermon content, however, shows that healthcare, terrorism, international trade, and even climate change are relatively less likely than other issues to be included in sermons (Boussalis et al. Reference Boussalis, Coan and Holman2021). The significant effects of worship attendance on approval of Obama’s handling of these four policy areas may therefore reflect some embeddedness in religious social networks other than the influence of what clergy say.Footnote 23
As for our control variables, we find strong, statistically significant effects of both partisanship and ideology across most, if not all, models. Not surprisingly, respondents who identify as Democrats are significantly more supportive of Obama’s handling of all policy areas than are Republicans. We also find that Independents (who report no partisan leaning) are significantly more supportive than Republicans are of Obama’s handling of healthcare, international trade, and climate change.Footnote 24 Likewise, liberal respondents are significantly more supportive than moderates and conservatives are of Obama’s handling of every policy area except for climate change. These strong, consistent findings with respect to partisanship and ideology are unsurprising given the wide partisan presidential approval gaps that have emerged in recent years (Donovan et al. Reference Donovan, Kellstedt, Key and Lebo2020; Lebo and Cassino Reference Lebo and Cassino2007).
Finally, religion variables perform significantly better than most demographic measures do in all models; our demographic control variables are rarely significant. This result provides additional support for Converse’s claim: “there is fair reason to believe that [religious differences] are fully important, if not more important, in shaping mass political behavior than class differences” (1964: 248). Male respondents are significantly more supportive of Obama’s handling of the economy and terrorism but are statistically indistinguishable from female respondents concerning the other five policy areas. We find no effect of age or education on support for Obama’s handling of any of the seven policy areas. We also find that income is significant and positive in our economy, overall foreign policy, and terrorism models; higher-income respondents have higher levels of approval of Obama’s handling of these issues.
We generate predicted probabilities using CLARIFY (Tomz et al. Reference Tomz, Wittenberg and King2001) and plot the means and 95 percent confidence intervals for each of our nine religion variables in separate figures for each dependent variable.Footnote 25 Due to space considerations, we include only one domestic and one foreign policy figure (see Figures 1 and 2). For each figure, we set party identification as Democratic and all controls at their mean or modal values. The overall level of predicted support for Obama is buoyed by our choice to set party identification as Democratic, but the patterns and differences would be the same (albeit at lower levels) if we set it as Independent or Republican.
Figure 1. Predicted probability of Obama support on healthcare.
Figure 2. Predicted probability of Obama support on immigration.
Our results demonstrate that in addition to finding statistically significant effects of religion—especially “belonging”—on policy-focused presidential approval as shown above, we also find that many of these effects are substantively large for most of our dependent variables. Figures 1 and 2, and again in the figures presented in the Appendix, show the differences in expected support for Obama’s handling of policy among various religious groups. Black Protestants have a roughly 20 to 33 percent higher expected probability of reporting support for Obama on healthcare, the economy, overall foreign policy, immigration, and climate change than evangelical Protestants (the effects on terrorism and trade are not statistically significant and substantively smaller).
Our results also demonstrate consistent differences in support between non-Hispanic Catholics and Black Protestants. These differences are significant in assessments of Obama’s performance on healthcare, the economy, foreign policy, and immigration. These differences range from an expected 13 percent difference in support for Obama’s handling of healthcare to a roughly 21 percent difference in expected support for Obama on foreign policy. As we note above, there is only one area in which another religious group is significantly more approving of Obama’s policy performance than Black Protestants: Hispanic Catholics on trade. That effect is substantively large—the expected probability of support among Hispanic Catholics is roughly 22 percent higher than support among Black Protestants.
The discussion above focuses on differences in support for Obama’s policy performance between the Black Protestants and each other religious group. In analyses not reported, we have run models that exclude each of the other religious groups. Our large number of dependent variables and religious groups makes it impractical to present all those models here. We present two parallel sets of models for Table 1 in our online Appendix, where the excluded groups are (a) evangelical Protestants (Table A4) and (b) religiously unaffiliated (Table A5). Those results also lend support for our hypotheses overall.
Conclusion
Does religious affiliation play a role in shaping the public’s evaluations of the president’s policy performance? The goal of this study has been to answer this question by analyzing support for President Barack Obama’s handling of a range of domestic and foreign policy areas. Using survey data from the Pew Research Center and the Council on Foreign Relations, we have investigated the roles religious variables—“belonging” (religious affiliation) and “behaving” (frequency of worship attendance)—play alongside other factors that drive support for Obama’s handling of seven policy areas. Our findings demonstrate that religious affiliation has a powerful and independent impact on the public’s evaluation of President Obama’s policy performance. Compared to Black Protestants, evangelical Protestants, non-Hispanic Catholics, and (to a lesser extent) mainline Protestants and seculars espouse significantly less-approving attitudes regarding Obama’s policy performance. Across seven models, we find only one instance where a religious group—Hispanic Catholics—is more approving than Black Protestants are of Obama’s management of policy (international trade). In addition to the effects of religious affiliation, frequency of worship attendance predicts greater approval of Obama’s policy performance in four of our models.
The effects of religion on support for Obama’s handling of policy are substantively large, lending credence to the argument that people rely on religious identities as heuristics in evaluating politics (Madrid et al. Reference Madrid, Merolla, Ruiz and Schroedel2022; Westfall and Russell Reference Westfall and Çelik Russell2019). Our findings also support the argument that the political significance of religious affiliation intersects powerfully with race and ethnicity (Wilde Reference Wilde2018; Wilde and Glassman Reference Wilde and Glassman2016). Evangelical Protestants and non-Hispanic Catholics stand out for their distaste for Obama’s handling of public policy in comparison with non-White-Christian religious groups. Taken together, our findings underline the central role of social identities—religious, racial/ethnic, partisan, and ideological—in the formation of political attitudes and in the reinforcement of religion’s importance to public opinion and US politics in general.
The most striking results in Table 1 (and Figure 1) concern the differences between Black Protestants and evangelicals, as the latter group is consistently opposed to Obama’s handling of most policy areas, whether domestic or international. Few evangelicals voted for Obama (Pew Research Center 2012), and to some degree they hold negative attitudes toward racial minorities (Davis and Perry Reference Davis and Perry2021; McDaniel et al. Reference McDaniel, Nooruddin and Shortle2022; Perry et al. Reference Perry, Whitehead and Davis2019) and Obama as an individual (Benson et al. Reference Benson, Geer and Merolla2012; Guth Reference Guth and Steven2011; Tope et al. Reference Tope, Rawlinson, Pickett, Burdette and Ellison2017; Williams Reference Williams2013). Non-Hispanic Catholics also display a pattern of disapproving of Obama’s handling of policy, including in areas that the Catholic Church highlights as priorities for progressive action: healthcare and immigration. In the aggregate, non-Hispanic Catholics voted against Obama both times he stood for election (Pew Research Center 2012), reflecting their movement away from the Democratic Party (Gayte et al. Reference Gayte, Chelini-Pont and Rozell2018). Another Majority-White group of American Christians, mainline Protestants, disapprove only of Obama’s handling of domestic policy. This result reflects previous research establishing mainline progressivism on foreign policy issues (Taydas and Olson Reference Taydas and Olson2022).
As important as these findings are for understanding the relationship between public opinion and religion, further research on the topic is needed. First, while we explore multiple dependent variables, we examine the effect of religion on attitudes about the president’s handling of public policy at a single point in time. Expanding this study to analyze data from other years and about other presidents to see whether our results hold in different contexts would be a fruitful avenue for future research. For example, examining the role religion plays in shaping people’s views about President Joe Biden’s handling of domestic and international policies would provide additional insights, as he served as Obama’s vice president for eight years.
The results of this study also indicate that the impact of religious affiliation on public opinion (at least in the context of policy-specific presidential support) is distinct from the impact of party affiliation. However, religion’s influence on public attitudes might be mediated by partisan orientations. We hope future studies will explore the possibility of this indirect effect. Moreover, our analysis is limited to the questions Pew and the Council on Foreign Relations chose to ask. Incorporating assessments of presidential performance in other policy areas such as education, taxation, and criminal justice would enrich our understanding. Finally, we could not test the impact of the strictness of religious beliefs due to data limitations. Those who design survey instruments should be aware of the importance of nuanced religious measures and incorporate them in survey questionnaires in the future.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/S1755048325100187
Financial support
None.
Competing interests
None.
Open access
