Post-Seculars - American Sociological Association

ASRXXX10.1177/0003122414558919American Sociological ReviewO’Brien and Noy
Traditional, Modern, and
Post-Secular Perspectives on
Science and Religion in the
United States
American Sociological Review
2015, Vol. 80(1) 92­–115
© American Sociological
Association 2014
DOI: 10.1177/0003122414558919
Timothy L. O’Briena and Shiri Noyb
Using General Social Survey data, we examine perspectives on science and religion in the
United States. Latent class analysis reveals three groups based on knowledge and attitudes
about science, religiosity, and preferences for certain religious interpretations of the world.
The traditional perspective (43 percent) is marked by a preference for religion compared to
science; the modern perspective (36 percent) holds the opposite view. A third perspective,
which we call post-secular (21 percent), views both science and religion favorably. However,
when faced with competing accounts of events such as creation and evolution, post-seculars
root their views in religion rather than in mainstream science. Regression models indicate that
perspectives on science and religion do not simply mirror other denominational or ideological
differences. Furthermore, religio-scientific perspectives shape attitudes about political
issues where scientific and some religious communities diverge, including on abortion
rights and stem cell research. Overall, most individuals favor either scientific or religious
ways of understanding, but many scientifically inclined individuals prefer certain religious
accounts. This suggests that public divisions related to science and religion are cultural and
epistemological. This article underscores the complexity of the boundary between reason and
faith and highlights the roots of political conflict in perspectives on science and religion in
the United States.
science, religion, political culture, secularization, cultural conflict
The relationship between science, religion,
and society has been the focus of sociological
inquiry since the discipline’s earliest years.
Historically, many sociologists have presumed that as societies develop, science and
reason replace religion and faith as bases for
understanding the world. For example, Comte
(1998), Marx (1978), Weber ([1904] 1930),
and Durkheim ([1912] 1995) each forecast
that in modern society, positivism, rationality,
and science would displace the theological,
enchanted, religiously oriented perspective
held by earlier civilizations. However, more
University of Evansville
University of Wyoming
Corresponding Author:
Timothy L. O’Brien, Department of Law, Politics,
and Society, University of Evansville, 1800
Lincoln Avenue, Evansville, IN 47722
E-mail: [email protected]
O’Brien and Noy
than a century later, the prediction that a modern, scientifically oriented worldview would
dislocate a traditional, religiously inclined
one has not borne out. Continued high-profile
confrontations between leaders of religious
and scientific communities over issues such
as embryonic stem cells, genetic modification, and family planning illustrate that science has not simply replaced religion in the
modern world (Evans 2010; Nelkin 2004).
Popular accounts typically downplay doctrinal and institutional differences among religious traditions and presume a broadly
antagonistic relationship between science and
religion (Dawkins 2006). Contemporary
social scientific research often shares this
assumption of conflict (Barbour 2000; Evans
and Evans 2008). As Western society is
increasingly characterized by reason and science, religious authority has ceded influence.
Some secularization theories explain this shift
as the result of deep-seated incompatibility
between scientific and religious authority
(Chaves 1994; Gorski 2000). Although secularization is apparent among many Western
public institutions, it is less clear whether
individuals see science and religion as providing contradictory views of the world (Ecklund 2010; Taylor 2007).
Research on science, religion, and the public is extensive, but it contains two important
gaps. First, many studies treat attitudes about
science as either outcomes or antecedents of
religious attitudes and behaviors (Evans 2011;
Freeman and Houston 2011; Sherkat 2011).
Yet conceptualizing the relationship in this
way assumes that views of science and religion are causally related. In light of the strong
appreciation of science and continued religiosity in the United States (Gauchat 2012;
Stark 1999), we believe that such an assumption warrants further investigation.
Second, previous studies often examine
science and religion in uniform terms that do
not reflect their conceptual richness. Research
into the public understanding of science finds
that individuals view science as interrelated
dimensions of affect, knowledge, and understanding (Allum et al. 2008; Miller 2004).
Uni-dimensional analyses of science attitudes
or knowledge may therefore miss important
details about how individuals use science to
make sense of the world. Furthermore,
research on science and religion is typically
Christian-centric, measuring religion with
denominational and religious attendancebased items that may not capture differences
in how individuals use religion to orient their
lives (Bender et al. 2013).
We pursue a theoretically guided investigation that incorporates multiple dimensions
of individuals’ views of science and religion.
We refer to science and religion as distinctive
sets of ideas, information, beliefs, and practices that provide explanations and understandings of certain phenomena. For science,
these explanations and understandings are
based on reason and logic; for religion, they
are based on faith and tradition. In this article,
we examine individuals’ perspectives on science and religion and how they serve as interpretative frameworks that guide individuals’
understandings of the world. Perspectives on
science and religion are multifaceted, simultaneously consisting of and serving as sources
of knowledge and values. They are affective
and may be favorable, unfavorable, or ambivalent. By examining perspectives on science
and religion in tandem, we provide a unique
analysis that avoids conceptualizing views
about these two ways of understanding as
cause and consequence of one another. Findings suggest that the issue is more complicated than is often acknowledged and that
perspectives on science and religion mark
epistemological and cultural divides.
We use General Social Survey data to
examine whether a scientifically inclined
worldview precludes a religiously inclined
one, or if favorable perspectives on science
and religion can coexist in the minds of the
U.S. public. Latent class analysis suggests
that most people hold either modern or traditional perspectives, preferring either scientific or religious ways of understanding the
world. We also identify a third, post-secular
perspective held by roughly one in five individuals that sees conflict between science and
religion as limited to a narrow but important
set of issues. This third perspective is not
middle ground between a preference for science and one for religion, but a distinctive
worldview that reconciles science and religion in all but a few ways. We also investigate
how religio-scientific perspectives correspond to sociodemographic characteristics
and attitudes about political issues where
scientific and certain religious interpretations
diverge. Results indicate that perspectives on
science and religion do not necessarily replicate
denominational, ideological, or sociodemographic differences.
Theoretical Perspectives
Conflict between Science and
Contemporary accounts of conflict between
science and religion are often rooted in secularization theories, which hold that as societies modernize, individuals and public
institutions increasingly rely on rationality
rather than faith to organize daily life. Secularization occurs when religious authority is
differentiated from other sources of authority
and is no longer needed to legitimize institutions such as family, government, and education. In modern society, reason and science
are used to legitimize social arrangements
once considered the domain of religious
authority. Although different secularization
theories disagree on why the cultural roles of
science and religion have shifted (Gorski
2000), many scholars suggest that science and
religion provide contradictory frameworks, in
which an inclination for one prohibits an
inclination for the other (Binder 2002; Lienesch 2007; Smith 2003; Toumey 1994).
Observers agree that secularization is evident in many Western institutions (Drori et al.
2003; Evans and Evans 2008). Additionally,
waning religious participation in Europe and
declining public belief in the supernatural are
sometimes seen as indicative of religion’s
declining social influence (Bruce 2011;
Dobbelaere 1999; Wilson 1975). Together
with declining religiosity, strong support for
science among Europeans is consistent with
conflict-based accounts (National Science
American Sociological Review 80(1)
Board 2012). Furthermore, when asked
directly, many people in the United States
view science and religion as broadly conflicting (Baker 2012; Longest and Smith 2011),
suggesting that for these individuals an inclination toward scientific authority is incompatible with one for religious authority.
Overall, prior research suggests that individuals who are more knowledgeable about and
appreciative of science are more skeptical of
religious interpretations of the world. Alternatively, individuals who favor certain religious ways of knowing may be less
knowledgeable and appreciative of science.
A growing body of scholarship indicates
that perceived conflict between science and
religion may be limited to a few specific
issues. Religious individuals are less likely to
accept mainstream scientific theories of evolution and the big bang but are not necessarily
less knowledgeable about uncontested science (Evans 2011; Lawson and Worsnop
1992; Verhey 2005). This suggests that some
people see conflict between science and religion as localized and rely on both science and
religion to the extent that they do not provide
contradictory explanations. Religious people
in the United States are not necessarily averse
to science, but they may choose to interpret
some events in a religious light. Furthermore,
preferences for certain religious rather than
scientific accounts may reflect differences in
beliefs about the boundaries of legitimate science rather than an assessment of the value of
scientific knowledge (Plantinga 2011). If so,
an individual’s decision to diverge from
mainstream science on certain issues may not
reflect perceived conflict between reason and
Compatibility between Science
and Religion
Science and religion may be more compatible
than the conflict thesis suggests. Merton
([1938] 1973) identified the importance of
Puritanism in the scientific revolution, and
scholars have since elaborated the role of
religious values in early scientific discovery
and the professionalization of science (Shapin
O’Brien and Noy
1996). Recent studies of religious scientists
illustrate that scientific careers do not preclude favorable views of religion (Collins
2006; Ecklund 2010). However, these studies
emphasize non-institutionalized religious
beliefs and practices such as spirituality, raising questions about how scientifically inclined
individuals define religion (Ecklund and
Long 2011).
Despite declining trust in science among
political conservatives, the U.S. public has
expressed favorable attitudes about science
for at least the past four decades (Gauchat
2012). In contrast with Europe, religious participation in the United States was relatively
stable over this period, a trend at odds with
secularization theories (Inglehart and Baker
2000). Taken together, public support for science and continued religiosity among the
U.S. public suggests that many individuals do
not perceive the conflict between science and
religion presumed by many scholarly and
popular accounts.
Theories of a post-secular society indicate
that despite institutional differentiation, individuals often view the world through religious lenses. Although many aspects of
modernity are marked by secular authority,
some argue that religion’s influence on society has been transformed rather than replaced
by science (Casanova 2010; Habermas 2008;
Taylor 2007). While civic, political, and economic institutions stake their authority to
secular ground, many individuals continue to
use religious framings to make decisions
about daily life (Gorski et al. 2012). In this
view, although religion has ceded formal
authority, it continues to influence society
through individuals’ choices about what to
believe and how to behave.
In this account, individuals purposefully
construct worldviews by combining elements
of scientific, religious, and other ways of
knowing (Casanova 2010). Hence, a postsecular perspective entails recognition of the
value and utility of multiple belief systems. A
post-secular worldview rejects the strict
adherence to science characteristic of modernity. Instead, it blends scientific, religious,
and other authorities to provide a personally
compelling narrative of the world. In essence,
a post-secular perspective views any singular
interpretative framework, such as science or
religion, as only a partial explanation of reality. Thus, the promise of post-secular theories
is not to anticipate uniform preferences for
science or religion, but to clarify the circumstances under which individuals prefer different kinds of explanations. In summary,
existing research leads us to expect that some
individuals have generally favorable perspectives on both science and religion.
The Waning Importance of Science
and Religion
Alternatively, individuals may reject both science and religion as sources of understanding.
The increasing complexity and bureaucracy
of modern institutions isolate and alienate
individuals from many social, economic, and
political processes (Giddens 1991; Habermas
1989). Scientific and religious authority may
each suffer as a result. A postmodern view of
society entails a strong relativist epistemology, in which truth claims are evaluated individually, subjectively, and without consistent
reference to broader interpretative frameworks. While post-secularism contends that
individuals blend epistemological positions to
orient their lives, the postmodern view implies
that social institutions are incapable of providing enduring meaning or truth (Triandis
1995). Thus, some individuals may dismiss
both science and religion as ways to ground
their understanding of the world.
Overall, perspectives on science and religion may be characterized by one of four
patterns. Table 1 outlines these ideal types.
According to conflict theories, we should
identify two worldviews, one oriented toward
science and away from religion and one with
the opposite preference. We refer to these
contrasting perspectives as modern and traditional.1 We may also observe a post-secular
perspective, marked by generally favorable
views of both science and religion. Finally,
some respondents may be skeptical of both
scientific- and religious-based understandings, a perspective we refer to as postmodern.
American Sociological Review 80(1)
Table 1. Ideal Types of Perspectives on Science and Religion
Although this schema necessarily distills the
U.S. public’s views, our framework advances
scholarship in this area by allowing perceptions of science and religion to vary independently of one another.
Social Bases of Perspectives on
Science and Religion
Although previous research has not focused
directly on religio-scientific perspectives,
scholars have linked attitudes and behaviors
related to science and religion to a variety of
sociodemographic characteristics. We therefore examine how religious traditions, gender,
race, class, and other traits correspond to
perspectives on science and religion. In general, we expect individuals who attend religious services regularly, regardless of their
faith traditions, to have especially favorable
views of religious interpretations of the world.
Furthermore, institutional and doctrinal differences among religious traditions suggest
that different faiths promote systematically
different perspectives on science and religion.
Some traditions accept scientifically favored
theories of life, whereas others reject certain
scientific accounts. Further complicating
matters is the potential decoupling of institutional and individual views. For example,
while the Catholic Church has formally reconciled with scientific theories of the big
bang (Pope Pius XII 1951; Ratzinger 1988),
some Catholics may still favor young-earth
interpretations of creation. Nonetheless, we
expect that members of conservative Protestant traditions are more likely than others to
tie their worldviews to religious explanations
incompatible with mainstream science (Hoffman and Johnson 2005; Nelsen, Guth, and
Fraser 2001; Scheufele et al. 2009). We also
anticipate that individuals unaffiliated with a
faith tradition will be most likely to favor a
scientific worldview and reject a religiously
inclined one.
Perspectives on science and religion may
also vary by gender. Women are socialized
from a young age to be less scientifically oriented than men (Correll 2004; Xie and Shauman 2003), and women also tend to be more
religious, although the gender gap in religiosity disappears among scientists (Ecklund,
Park, and Veliz 2008; Miller and Hoffman
1995). This suggests that, compared to men,
women will have less favorable perspectives
on science and more favorable perspectives
on religion. Furthermore, African Americans
and Latinos have less favorable attitudes
about science and medicine compared to
whites (Pew Research Center 2009; Schnittker, Freese, and Powell 2000) and are, on
average, more religious (Hunt and Hunt
2001). We therefore anticipate racial and ethnic differences in religio-scientific perspectives, with non-whites viewing science less
favorably and religion more favorably compared to whites.
Social class may also shape perspectives
on science and religion. Education generally
corresponds to knowledge of and support for
science (Allum et al. 2008; Miller 2004).
Additionally, higher socioeconomic status is
linked to lower levels of religiosity (Schieman 2010; Smith and Faris 2005). Thus, we
expect social class, as measured by education,
income, and occupational status, to correspond to positive perspectives on science and
negative perspectives on religion.
Views on science and religion may relate
to other cleavages as well. Compared to
younger individuals, older people may be
more oriented toward religion and less toward
O’Brien and Noy
science (Argue, Johnson, and White 1999;
Bak 2001). Because the U.S. South is characterized by unique cultural features, including
a stronger historical identification with the
anti-evolution movement, we control for geographic location. Finally, we anticipate differences associated with political ideology, with
liberals holding more favorable views of science and conservatives holding more favorable views of religion (Gauchat 2012).
Science, Religion, and Political
If perspectives on science and religion are
distinct from denominational, ideological,
and sociodemographic differences, then they
may also have distinct effects on attitudes
about political conflicts. In particular, religioscientific perspectives may help shed light on
issues where science and religion have overlapping jurisdictions. For example, the issue
of abortion rights has long been within the
reach of both medical/scientific and religious
authorities (Luker 1984). More recently, topics such as embryonic stem cell research and
genetic modification have posed “ideological
dilemmas” for the public (Locke 1999). Science is often associated with progressive politics, whereas religious adherence is more
closely aligned with social conservatism
(Davis and Robinson 1999; Gauchat 2012).
We therefore expect that an inclination for
science is associated with support for liberal
public policies, whereas an inclination for
religion is linked to support for a conservative
agenda. We do not, however, expect perspectives on science and religion to correspond
directly to political attitudes about issues
where scientific and religious communities
do not provide direct counterpoints.
Data, Measures, and
To examine perspectives on science and religion, we analyze cross-sectional data from
the 2006, 2008, and 2010 waves of the
General Social Survey (GSS).2 The GSS is a
nationally representative biennial survey of
households in the United States that uses a
multi-stage area-probability sampling frame
(Smith et al. 2011). These waves of the GSS
included a special topics module, which
asked respondents about their knowledge of
and attitudes about science. Along with questions about religion asked in each survey
wave, these variables are the focus of our
GSS data also contain background information on respondents, which we use to
examine sociodemographic and attitudinal
correlates of perspectives on science and religion. In 2006, 2008, and 2010, the science
module was administered to 1,864, 1,526, and
691 respondents, respectively. Accounting for
the survey’s split ballot design and missing
cases, results presented here are based on
analyses of 2,901 cases (1,563 from 2006;
988 from 2008; and 350 from 2010).
Measuring Perspectives on Science
and Religion
Perspectives on science encompass attitudes
about science as well as knowledge of scientific concepts and methods. We examine attitudes about science using survey questions
commonly used to measure public appreciation of science (Miller 2004). These items
asked whether (1) science creates more
opportunities for the next generation, (2) science makes life move too fast (reverse coded),
(3) science should be supported by government funding, and (4) the benefits of science
outweigh its costs. We model these items as
discrete ordinal variables on four- and fivepoint scales, where higher scores indicate
more favorable attitudes. Table 2 contains
unadjusted descriptive information for the
variables we use to measure perspectives on
science and religion.
We examine knowledge of science using a
series of 14 quiz-style questions modeled as
binary variables. These questions focus on
uncontroversial aspects of scientific knowledge, including radioactivity, subatomic particles, and experimental design. This series
Source: 2006, 2008, and 2010 GSS.
Significantly different from traditional, p < .05 (two-tailed t-test).
Significantly different from post-secular, p < .05 (two-tailed t-test).
Significantly different from modern, p < .05 (two-tailed t-test).
Class Size
Science Knowledge (scientifically correct answer equals one, else equals zero)
Center of the Earth is very hot?
All radioactivity is man-made?
The father’s gene decides whether the baby is boy or girl?
Lasers work by focusing sound waves?
Electrons are smaller than atoms?
Antibiotics kill viruses as well as bacteria?
Does the Sun go around the Earth or the Earth around the Sun?
The continents have been moving for millions of years and will move in the future?
The Universe began with huge explosion?
Human beings developed from earlier species of animals?
Does a one in four chance of inherited illness mean that if the first child has the illness,
the next three will not?
Does a one in four chance of inherited illness mean that each child has the same risk of
having the illness?
Understand experimental research design?
Clear understanding of what it means to study something scientifically?
Science Attitudes
Science and technology create more opportunities for the next generation (1 = strongly
disagree, 4 = strongly agree).
Science makes our way of life change too fast (1 = strongly agree, 4 = strongly disagree).
Scientific research that advances the frontiers of knowledge is necessary and should be
supported by the federal government (1 = strongly disagree, 4 = strongly agree).
Do the benefits of scientific research outweigh the harmful results (0 = harm strongly
outweighs benefits, 2 = harm and benefits about equal; 4 = benefits strongly outweigh
Religion Indicators
Bible is the actual word of God (0 = no, 1 = yes)?
Bible is inspired by the word of God (0 = no, 1 = yes)?
Bible is a book of myths and fables (0 = no, 1 = yes)?
Strength of religious affiliation (1 = none, 4 = very strong).
Overall Sample
(n = 2,901)
Table 2. Description of Variables Used to Measure Perspectives on Science and Religion
(n = 1,041)
(n = 1,238)
(n = 622)
Conditional Means by Latent Class
O’Brien and Noy
also includes questions about the big bang
and evolution, which have been criticized for
confounding knowledge of science with preferences for certain religious accounts (Roos
2012). However, these survey questions are
central to our analysis because they capture
the cultural distinctions we seek to investigate. Additionally, analyzing questions about
the big bang and evolution in combination
with questions about less controversial topics
like probability provides unique insight into
the U.S. public’s preferred sources of knowledge and values (Toumey et al. 2010).
To measure perspectives on religion, we
analyze responses to a question that asked
whether the Bible is (1) the actual word of
God, (2) inspired by the word of God, or (3)
filled with myths and fables. This item, which
we model as a nominal variable, is frequently
used to measure views of certain religious
interpretations of the world (Davis and Robinson 1999). We also analyze individuals’
religiosity as an ordinal variable based on a
question that asked respondents to rate the
strength of their religious beliefs on a fourpoint scale, where higher scores correspond
to stronger belief. Although the GSS contains
other behavioral and attitudinal measures of
religion, we focus on these items because
they begin to tap individual preferences for
religion and religious-based accounts of the
world despite institutional and intellectual
differences among faiths.3
Sociodemographic Characteristics
We examine how several respondent characteristics relate to perspectives on science and
religion. Table 3 summarizes these variables.
We measure religious traditions using the
typology developed by Steensland and colleagues (2000). We analyze the variable as a
set of binaries for conservative, mainline, and
Black Protestants; Catholics; Jews; followers
of other faiths; and individuals not associated
with organized religion.4
We examine religious attendance using an
eight-category ordinal variable ranging from
“never attends” to “attends more than once
per week.”5 We measure race and ethnicity
using binary variables for Latino, non-Latino
African American, non-Latino white, and
non-Latino other race identification. We
measure age in years, divided by 10. We
examine geographic location using a binary
measure for residents of the South. We measure political ideology using an ordinal sevenpoint scale, where 1 refers to extremely liberal
and 7 to extremely conservative. We measure
education in years and income as a natural log
transformation of household income category
midpoints.6 Finally, we measure occupations
using a five-class specification of the EriksonGoldthorpe-Portocarero scheme created
according to the procedure described by
Alderson, Junisbai, and Heacock (2007) (I +
II = service class; IIIa + b = routine nonmanual; IVa + b + c = petty bourgeoisie/
farmer; V + VI = skilled workers and foremen; and VIIa + b = non-skilled worker).7
Political Attitudes
To begin to assess how perspectives on science and religion relate to U.S. political culture, we analyze attitudes about several issues
where scientific and religious communities
each have claims to authority. We examine
respondents’ opinions about women’s right to
choose abortion with a binary variable. We
measure support for government funding for
stem cell research with a four-point ordinal
variable. We analyze a three-point ordinal
variable that asked the degree to which respondents are comfortable consuming genetically
modified food. To test whether perspectives
on science and religion also predict attitudes
about issues where science and religion may
not directly compete, we examine a five-point
ordinal variable that asked about support for
requiring greater fuel economy from automakers and a four-point ordinal variable about
support for nuclear energy production. Higher
scores correspond to greater support on each
of these attitudinal measures (see Table 3).
Analytic Technique
To identify underlying perspectives on science and religion, we use latent class analysis
Religious Tradition
Mainline Protestant
Conservative Protestant
Black Protestant
Other faith
No religious affiliation
Religious Attendance (0 = never, 8 = more than once per week)
African American (non-Latino)
Other race (non-Latino)
White (non-Latino)
Education (in years)
Income (natural log transformation of household income category midpoints)
Erikson-Goldthorpe-Portocarero Class Scheme
I + II = service class
IIIa + b = routine non-manual
IVa + b + c = petty bourgeoisie/farmer
V + VI = skilled workers and foremen
VIIa + b = non-skilled workers
Political Views (1 = extremely liberal, 7 = extremely conservative)
Lives in South
Age (in years, divided by 10)
Table 3. Description of Independent Variables and Political Attitudes
Overall Sample
Conditional Means by Latent Class
Overall Sample
Source: 2006, 2008, and 2010 GSS.
Note: Latent class sizes vary for political attitudes due to differences in sample sizes for these items.
Significantly different from traditional, p < .05 (two-tailed t-test).
Significantly different from post-secular, p < .05 (two-tailed t-test).
Significantly different from modern, p < .05 (two-tailed t-test).
Political Attitudes
Should it be possible for a pregnant woman to obtain a legal abortion for any
reason she wants? (0 = no, 1 = yes)
Should U.S. government fund scientific research using embryonic stem cells?
(1 = definitely should not, 4 = definitely should)
Will you eat genetically modified (GM) food? (1 = will not eat GM food,
3 = will eat GM food)
Favor or oppose requiring automakers to make cars and trucks that use less
gasoline? (1 = strongly oppose, 5 = strongly favor)
Favor or oppose increasing electricity in U.S. produced from nuclear power?
(1 = strongly oppose, 4 = strongly favor)
Table 3. (continued)
Conditional Means by Latent Class
American Sociological Review 80(1)
Table 4. Fit Statistics for Latent Class Analysis
Number of
Percent Reduction
in BIC
Source: 2006, 2008, and 2010 GSS; n = 2,901.
Note: p is p-value from Lo-Mendell-Rubin likelihood-ratio test; BIC is Bayesian information criterion;
LL is log likelihood; and df is degrees of freedom; bolded text indicates preferred model.
(LCA) (Goodman 1974; Lazarsfeld and
Henry 1968; Magidson and Vermunt 2001).
LCA assumes that responses to conceptually
similar questions share underlying associations. LCA identifies the number of latent
classes (T ) needed to best account for
response patterns in the observed manifest
variables, which in this article are attitudes,
affect, and knowledge about science and religion. Conventionally, LCA entails fitting an
independence model in which each case is
assigned to a single latent class (T = 1).
Assuming that an underlying association
exists among manifest variables, model fit
will improve when T = 2,…, n, until the
underlying associations among variables are
identified. Respondents are assigned to latent
classes based on their greatest posterior probability of class membership.8 For example, in
a three-class model where the probabilities of
class membership are .7, .2, and .1, the individual is assigned to the first class.
The second part of the investigation uses
multinomial logistic regression to examine
sociodemographic characteristics of the
groups identified by the latent class analysis.
Our sample in the second stage of the analysis
consists of the 2,331 individuals with complete information for sociodemographic variables of interest.9 Finally, we use binary and
ordinal logistic regression models to investigate how perspectives on science and religion
relate to political attitudes. Sample sizes for
this final set of analyses vary according to the
GSS ballot design. We conducted the LCA
using Mplus and descriptive and regression
analyses using Stata.10
Perspectives on Science and Religion
Table 4 presents LCA results. We examined
models that grouped respondents into between
one and seven latent classes. To determine the
best fitting model, we relied on Lo-MendellRubin likelihood-ratio tests (LMR) and the
Bayesian Information Criterion (BIC), common measures of model fit in LCA (Asparouhov and Muthén 2012; Nylund,
Asparouhov, and Muthén 2007). A significant
LMR test indicates that the model provides a
better fit than a model with one less latent
class. An insignificant test statistic indicates
no improvement in fit. In Table 4, the insignificant LMR test from the four-class model
suggests the three-class model best fits the
data. However, the BIC’s minimum value is
sometimes used to select the number of
classes in LCA, and in Table 4 the BIC is lowest in the six-class model.
When fit statistics point to different latent
class solutions, scholars suggest considering
the kinds of variables being analyzed (Nylund
et al. 2007). While the BIC may be preferable
for LCA models with continuous outcomes,
the LMR test more reliably selects the correct
number of classes for models with categorical
outcomes, like those in this analysis (Lo,
Mendell, and Rubin 2001). Furthermore, the
O’Brien and Noy
BIC’s reliability has been criticized when
there are a small number of classes, especially
when class sizes are unequal (Nylund et al.
2007). Because there is no consensus about
which statistic should be used to select the
number of latent classes, the decision typically rests on a combination of statistical,
substantive, and theoretical considerations.
Based on LCA results, analyses of different
latent class models, and theoretical expectations, we focus on results from the three-class
The final three columns in Table 2 contain
conditional means and proportions of manifest variables. These statistics informed our
choices of category labels and give meaning
to the groups identified in the LCA. Superscripts designate statistically significant differences between groups as indicated by
two-tailed t-tests. The top row of Table 2
indicates that the largest category is characterized by a traditional perspective on science
and religion and contains 43 percent of the
sample. A second class is marked by a modern view and contains 36 percent of the sample. Finally, 21 percent of respondents hold a
post-secular perspective on science and
Overall, these results suggest that most
people in the United States have favorable
orientations toward either science or religion
but not both. This is consistent with theories
of broad conflict between these two ways of
knowing. Furthermore, although the postsecular perspective entails high levels of science knowledge as well as favorable views of
science and religion, responses to questions
about evolution and the big bang suggest that
even for this most accommodating group, science and religion sometimes conflict. When
asked about these issues, the post-secular
latent class almost unanimously aligned their
views with particular religious accounts. So,
rather than the four ideal types outlined earlier, Table 2 identifies three unique religioscientific perspectives, each supporting a
variation of the conflict thesis.
Members of the traditional class have significantly lower scores than do the other
groups for all science literacy and attitude
items. For example, 47 percent of the traditional class, compared to 92 and 90 percent of
the modern and post-secular groups, respectively, correctly answered that radioactivity
occurs naturally. Although the traditional
class reported lower levels of religiosity than
did the post-secular class, its religious affiliation strength is higher than average and significantly higher than that of the modern
class. Likewise, 46 percent of the traditional
group, compared to 31 percent of the overall
sample, responded that the Bible is the literal
word of God.
The modern perspective, shared by 36 percent of the sample, stands in stark opposition
to a traditional view of science and religion.
Members of the modern category are knowledgeable about science and the most appreciative of its uses, and they are the least
religious. This perspective is not unique in its
high science literacy, but it is distinctive in its
optimism about science in society and its low
levels of religiosity. Unlike other perspectives, large majorities of moderns responded
that the universe began with a big bang and
that humans evolved from other animals ( p <
.05). Moreover, only 3 percent of the modern
group responded that the Bible is the literal
word of God. Among moderns, 41 percent
indicated that the Bible is a book of myths
and fables, compared to 19 percent of the
overall sample.
The post-secular perspective, held by 21
percent of respondents, is characterized by
less favorable attitudes about science than
those of the modern class, but their attitudes
toward science are significantly more favorable than those of the traditional class. Moreover,
science literacy scores for the post-secular
and modern classes are statistically indistinguishable on nearly half of these items. In
contrast to moderns, however, 48 percent of
post-seculars reported that the Bible is the
literal word of God, and none reported that
the Bible is a book of myths and fables. Furthermore, post-seculars’ mean religious affiliation strength is significantly higher than that
of each other latent class ( p < .05).
Despite their generally favorable outlook
on science, members of the post-secular
category were substantially and significantly
less likely than the other classes to respond
that the universe began with a big bang (6
percent) and that humans evolved from other
animals (3 percent). Post-seculars were also
less likely than moderns to respond that the
continents have been moving for millions of
years ( p < .05). These findings suggest the
post-secular perspective recognizes limited
conflict between science and religion. Analyses of additional science knowledge questions
asked to a subset of respondents in 2008 further suggest that members of the post-secular
latent class are knowledgeable about science
yet choose to root their understandings of
certain events in religious belief.11
In summary, our LCA identifies three distinct perspectives on science and religion.12
Findings suggest that most people in the
United States are inclined toward either science or religion, but not both. We also find a
narrow swath of the public with generally
favorable perspectives on both scientific and
religious understandings. But rather than a
fully compatible view, even these individuals
cannot reconcile some scientific and religious
accounts. This third perspective is consistent
with recent findings that many religious individuals are scientifically literate yet prefer
some religious explanations to scientific ones
(Evans 2011). Our results suggest that approximately one-in-five members of the U.S. public see science and religion in this light.
Sociodemographic Differences in
Perspectives on Science and Religion
The final columns of Table 3 present
sociodemographic characteristics of each
latent class. As expected, women, African
Americans, Latinos, and individuals in lower
social classes are significantly overrepresented in the traditional compared to the
modern category. The post-secular and modern classes have similar income, despite postseculars’ slightly lower educational attainment
and occupational status. Furthermore, the
post-secular perspective is held disproportionately by respondents who are older, politically conservative, and reside in the South.
American Sociological Review 80(1)
Table 3 also underscores the importance of
religious traditions for individuals’ perspectives on science and religion. Conservative
Protestants are split nearly evenly between the
post-secular and traditional latent classes.
While conservative Protestants in the postsecular and traditional classes differ significantly in their gender, race, class, age, and
political views, these gaps tend to reflect more
general differences between perspectives.13
To examine these patterns in a multivariate
setting, we estimated a multinomial logistic
regression model. Table 5 indicates that many of
the group differences identified in Table 3
remain statistically significant net of one
another. Compared to mainline Protestants, conservative Protestants are 3.618 ( p < .001) and
1.663 ( p < .01) times more likely to hold the
post-secular rather than modern or traditional
perspectives. Religiously unaffiliated respondents are less likely to hold post-secular rather
than modern (.144; p < .001) or traditional
views (.359; p < .01).14 Also, members of the
post-secular class attend religious services more
frequently than do members of the other latent
classes, net of other differences ( p < .001).
Gender and race differences persist in the
multivariate context. Women are more than
twice as likely as men to hold post-secular
rather than modern perspectives (2.037, p <
.001). Whites are more likely to express postsecular (4.706; p < .001) or modern (8.736;
p < .001) rather than traditional views.15 Furthermore, Table 5 indicates that the postsecular and modern classes have similar
income and occupational status despite a statistically significant education gap.16 Finally,
Table 5 indicates that respondents who hold a
post-secular perspective are significantly
more politically conservative than the other
respondents, other differences aside.
Traditional, Modern, and PostSecular Perspectives and Political
Finally, we examine how religio-scientific
perspectives relate to beliefs about issues
where science and religion have each been
mobilized by political interests. Table 3
O’Brien and Noy
Table 5. Odds Ratios for Multinomial Logistic Regression of Science-Religion Perspectives
on Independent Variables
versus Modern
Religious Traditiona
Conservative Protestant
Black Protestant
Other faith
No religious affiliation
Religious Attendance
White (non-Latino)
Other race (non-Latino)
Erikson-Goldthorpe-Portocarero Class Schemec
I + II = service class
IIIa + b = routine non-manual
IVa + b + c = petty bourgeoisie/farmer
V + VI = skilled workers and foremen
Political Views
Lives in South
Age (years, divided by 10)
Log likelihood
Modern versus
versus Traditional
Source: 2006, 2008, and 2010 GSS; n = 2,331.
Note: Standard errors are in parentheses.
Wald X2 = 150.315; p < .001; referent is mainline Protestant.
Wald X2 = 109.070; p < .001; referent is non-Latino African American.
Wald X2 = 33.247; p < .001; referent is VIIa + b (non-skilled workers).
*p < .05; **p < .01; ***p < .001 (two-tailed test).
suggests that perspectives on science and
religion are related to attitudes about some
but not all political conflicts. For example,
the overall sample is closely divided on the
question of abortion, yet attitudes vary sharply
across latent classes. The post-secular class is
most strongly opposed to women’s right to
choose abortion, significantly more so than
the traditional class ( p < .05). Only the modern latent class reports majority support for
abortion rights ( p < .05). Attitudes about
stem cell research and genetically modified
food follow a similar pattern. For both issues,
post-seculars and traditionals express disproportionate support for positions associated
most closely with organized religion, whereas
moderns report greater support for positions
associated with mainstream science. Differences in attitudes about fuel economy standards and nuclear energy production are
largely statistically insignificant, suggesting
that religio-scientific perspectives are most
salient in conflicts over issues where scientific and religious institutions each claim
authority. Debates related to environmental
and energy policy are politically contentious,
but they do not engage the same scientific and
religious meanings as questions related to
life, such as when it begins and how it can be
manipulated by humans.
Attitudes about socio-political issues are
often attributed to denominational, ideological, and other social cleavages. However,
regression results presented in Table 6 indicate that perspectives on science and religion
are significantly associated with attitudes
about abortion rights, stem cell research, and
genetically modified food net of other differences among respondents. A modern perspective is associated with greater support for all
of these issues ( p < .05). However, while
moderns are more willing than post-seculars
to consume genetically modified food, the
difference is only marginally statistically significant ( p < .10). Additionally, the postsecular class is significantly more opposed
than the traditional class to abortion rights
and public funding for stem cell research
American Sociological Review 80(1)
( p < .05; results not shown). Overall, these
results further suggest that religio-scientific
perspectives are particularly salient for issues
where humans intervene in natural processes
of life, even in cases of non-human life as
illustrated by attitudes about genetically modified food. Religio-scientific perspectives
may therefore also be useful for understanding political conflict concerning human cloning and genomic research, assisted
reproductive technology, and end-of-life
issues including capital punishment.
Several control variables in Table 6 have
statistically significant effects in anticipated
directions. Other differences aside, political
conservatives and respondents who attend
religious services more frequently are more
likely to oppose abortion rights and government funding for stem cell research ( p <
.001), and women are more likely to support
abortion rights ( p < .05). Most important to
this article, however, is that after controlling
for a wide range of respondent characteristics,
significant differences remain in traditional,
modern, and post-secular views on political
debates where science and religion compete
for cultural and epistemic priority.
Our results suggest that the discursive
framing of political debates is crucial for
determining whether and how they become
embedded within broader perspectives on science and religion. Political contests over
reproductive rights, stem cells, and genetic
modification often center on direct conflict
between certain scientific and religious
authorities. Consequently, public attitudes are
entwined with wider cultural sentiments
related to reason and faith. Political debates
about fuel economy standards and nuclear
energy are not imbued with the same religious
meaning as controversies related to life, especially human life. Hence, attitudes about
energy policy are not strongly related to
perspectives on science and religion. In summary, these findings suggest that religioscientific perspectives are most useful for
understanding political controversies where
the jurisdictions of science and religion
Perspectives on Science and Religiona
Traditional perspective
Post-secular perspective
Wald X2 for joint significance of science-religion
Religious Traditionb
Conservative Protestant
Black Protestant
Other faith
No religious affiliation
Religious Attendance
Other race (non-Latino)
Support Funding
for Stem Cell
Support Right to
Choose Abortion
Support for GM
Support Fuel
Support Expanded
Economy Standards Nuclear Energy
Table 6. Odds Ratios for Logistic Regressions of Political Attitudes on Science-Religion Perspectives and Independent Variables
Support for GM
Support Funding
for Stem Cell
Support Right to
Choose Abortion
Support Fuel
Support Expanded
Economy Standards Nuclear Energy
Source: 2006, 2008, and 2010 GSS.
Note: Standard errors are in parentheses. Abortion rights model is binary logit, all others are ordinal logit (cut points not shown).
Referent is modern perspective.
Referent is mainline Protestant.
Referent is non-Latino African American.
Referent is VIIa + b (non-skilled workers).
*p < .05; **p < .01; ***p < .001 (two-tailed test).
White (non-Latino)
Erikson-Goldthorpe-Portocarero Class Schemed
I + II = service class
IIIa + b = routine non-manual
IVa + b + c = petty bourgeoisie/farmer
V + VI = skilled workers and foremen
Political Views
Lives in South
Age (years, divided by 10)
Model log likelihood
Model BIC
Table 6. (continued)
O’Brien and Noy
overlap, and where these sources of authority
have each been enlisted in the service of
political interests.
This article examined views of science and
religion in the United States and how these
perspectives correspond to sociodemographic
attributes and socio-political attitudes. The
traditional and modern perspectives illustrate
a fundamental fissure among the U.S. public’s cognitive framings of the world: one
favors science over religion, the other holds
the opposite view. The post-secular perspective indicates that roughly one-in-five individuals view science and religion more
circumspectly. However, rather than the fully
compatible worldview we anticipated, the
post-secular perspective sees conflict between
science and religion as limited to particular
issues related to life. In these domains, the
post-secular perspective is associated with a
tendency to use religion to ground one’s
Conservative Christians are not uniformly
opposed to science, but they are more likely
to reject scientific theories of the big bang
and human evolution (Evans 2011). However,
our analysis reveals that the selective rejection of certain scientific theories cannot be
accounted for by denominational differences
alone. Individuals from across religious traditions accept some facets of science and reject
others, suggesting that although faith traditions are related to religio-scientific perspectives, they are not constitutive of them. Thus,
we find that religion shapes individuals’
worldviews in ways not captured by conventional measures of religion such as denominational affiliation or religious attendance. This
finding bolsters recent criticisms of social scientific work on religion as overly congregationbased and Protestant-centric (Bender et al.
2013). Furthermore, it illustrates how views
of religion matter for social attitudes beyond
strictly religious contexts.
Post-seculars’ generally favorable views of
science and their rejection of evolution and
the big bang may indicate that science and
religion coexist for these individuals, but that
evolution and the big bang are not viewed as
legitimate science. By defining a specific
component of evolution—natural selection—
as outside the boundaries of science, other
aspects of the theory, such as an ancient earth
and common ancestry, can be reconciled with
religious texts (Plantinga 2011). The disproportionate number of conservative Protestants
in the post-secular class supports this interpretation. Historically, some conservative
Christian traditions have viewed mainstream
scientific theories of evolution and the big
bang as corruptions rather than limitations of
science (Iannaccone 1993; Marsden [1980]
2006, 1991). If these theories are perceived as
non-science, then individuals who dismiss
them may not see their views on science and
religion as conflicting in even a narrow sense.
Perspectives on science and religion are an
important component of U.S. political culture. Shared meanings of expertise, credibility, and authority are ingrained within
society’s collective decisions about nature,
knowledge, and politics (Jasanoff 2005).
Competing worldviews among publics stoke
fundamental disagreements about questions
regarding life, such as when it begins, how it
can end, and whether it should be manipulated by technology. We find that traditional,
modern, and post-secular perspectives on science and religion differentiate U.S. public
opinion on the big bang, evolution, abortion
rights, stem cell research, and genetically
modified food. The overlapping jurisdictions
of science and religion in these areas create
ideological dilemmas, which force individuals to choose among different interpretations
of the world (Locke 1999). Ideological dilemmas are not limited to matters of reason and
faith, but the absence of a fourth perspective,
one that rejects both science and religion,
underscores that science and religion are central pillars in U.S. life. Therefore, rather than
appealing to other sources of authority,
including common sense or morality, political
debates often stress direct opposition between
scientific and certain religious communities.
The social profiles of religio-scientific
perspectives further inform understandings of
political culture. In terms of education,
income, and occupations, individuals with a
traditional perspective are more socially marginalized than people in other classes, and
marginalization is often linked to conservative religious commitment (Stark and Bainbridge 1987). However, we show that the
post-secular perspective is the most religiously committed despite its association
with relatively high socioeconomic status.
Additionally, favorable views of science are
typically associated with progressive sociopolitical attitudes. However, the post-secular
perspective suggests this linkage is not inevitable. In other words, this analysis provides
evidence that social marginalization and religious commitment are not necessarily bound
together, nor are positive views of science and
liberal political preferences. Future work
should examine the implications of this finding in greater depth.
Modernity is often defined by rationalization (Weber [1904] 1930), information (Castells 1996), quantified risk (Beck 1992), and
other kinds of scientific knowledge. However, we find that more than three in every
five members of the U.S. public are associated with perspectives that diverge from
mainstream science on questions of creation
and evolution. Perhaps Western society has
never been marked by a widespread scientific
orientation (Latour 1993). In the twenty-first
century, a majority of the U.S. public believes
in miracles, just as they did throughout the
twentieth century (Shapin 2008). This suggests that religion continues to shape individuals’ worldviews, which in turn shape
political culture more generally. Nonetheless,
this article provides a more refined understanding of science, religion, and the public
than is often assumed. Although much of the
U.S. public prefers scientific to religious
ways of understanding, neither the postsecular nor the traditional perspectives can be
easily dismissed as anti-scientific.
Individuals’ daily choices—about who to
trust, what to believe, and how to behave—
are simultaneously constrained by and constitutive of society’s macro-level contours
(Giddens 1991). This article advances
American Sociological Review 80(1)
sociological theory by showing how the U.S.
public clusters into widely held interpretative
positions that correspond to an array of social
cleavages and provide a setting for deepseated political conflict. Although religioscientific perspectives may not correspond to
all political attitudes, additional corollaries of
perspectives on science and religion, including voting patterns, interpersonal behaviors,
and socioeconomic attitudes, are a promising
avenue for future research.
This article relies on a unique set of survey
questions fielded in select waves of the GSS
to subsamples of respondents, and many of
the items used in this analysis are not contained in existing datasets. Our findings highlight the importance of variables measuring
beliefs about the Bible, benefits of scientific
research, human evolution, and the big bang
for distinguishing perspectives on science and
religion. If future research can identify reliable ways of capturing religio-scientific perspectives using a smaller number of survey
questions, these items may be worth including in other national and cross-national surveys. Such data would facilitate studying
religio-scientific perspectives across domains,
comparatively, and over time.
Religion’s diminishing influence over certain public affairs has coincided with an
increase in spirituality among individuals,
even among people who are scientifically
inclined (Ecklund and Long 2011; Marler and
Hadaway 2002). Moreover, our findings indicate there is substantial decoupling between
individual and institutional perspectives on
science and religion. For example, despite the
Catholic Church’s acceptance of modern science’s origin theory, some Catholics in our
sample did not support the scientific theory of
the big bang (Pope Pius XII 1951; Ratzinger
1988). Overall, the finding that individual
views of science and religion often differ
from the official positions held by religious
authorities points to the importance of cultural framings in addition to institutional ones
for understanding individual choices about
what to believe and when.
This article suggests that the post-secular
perspective emerged from traditional and
O’Brien and Noy
modern views as a way for individuals to
reconcile competing frameworks (Casanova
2010). An alternative is that traditional and
modern perspectives arose from the postsecular view. This possibility is supported by
Merton’s ([1938] 1973) account of the role of
Puritanism in the professionalization of science and the relatively recent schisms (i.e.,
twentieth century) between scientific and
some religious institutions over evolution.
Cross-sectional analyses risk inaccurately
portraying the boundary between science and
religion as static. Therefore, additional study
is needed to consider how perspectives on
science and religion emerge historically and
Although our findings support theories of
conflict between science and religion, they
also show that many individuals see reason
and faith as more compatible than is often
acknowledged. We advance theory and
research in this area of continued sociological
interest by providing evidence for traditional,
modern, and post-secular perspectives on science and religion. The post-secular view is
not a midpoint between modern and traditional, but a distinct way of using science and
religion to interpret the world. These results
highlight the complexity of this cultural
terrain and the value of examining religioscientific perspectives in understanding divisions in U.S. society.
We thank Patricia McManus, Brian Powell, Brian
Steensland, Mike Vasseur, members of Indiana University’s Politics, Economy, and Culture Workshop, and
ASR’s editors and anonymous reviewers for helpful feedback on this article.
Funding provided by grant number SRS 0935815 from
the National Science Foundation.
1. We prefer the label modern to secular because secularization theories do not necessarily presume conflict between science and religion (Gorski 2000).
2. Many variables for this analysis come from questions asked only to subsamples of respondents in
these survey years. To maximize the sample size
for analysis, we pooled data across GSS waves.
Mean levels for variables of interest are similar
across survey years, and in supplemental regression
models we included binary controls for survey year.
Findings from these supplemental analyses do not
change the conclusions presented in this article.
3. We computed supplemental latent class analysis
models that include manifest variables measuring:
(1) religious attendance, (2) prayer frequency, (3)
belief in an afterlife, and (4) confidence in clergy
(see Table S1 in the online supplement [http://asr]). Additional religion
indicators produced latent classes similar to the
ones presented here in their meanings and sizes, the
attributes of respondents assigned to them, and their
relationships with political attitudes. Ultimately, we
excluded these manifest variables from the final
analysis because of missing data issues.
4. Although we adopt Steensland and colleagues’
(2000) classification, which includes fundamentalist, Pentecostal, charismatic, and evangelical Christian traditions in a single category, we recognize
important historical differences among these traditions (Smith et al. 1998; Woodberry et al. 2012). We
therefore prefer the term conservative Protestant to
evangelical Protestant.
5. In supplemental latent class analysis models, we
analyzed religious attendance as a manifest variable
in combination with religious affiliation strength.
Conclusions from these additional analyses are
similar to those presented here. We do not assume
a causal relationship between religious attendance
and perspectives of science and religion. Rather,
religious attendance is a social characteristic we
examine in relation to religio-scientific perspectives.
6. Household income category midpoints are (in dollars) 500; 2,000; 3,500; 5,500; 6,500; 7,500; 9,000;
11,250; 13,750; 16,250; 18,750; 21,250; 23,750;
27,500; 32,500; 37,500; 45,000; 55,000; 67,500;
82,500; 100,000; 120,000; 140,000; and 172,500.
7. We thank Art Alderson for sharing details of his
analysis with us.
8. We also examined alternative class assignments
obtained using a pseudo-likelihood estimator (Vermunt 2010). Class assignments and results of analyses of manifest and independent variables were
similar based on each class assignment method. We
therefore report results from the more conventional
posterior probability method.
9. We also examined cases excluded due to missing
sociodemographic data. The percentage of post-seculars is not significantly different between the full
and restricted samples (22 and 21 percent, respectively; two-tailed t-test). However, the traditional
class is significantly larger in the restricted compared to the full sample (50 percent compared to 43
percent, p < .05). The restricted sample’s modern
class is correspondingly smaller (28 percent compared to 36 percent, p < .05). Although respondents
in the restricted sample overall have significantly
lower scores on several science knowledge items,
and significantly higher scores on religion items,
levels of manifest variables are similar between corresponding classes in the full and restricted samples.
10. We employ a two-stage rather than a simultaneous estimation strategy for the LCA and regression
analyses. Simultaneous estimation allows independent variables from the regression equation to affect
the formation of latent classes, which we wish to
avoid for the theoretical and substantive reasons
discussed earlier. Furthermore, we prefer Stata’s
post-estimation and prediction capabilities. We
therefore use Mplus for the LCA (not supported by
Stata) and Stata for subsequent analyses.
11. The 2008 survey included 15 science knowledge
questions about uncontested topics such as erosion
and aerodynamics, designed to minimize cultural
bias. In supplemental analyses, we compared additional science knowledge items across latent classes.
Post-seculars and moderns each had significantly
higher scores than traditionals on each item ( p <
.05, two-tailed t-tests). Post-seculars and moderns
differed significantly only on a question about litmus paper. Although post-seculars had a lower mean
score than moderns on an additive scale of correct
responses, the difference was only .54 points. This
further suggests that the post-secular perspective
entails high science literacy and disagreement with
only a few particular scientific theories.
12. Table S2 in the online supplement summarizes
manifest variables for models with four, five, and
six latent classes. Each model includes a class
of post-seculars containing roughly one-fifth of
respondents. Table S3 in the online supplement
examines how respondents were assigned to latent
classes across models. Of post-seculars in the threeclass model, 95 percent were assigned to a common
class in the four-class model. Additionally, 87 and
76 percent of the three-class post-seculars were
assigned to common classes in the five- and sixclass models, respectively. LCA models with more
than three classes created new classes along a continuum defined by traditional and modern perspectives at its poles (see Table S2). The traditional and
modern perspectives are further subdivided by adding more classes; the post-secular class is mostly
unchanged. A unique, postmodern, fourth perspective does not emerge. Overall, these results provide
support for the three-class solution.
13. The most notable denominational differences
between post-secular and traditional conservative
Protestants are the larger share of traditionals who
are Baptist (21 versus 10 percent) or Pentecostal (11
versus 6 percent). Furthermore, nondenominational
conservative Protestants are overrepresented in the
post-secular class compared to the traditional latent
class (22 versus 14 percent).
14. Religious traditions are jointly significant (X 2 =
346.94; p < .001). Supplementary stepwise regressions reveal that relationships between religious
American Sociological Review 80(1)
traditions and latent classes are largely independent of
other sociodemographic characteristics. One exception is that in the nested model, Black Protestants were
significantly more likely to hold traditional rather than
modern perspectives. However, after controlling for
race, the effect of Black Protestantism on the odds of
holding a traditional rather than modern perspective
loses statistical significance. This is consistent with
Shelton and Emerson’s (2012) finding that although
Black Protestants share doctrine with other conservative Christian traditions, the two groups have different approaches to the boundary between science and
religion. Our findings suggest that while conservative
Christian traditions shape religio-scientific perspectives independently of race, the same cannot be said
for Black Protestant traditions.
15. Race variables have a jointly significant effect on
the model (X 2 = 86.77; p < .001).
16. Occupational status variables have a jointly significant effect on the model (X 2 = 33.49; p < .001).
Alderson, Arthur, Azamat Junisbai, and Isaac Heacock.
2007. “Social Status and Cultural Consumption in the
United States.” Poetics 35(2):191–212.
Allum, Nick, Patrick Sturgis, Dimitra Tabourazi, and
Ian Brunton-Smith. 2008. “Science Knowledge and
Attitudes across Cultures: A Meta-Analysis.” Public
Understanding of Science 17(1):35–54.
Argue, Amy, David R. Johnson, and Lynn K. White.
1999. “Age and Religiosity: Evidence from a ThreeWave Panel Analysis.” Journal for the Scientific
Study of Religion 38(3):423–35.
Asparouhov, Tihomir and Bengt Muthén. 2012. “Using
Mplus TECH11 and TECH14 to Test the Number of
Latent Classes.” Mplus Web Notes: No. 14 (https://www
Bak, Hee-Je. 2001. “Education and Public Attitudes
toward Science: Implications for the ‘Deficit Model’
of Education and Support for Science and Technology.” Social Science Quarterly 82(4):779–95.
Baker, Joseph O. 2012. “Public Perceptions of Incompatibility between ‘Science and Religion.’” Public
Understanding of Science 21(3):340–53.
Barbour, Ian G. 2000. When Science Meets Religion:
Enemies, Strangers, or Partners? New York: HarperCollins.
Beck, Ulrich. 1992. Risk Society: Towards a New Modernity, Vol. 17. Thousand Oaks, CA: Sage Publications
Bender, Courtney, Wendy Cadge, Peggy Levitt, and
David Smilde, eds. 2013. Religion on the Edge: Decentering and Re-centering the Sociology of Religion.
New York: Oxford University Press.
Binder, Amy J. 2002. Contentious Curricula: Afrocentrism and Creationism in American Public Schools.
Princeton, NJ: Princeton University Press.
Bruce, Steve. 2011. Secularization: In Defense of an
Unfavorable Theory. Oxford, UK: Oxford University
O’Brien and Noy
Casanova, Jose. 2010. “A Secular Age: Dawn or Twilight.” Pp. 265–81 in Varieties of Secularism in a
Secular Age, edited by M. Warner, J. VanAntwerpen,
and C. Calhoun. Cambridge, MA: Harvard University
Castells, Manuel. 1996. The Rise of the Network Society,
the Information Age: Economy, Society and Culture,
Vol. I. Cambridge, MA; Oxford, UK: Blackwell.
Chaves, Mark. 1994. “Secularization as Declining Religious Authority.” Social Forces 72(3):749–74.
Collins, Francis S. 2006. The Language of God: A Scientist Presents Evidence for Belief. New York: Free
Comte, Auguste. 1998. Auguste Comte and Positivism:
The Essential Writings, edited by G. Lenzer. Piscataway, NJ: Transaction Books.
Correll, Shelley J. 2004. “Constraints into Preferences:
Gender, Status and Emerging Career Aspirations.”
American Sociological Review 69(1):93–113.
Davis, Nancy J., and Robert V. Robinson. 1999. “Their
Brothers’ Keepers? Orthodox Religionists, Modernists, and Economic Justice in Europe.” American
Journal of Sociology 104(6):1631–65.
Dawkins, Richard. 2006. The God Delusion. New York:
Houghton Mifflin.
Dobbelaere, Karel. 1999. “Towards an Integrated Perspective of the Processes Related to the Descriptive
Concept of Secularization.” Sociology of Religion
Drori, Gili S., John W. Meyer, Francisco O. Ramirez, and
Evan Schofer. 2003. Science in the Modern World
Polity: Institutionalization and Globalization. Stanford, CA: Stanford University Press.
Durkheim, Emile. [1912] 1995. The Elementary Forms of
Religious Life. New York: Free Press.
Ecklund, Elaine Howard. 2010. Science vs. Religion:
What Scientists Really Think. New York: Oxford University Press.
Ecklund, Elaine Howard, and Elizabeth Long. 2011.
“Scientists and Spirituality.” Sociology of Religion
Ecklund, Elaine Howard, Jerry Z. Park, and Phil Todd
Veliz. 2008. “Secularization and Religious Change
among Elite Scientists.” Social Forces 86(4):1805–
Evans, John. 2010. Contested Reproduction: Genetic
Technologies, Religion, and Public Debate. Chicago:
University of Chicago Press.
Evans, John. 2011. “Epistemological and Moral Conflict
between Religion and Science.” Journal for the Scientific Study of Religion 50(4):707–727.
Evans, John H., and Michael S. Evans. 2008. “Religion
and Science: Beyond the Epistemological Conflict
Narrative.” Annual Review of Sociology 34:87–105.
Freeman, Patricia K., and David J. Houston. 2011.
“Rejecting Darwin and Support for Science Funding.” Social Science Quarterly 92(5):1151–68.
Gauchat, Gordon. 2012. “Politicization of Science in the
Public Sphere: A Study of Public Trust in the United
States, 1974 to 2010.” American Sociological Review
Giddens, Anthony. 1991. Modernity and Self-Identity:
Self and Society in the Late Modern Age. Stanford,
CA: Stanford University Press.
Goodman, Leo A. 1974. “Exploratory Latent Structure
Analysis Using both Identifiable and Unidentifiable
Models.” Biometrika 61(2):215–31.
Gorski, Philip S. 2000. “Historicizing the Secularization
Debate: Church, State, and Society in Late Medieval
and Early Modern Europe, ca. 1300 to 1700.” American Sociological Review 65(1):138–67.
Gorski, Philip, David Kyuman Kim, John Torpey, and
Jonathan VanAntwerpen. 2012. The Post-Secular in
Question: Religion in Contemporary Society. New
York: New York University Press.
Habermas, Jurgen. 1989. On Society and Politics:
A Reader, edited by S. Seidman. Boston: Beacon
Habermas, Jurgen. 2008. “Secularism’s Crisis of Faith:
Notes on Post-Secular Society.” New Perspectives
Quarterly 25(4):17–29.
Hoffman, John, and Sherrie Mills Johnson. 2005. “Attitudes toward Abortion among Religious Traditions in
the United States: Change or Continuity?” Sociology
of Religion 66(2):161–82.
Hunt, Larry, and Matthew Hunt. 2001. “Race, Religion
and Religious Involvement: A Comparative Study
of Whites and African Americans.” Social Forces
Iannaccone, Lawrence. 1993. “Heirs to the Protestant
Ethic? The Economics of American Fundamentalists.” Pp. 342–66 in Fundamentalisms and the State,
edited by M. Marty and S. Appleby. Chicago: University of Chicago Press.
Inglehart, Ronald, and Wayne Baker. 2000. “Modernization, Cultural Change, and the Persistence of
Traditional Values.” American Sociological Review
Jasanoff, Sheila. 2005. Designs on Nature: Science and
Democracy in Europe and the United States. Princeton, NJ: Princeton University Press.
Latour, Bruno. 1993. We Have Never Been Modern. Cambridge, MA: Harvard University Press.
Lawson, Anton, and William Worsnop. 1992. “Learning about Evolution and Rejecting a Belief in Special Creation: Effects of Reflective Reasoning Skill,
Prior Knowledge, Prior Belief and Religious Commitment.” Journal of Research in Science Teaching
29(2): 143–66.
Lazarsfeld, Paul Felix, and Neil W. Henry. 1968. Latent
Structure Analysis. Boston: Houghton Mifflin.
Lienesch, Michael. 2007. In the Beginning: Fundamentalism, the Scopes Trial, and the Making of the
Antievolution Movement. Chapel Hill: University of
North Carolina Press.
Lo, Yungtai, Nancy R. Mendell, and Donald B. Rubin.
2001. “Testing the Number of Components in a Normal Mixture.” Biometrika 88(3):767–78.
Locke, Simon. 1999. “Golem Science and the Public
Understanding of Science: From Deficit to Dilemma.”
Public Understanding of Science 8(2):75–92.
Longest, Kyle, and Christian Smith. 2011. “Conflicting
or Compatible: Beliefs about Religion and Science
among Emerging Adults in the United States.” Sociological Forum 26(4):846–69.
Luker, Kristin. 1984. Abortion and the Politics of Motherhood. Berkeley: University of California Press.
Magidson, Jay, and Jeroen K. Vermunt. 2001. “Latent
Class Factor and Cluster Models: Bi-plots and
Related Graphical Displays.” Sociological Methodology 31(1):223–64.
Marler, Penny Long, and C. Kirk Hadaway. 2002.
“‘Being Religious’ or ‘Being Spiritual’ in America:
A Zero-Sum Proposition?” Journal for the Scientific
Study of Religion 41(2):289–300.
Marsden, George. [1980] 2006. Fundamentalism and
American Culture: The Shaping of Twentieth-Century
Evangelicalism 1870–1925. New York: Oxford University Press.
Marsden, George. 1991. Understanding Fundamentalism and Evangelicalism. Grand Rapids, MI: Wm. B.
Eerdmans Publishing Co.
Marx, Karl. 1978. “Critique of Hegel’s Philosophy of
Right: Introduction.” Pp. 53–65 in The Marx-Engels
Reader, 2nd ed., edited by R. Tucker. New York:
Merton, Robert. [1938] 1973. “The Puritan Spur to Science.” Pp. 228–53 in The Sociology of Science:
Empirical and Theoretical Investigations, edited by
N. W. Storer. Chicago: University of Chicago Press.
Miller, Alan S., and John P. Hoffman. 1995. “Risk and
Religion: An Explanation of Gender Differences in
Religiosity.” Journal for the Scientific Study of Religion 34(1):63–75.
Miller, Jon. 2004. “Public Understanding of and Attitudes toward Scientific Research: What We Know
and What We Need to Know.” Public Understanding
of Science 13(3):273–94.
National Science Board. 2012. Science and Engineering
Indicators, 2012. Arlington, VA: National Science
Nelkin, Dorothy. 2004. “God Talk: Confusion between
Science and Religion Posthumous Essay.” Science,
Technology & Human Values 29(2):139–52.
Nelsen, Brent, James Guth, and Cleveland Fraser. 2001.
“Does Religion Matter? Christianity and Public Support for the European Union.” European Union Politics 2(2):191–217.
Nylund, Karen L., Tihomir Asparouhov, and Bengt O.
Muthén. 2007. “Deciding on the Number of Classes
in Latent Class Analysis and Growth Mixture Modeling: A Monte Carlo Simulation Study.” Structural
Equation Modeling 14(4):535–69.
Pew Research Center. 2009. “Public Praises Science;
Scientists Fault Public, Media.” (
American Sociological Review 80(1)
Plantinga, Alvin. 2011. Where the Conflict Really Lies:
Science, Religion, and Naturalism. New York:
Oxford University Press.
Pope Pius XII. 1951. “The Proofs for the Existence
of God in the Light of Modern Natural Science.”
Address by Pope Pius XII to the Pontifical Academy
of Sciences, pp. 182–92.
Ratzinger, Joseph C. 1988. Church, Ecumenism and Politics. New York: The Crossroad Publishing Company.
Roos, J. Micah. 2012. “Measuring Science or Religion?
A Measurement Analysis of the National Science
Foundation Sponsored Science Literacy Scale 2006–
2010.” Public Understanding of Science 23(7):797–
Scheufele, Dietram, Elizabeth A. Corley, Tsung-jen
Shih, Kajsa E. Dalrymple, and Shirley S. Ho. 2009.
“Religious Beliefs and Public Attitudes toward Nanotechnology in Europe and the United States.” Nature
Nanotechnology 4(2):91–94.
Schieman, Scott. 2010. “Socioeconomic Status and
Beliefs about God’s Influence in Everyday Life.”
Sociology of Religion 71(1):25–51.
Schnittker, Jason, Jeremy Freese, and Brian Powell.
2000. “Nature, Nurture, Neither, Nor: Black-White
Differences in Beliefs about the Causes and Appropriate Treatment of Mental Illness.” Social Forces
Shapin, Steven. 1996. The Scientific Revolution. Chicago:
University of Chicago Press.
Shapin, Steven. 2008. “Science and Modern Society.” Pp.
433–48 in The Handbook of Science and Technology
Studies, 3rd ed., edited by E. Hackett, O. Amsterdamska, M. Lynch, and J. Wajcma. Cambridge, MA: The
MIT Press.
Shelton, Jason E., and Michael O. Emerson. 2012. Blacks
and Whites in Christian America. New York: New
York University Press.
Sherkat, Darren. 2011. “Religion and Scientific Literacy in the United States.” Social Science Quarterly
Smith, Christian. 2003. The Secular Revolution: Power,
Interests, and Conflict in the Secularization of American
Public Life. Berkeley: University of California Press.
Smith, Christian, with Michael Emerson, Sally Gallagher,
Paul Kennedy, and David Sikkink. 1998. American
Evangelicalism: Embattled and Thriving. Chicago:
University of Chicago Press.
Smith, Christian, and Robert Faris. 2005. “Socioeconomic
Inequality in the American Religious System.” Journal for the Scientific Study of Religion 44(1):95–104.
Smith, Tom W., Peter Marsden, Michael Hout, and Jibum
Kim. 2011. General Social Surveys, 1972–2010
[machine-readable data file]. Principal Investigator,
Tom W. Smith; Co-Principal Investigator, Peter V.
Marsden; Co-Principal Investigator, Michael Hout;
Sponsored by National Science Foundation, NORC.
Chicago: National Opinion Research Center [producer]; Storrs, CT: The Roper Center for Public Opinion Research, University of Connecticut [distributor].
O’Brien and Noy
Stark, Rodney. 1999. “Secularization, R.I.P.” Sociology
of Religion 60(3):249–74.
Stark, Rodney, and William Bainbridge. 1987. A Theory
of Religion. New York: Peter Lang Publishing.
Steensland, Brian, Jerry Z. Park, Mark D. Regnerus,
Lynn D. Robinson, W. Bradford Wilcox, and Robert D. Woodberry. 2000. “The Measure of American
Religion: Toward Improving the State of the Art.”
Social Forces 79(1):291–318.
Taylor, Charles. 2007. A Secular Age. Cambridge, MA:
Belknap Press.
Toumey, Chris. 1994. God’s Own Scientists: Creationists in a Secular World. New Brunswick, NJ: Rutgers
University Press.
Toumey, Chris, John Besley, Meg Blanchard, Mark
Brown, Michael Cobb, Elaine Howard Ecklund,
Margaret Glass, Thomas M. Guterbock, A. Eamonn
Kelly, and Bruce Lewenstein. 2010. “Science in the
Service of Citizens and Consumers.” USC NanoCenter. Retrieved June 15, 2013 (
Triandis, Harry C. 1995. Individualism & Collectivism.
Boulder, CO: Westview Press.
Verhey, Steven D. 2005. “The Effect of Engaging Prior
Learning on Student Attitudes toward Creationism
and Evolution.” BioScience 55(11):996–1003.
Vermunt, Jeroen K. 2010. “Latent Class Modeling with
Covariates: Two Improved Three-Step Approaches.”
Political Analysis 18(4):450–69.
Weber, Max. [1904] 1930. The Protestant Ethic and the
Spirit of Capitalism. New York: Scribner’s.
Wilson, Bryan. 1975. Magic and the Millennium. St.
Albans, UK: Paladin.
Woodberry, Robert D., Jerry Z. Park, and Lyman A. Kellstedt with Mark D. Regnerus and Brian Steensland.
2012. “The Measure of American Religious Traditions: Theoretical and Measurement Considerations.”
Social Forces 91(1):65–73.
Xie, Yu, and Kimberlee A. Shauman. 2003. Women in
Science: Career Processes and Outcomes. Boston:
Harvard University Press.
Timothy L. O’Brien is an Assistant Professor of Sociology at the University of Evansville. His research focuses
on the intersection of science, religion, law, and politics.
One of his overarching interests is in how authority, expertise, and credibility are recognized in different settings. In
another project, he examines the use of scientific, medical,
and other kinds of expert knowledge in U.S. courts.
Shiri Noy is an Assistant Professor of Sociology at the
University of Wyoming. Her research is primarily in
political sociology, and centered on globalization, political culture, institutions, and development. In addition to
her research on science, religion, and public opinion, her
current research focuses on the dynamics of globalization
and health policy reform in Latin America.