Content uploaded by Eugenie Carol Scott
Author content
All content in this area was uploaded by Eugenie Carol Scott on Aug 16, 2021
Content may be subject to copyright.
https://doi.org/10.1177/09636625211035919
Public Understanding of Science
1 –16
© The Author(s) 2021
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/09636625211035919
journals.sagepub.com/home/pus
P U S
Public acceptance of evolution
in the United States, 1985–2020
Jon D. Miller
University of Michigan, USA
Eugenie C. Scott
National Center for Science Education, USA
Mark S. Ackerman
University of Michigan, USA
Belén Laspra
University of Oviedo, Spain
Glenn Branch
National Center for Science Education, USA
Carmelo Polino
University of Oviedo, Spain; Centro Redes, Argentina
Jordan S. Huffaker
University of Michigan, USA
Abstract
The public acceptance of evolution in the United States is a long-standing problem. Using data from a series
of national surveys collected over the last 35 years, we find that the level of public acceptance of evolution
has increased in the last decade after at least two decades in which the public was nearly evenly divided on
the issue. A structural equation model indicates that increasing enrollment in baccalaureate-level programs,
exposure to college-level science courses, a declining level of religious fundamentalism, and a rising level of
civic scientific literacy are responsible for the increased level of public acceptance.
Corresponding author:
Jon D. Miller, Institute for Social Research, University of Michigan, 426 Thompson Street, Ann Arbor, MI 48106, USA.
Email: jondmiller@umich.edu
1035919PUS0010.1177/09636625211035919Public Understanding of ScienceMiller et al.
research-article2021
Research article
2 Public Understanding of Science 00(0)
Keywords
biology/evolution, public acceptance, structural equation model, time series, United States
Acceptance of evolution in the United States increased substantially in the last decade, following
two decades of parity in the acceptance or rejection of evolution by American adults. Since 1985,
national samples of US adults have been asked to agree or disagree with the statement “Human
beings, as we know them today, developed from earlier species of animals.” During the last decade,
the percentage of US adults agreeing with this statement increased from 40% to 54%—a majority.
The level of acceptance of evolution in the United States is atypically low for a developed nation.1
In a study of the acceptance of evolution in 34 developed nations in 2005, only Turkey—at 27%—
scored lower than the United States (Miller et al., 2006). Even with the level of acceptance of evolu-
tion at 54% in 2019–2020, it is likely that the United States remains low in international rankings.2
This article quantifies and analyzes recent trends in the relative influence of American funda-
mentalist religious beliefs, ideological polarization, college-level science education, and other fac-
tors affecting the public acceptance of evolution in the United States. To do this, we utilize a
three-decade-long time series of US national surveys. We provide a statistical description of the
pattern of acceptance of evolution and use a structural equation model (SEM) to identify the pri-
mary variables that predict the acceptance of biological evolution among American adults in 2019.
The data
The data for this time series analysis come from a combination of (a) several years of biennial
National Science Board surveys that were used to produce the Science and Engineering Indicators
reports to Congress, (b) several national surveys funded by other units of the National Science
Foundation, and (c) a new time series focused on adult civic scientific literacy (CSL) funded by the
National Aeronautics and Space Administration. During the three decades covered by this analysis,
the modes and methods of survey data collection have changed. The items used in this analysis
were collected by a mix of telephone interviews and online questionnaires, but there is no correla-
tion between the modes of data collection and acceptance of evolution. For a description of spon-
sorship and methods, see the Supplemental Material.
The patterns of evolution acceptance and rejection since 1985
There are no adequate probability surveys on American attitudes toward evolution prior to the
Second World War. Researchers of American views of evolution employed a variety of questions
over the postwar years (Plutzer and Berkman, 2008). Despite the variety of questions, American
opinion about evolution was divided during the second half of the twentieth century and fewer than
half of Americans accepted evolution in these surveys (Brenan, 2019; IPSOS, 2011; Kampourakis,
2020; Miller et al., 2006; Pew Research Center, 2013, 2015, 2019b; Plutzer and Berkman, 2008;
Swift, 2017).
Starting in 1985, a series of surveys used “Human beings, as we know them today, developed
from earlier species of animals” to investigate attitudes toward evolution. An examination of data
from these surveys indicates that Americans were evenly divided on the evolution question from
1985 through 2007 with the exception of the 1999 respondents,3 but more than half of American
adults now indicate that they accept evolution (see Figure 1). The increase in acceptance of evolu-
tion reflects a small decline in overt rejection and a larger decline in the proportion of American
adults who are unsure about the evolution issue.
Miller et al. 3
Gallup (Brenan, 2019; Swift, 2017) and the Pew Research Center (2019b) have used a similar
question that asks respondents to indicate whether they think that humans have evolved over time
and, if so, whether God had any role in this process. Plutzer and Berkman (2008), the Pew Research
Center (2019b), and Kampourakis (2020) have examined the impact of question wording and
found conflicting advantages and disadvantages to the inclusion of multiple dimensions in the
same question. We think that the simple question asking whether humans evolved over a long
period is a useful and clearer indicator of respondent acceptance or rejection of evolution.
To explain the observed increase in acceptance of evolution by American adults, we look at
several factors that we hypothesize are associated with the acceptance or rejection of evolution.
The factors predicting acceptance or rejection of evolution
Bivariate cross-tabulations allow us to examine the strength of each variable’s association with the
public acceptance of evolution and to see how this relationship changed between 1988 and 2019.
We first discuss the demographic variables (age, gender, educational attainment, college science
courses taken, minor children at home), in decreasing order of strength (as measured by the abso-
lute value of the ordinal correlation coefficient gamma) in the 2019 data (see Table 1).4
Figure 1. Public acceptance and rejection of evolution in the United States, 1985–2020.
The following question was used in all of the years in this analysis:
For each statement below, please indicate if you think that it is definitely true, probably true, probably false, or definitely
false. If you don’t know or aren’t sure, please check the “not sure” box.
Human beings, as we know them today, developed from earlier species of animals.
The number of respondents in each year and the confidence intervals are provided in the Supplemental Material in SI
Table 1.
4 Public Understanding of Science 00(0)
Completion of one or more college science courses was the strongest predictor of adult accept-
ance of evolution among the demographic variables (see Table 1). An important part of American
higher education is a unique requirement for students to complete a year of general education
courses, including science courses, to earn an accredited baccalaureate (Hyman et al., 1975; Muller,
2010). The gamma for this relationship was 0.23 in 1988 and 0.25 in 2019, indicating the stable and
substantial influence of college science courses.
The level of educational attainment (highest degree earned) has been positively associated with
the acceptance of evolution during the last 30 years (see Table 1). It is important to note that the
percentage of American adults who earned a baccalaureate increased from 20% in 1988 to 35% in
2018 (United States Census Bureau, 2019).
Over the last 30 years, age has had a negative relationship with the acceptance of evolution (see
Table 1). In 1988, the gamma for this relationship was –0.18 and 30 years later it was –0.17, indi-
cating that the influence of age on the acceptance of evolution has remained stable over the last
Table 1. The acceptance of evolution, by selected factors: 1988, 2019.
1988 2019
% accept
evolution
gamma % accept
evolution
gamma
All adults 46 – 54 –
Age
18–24 54 –0.18 68 –0.17
25–34 50 65
35–44 53 54
45–54 41 50
55–64 36 45
65 and above 37 45
Gender
Male 52 –0.23 57 –0.13
Female 41 51
Education
Less than high school 41 0.19 53 0.20
High school grad/GED 44 47
Associate degree 38 54
Baccalaureate 61 58
Graduate/prof degree 67 69
College science courses
None 42 0.23 47 0.25
1–3 courses 51 58
4 or more courses 59 65
Minor children at home
None 46 ns 54 –0.06
One or more 46 52
Number of cases 2,041 2,738
ns = not significant at the .05 level.
Although data concerning the proportion of American adults who accept evolution is available from 1985 through 2020,
not all of the predictor variables were asked in each of those years. Some items measuring religious fundamentalism
were not asked in 1985 and 1988. The year-to-year variation is not significant at the .05 level, so we utilize 1988 and
2019 as the end points of a three-decade distribution for illustration purposes.
Miller et al. 5
three decades. The negative sign indicates that younger respondents are slightly more likely to
accept evolution than older respondents.
Throughout the last three decades, a higher proportion of men than women have accepted evo-
lution, but the magnitude of this difference is declining (see Table 1). This narrowing of the gender
difference may reflect the growth of educational attainment during the same period—the percent-
age of women with a baccalaureate increased from 17% to 35% over the last 30 years (United
States Census Bureau, 2019).
Finally, another family demographic of interest is the number of minor children in the respond-
ent’s home. Although we often think of adults as transmitting science information to their children,
research in recent decades has shown that many parents learn new science information through
helping their children with homework, assisting with science fair projects, and answering questions
that occur during science museum visits or the viewing of science-related events on television
(Bengtson et al., 2002; Shumow and Miller, 2001). Increasingly, children may ask their parents
about issues like climate change. Educational scholars often refer to this as the science fair effect.
In this analysis, the results from 1988 indicated that the presence of minor children at home was
not related to the acceptance of evolution, but the 2019 results show a small negative relationship
(gamma = –.06), indicating that the parents of minor children were slightly less likely to accept
evolution than adults without minor children at home. We will return to this issue in our SEM
analysis below.
In addition to personal and family demographic variables, other relevant variables were consid-
ered (religious fundamentalism, civic scientific literacy [CSL], ideological partisanship, level of
interest in selected science policy issues, the use of informal science education resources, and pat-
terns of individual science information acquisition). We discuss them below in decreasing order of
strength (as measured by the absolute value of the ordinal correlation coefficient gamma) in the
2019 data.
First, any analysis of the acceptance of evolution by American adults must take into account the
impact of religious fundamentalism (Chaves, 2017; Emerson and Hartman, 2006; Putnam and
Campbell, 2010; Wilcox and Larson, 2006). All of the surveys used in this analysis include some
measures of religious fundamentalism, but they sometimes used slightly different questions. For
those years in which all five of the variables5 used to measure religious fundamentalism were avail-
able, we used a confirmatory factor analysis (CFA) to determine that all of the items loaded on a
single factor. The factor scores were used to compute a zero-to-10 index for each year. The gamma
for the relationship between the CFA-derived Index of Religious Fundamentalism and the accept-
ance of evolution was –0.71 in 2007 and –0.71 in 2019, indicating that there was no change in this
relationship over the last 12 years. A description of the 2007 and 2019 CFAs is included in the
Supplemental Material for this article.
The next strongest factor was CSL, a measure of the ability of citizens to read about and make
sense of public policy issues involving science and technology6 (Miller, 1983, 1987, 1998, 2000,
2010a, 2012; Miller et al., 1997; Shen, 1975). The number of college science courses taken and
educational attainment (discussed earlier) are strong predictors of CSL. CSL was assessed in the
1988 and 2019 surveys, and the results demonstrate a strong relationship to the acceptance of evo-
lution over the last 30 years, with gammas of 0.59 and 0.51, respectively (see Table 2). The small
decline in the strength of the association may reflect the growing proportion of American adults
who qualified as civic scientifically literate over the last three decades—an increase from 11% in
1988 to 30% in 2019 (Miller, 2010a).
For the last 100 years, the acceptance or rejection of evolution has been politicized in the United
States. The growth of ideological partisanship and affective partisanship in the postwar years and
its acceleration in recent decades has made evolution a component of the alliance of evangelical
6 Public Understanding of Science 00(0)
Table 2. The acceptance of evolution, by selected factors: 1988, 2019.
1988* 2019
% accept
evolution
gamma % accept
evolution
gamma
All adults 46 – 54 –
Religious fundamentalism (2007 and 2019)
Index score 0–1 78 –0.70 91 –0.71
Index score 2–3 74 69
Index score 4–6 39 39
Index score 7–8 11 18
Index score 9–10 8 32
Civic scientific literacy
Not scientifically literate 43 0.59 45 0.51
Scientifically literate 74 74
Ideological partisanship (2007 and 2019)
Conservative Republican 25 0.26 34 0.34
Moderate Republican 40 49
Conservative independent 37 41
Independent 41 51
Liberal independent 65 78
Moderate Democrat 35 53
Liberal Democrat 73 83
Science information acquisition
None 45 0.22 50 0.21
1–26 times/year 37 56
27–52 times/year 41 67
53–104 times/year 43 57
105 or more times/year 59 71
Science issue interest
No science issue interest 44 0.16 44 0.19
Very interested in one issue 36 49
Very interested in two issues 50 53
Very interested in three issues 51 56
Very interested in 4+ issues 56 65
Informal science learning resource use
0–12 times/year 46 0.08 50 0.07
13–24 times/year 44 57
25–36 times/year 42 56
37–48 times/year 47 53
49–60+ times/year 54 58
Number of respondents 2041 2738
*Although data concerning the proportion of American adults who accept evolution is available from 1985 through
2020, some of the predictor variables were not asked in each of those years. Items measuring religious fundamentalism
were not asked in 1985 and 1988 and are reported for 2007 instead.
Christianity and the Republican Party (Abramowitz., 2010, 2018; Levandusky, 2009; Williams,
2012); evolution has emerged as an ideological litmus test of conservatism. Using our seven-point
scale of ideological partisanship in 2019, we observe that 83% of liberal Democrats accept
Miller et al. 7
evolution compared to 34% of conservative Republicans (see Table 2). The relationship is roughly
linear. To assess the relative importance of ideological partisanship in the context of religious fun-
damentalism and college science course exposures, the same seven-point scale of ideological par-
tisanship is included in our SEM.
An increasingly important—and changing—measure concerns adult science information acqui-
sition. The traditional broadcast model of information dissemination is being replaced by a just-in-
time information acquisition system that takes advantage of the Internet and related digital
communication technologies (Case and Given, 2016; Miller, 2010b, 2010c; Miller et al., 2021).
Because of the changes in information technologies over the 30 years, our measures have varied to
reflect the dominant technologies and practices of the time. For example, the 1988 survey did not
ask about acquiring information from the Internet because the first home-based Internet service
was not introduced until 1989. In 2019, the relationship between the frequency of science informa-
tion acquisition (which includes the Internet, various wireless communication services, electronic
and print books and newspapers, and face-to-face and electronic discussions with friends and fam-
ily) and acceptance of evolution was relatively strong, with a gamma of 0.21 (see Table 2).
The level of individual interest in various science and technology policy issues might be
expected to be an important indicator to include in this analysis. Some individuals frequently read
about and pay attention to (a) new scientific discoveries, (b) new medical discoveries, (c) new
inventions and technologies, (d) space exploration, and/or (e) climate change. In this analysis,
respondents were asked about all five of these issues in 2019, but respondents to our 1988 survey
were not asked about climate change. In 1988, the major environmental issue concerned acid rain;
climate change was not a part of the public policy vocabulary for most American adults. Although
the measures are not identical, they are comparable, and the results indicate that adults who follow
science policy issues are only slightly more likely to accept evolution (see Table 2).
A related measure of adult involvement with science and technology is the frequency of use of
traditional science learning resources such as science museums, natural history museums, planetar-
iums, zoos, aquariums, and arboretums. The results indicate that the frequency of use of informal
science learning resources is positively related to the likelihood of accepting evolution, although
not strongly (see Table 2).
A model to predict the acceptance of evolution
The factors discussed in our bivariate analysis do not operate independently. The acceptance or
rejection of evolution does not occur spontaneously; it is a dependent belief mediated by mental
schemas and life experiences (Axelrod, 1973; Schank and Abelson, 1977; Sternberg and Ben-Zeev,
2001). Each individual has an age and a gender, has completed some level of education, has expe-
rienced or not experienced some college-level science courses, has some ideological partisan
views, and engages in various forms of adult learning and information acquisition. He or she may
or may not hold fundamentalist religious views. A SEM provides an analysis that can estimate the
relative influence of each of these factors in a multivariate context (Hayduk, 1987; Jöreskog and
Sörbom, 1993). We use our 2019 data.
A SEM is a path model that takes chronological and logical order into account in seeking to
understand the predictive power of independent variables in the model (Hayduk, 1987; Jöreskog
and Sörbom, 1993). In the diagram of a SEM, influence flows from left to right, reflecting chrono-
logical or logical order (see Figure 2). For example, the level of education that individuals attain
may be influenced by their age or gender or both, but their level of educational attainment cannot
influence or change their age or gender, so age and gender are placed to the left of educational
attainment. Whenever there is a statistically significant relationship between two variables in this
8 Public Understanding of Science 00(0)
Variables Total
Effect
Religious fundamentalism -0.60(.03)
Civic scientific literacy 0.32(.06)
College science courses 0.23(.03)
Education (highest degree attained) 0.19(.03)
Age (in six ordinal categories) -0.19(.03)
Ideological partisanship (liberal Democrat high) 0.18(.03)
Interest in science, technology, medical, space, or climate issues 0.08(.03)
Minor children in the home -0.08(.02)
Science information acquisition activities 0.07(.02)
Gender (female) -0.05(.01)
Use of informal science learning resources 0.00(.00)
R2 0.50
Chi-squares = 141.8; degrees of freedom = 31; Root Mean Square Error of
Approximation (RMSEA) = 0.009; the upper 10% confidence interval of
RMSEA = 0.019; N = 2,146.
Figure 2. A model to predict acceptance of evolution, 2019.
model, they are connected by an arrow reflecting the direction of the influence. Each arrow is
associated with a coefficient ranging from –1.0 to +1.0, a measure of the relative strength of that
relationship. The absence of an arrow means that the relationship between the variables was not
statistically significant at the .05 level.
Miller et al. 9
To estimate the total influence of each variable in the model on the outcome variable—here the
acceptance of evolution—the SEM computes a total effect, ranging from –1.0 to +1.0, which is the
sum of the products of all of the path coefficients in any path that runs from a specific variable to
the outcome variable (see the example in the Supplemental Material). While the total effect lacks
a Proportional Reduction in Error (PRE) interpretation, it is a good indicator of the relative influ-
ence of variables in any specific model.
The SEM for the 2019 data indicates that religious fundamentalism is the strongest predictor of
acceptance of evolution, with a total effect of –0.60 (see Figure 2). The negative effect means that,
holding constant the preceding variables in the model, individuals with a high level of religious
fundamentalism are unlikely to accept evolution.
The level of CSL is the second strongest predictor of the acceptance of evolution, with a total
effect of 0.32 (see Figure 2). The successful acquisition of CSL is strongly influenced by the num-
ber of college science courses completed and the highest level of education attained (Miller, 1998,
2000, 2010a), which are the third and fourth strongest predictors of the acceptance of evolution,
with total effects of 0.23 and 0.19, respectively (see Figure 2). Age, with a total effect of –0.19, is
the fourth strongest predictor.
Consistent with prior research, individuals with more formal education and more experience
with college science courses tend to have a higher level of interest in science and technology
broadly and in public policy regarding science and technology specifically. In the 2019 model,
individuals with more interest in science and technology issues were more likely to accept evolu-
tion (total effect = 0.08). Adults with higher levels of educational attainment and exposure to col-
lege science courses were significantly less likely to hold fundamentalist religious beliefs.
The level of liberal ideological partisanship is weakly predicted by female gender (0.10) and col-
lege science courses (0.07). The intensity of ideological partisanship is U-shaped, with the highest
level reported by conservative Republicans and liberal Democrats. This means that the small posi-
tive total effect between female gender and liberal ideological partisanship would translate into a
similar small positive relationship between male gender and conservative ideological partisanship.
The scale on which ideological partisanship is scored is nominal and has no substantive meaning per
se. But the two ends of the ideological partisanship scale have very different views of evolution with
a total effect of 0.18, meaning that individuals with liberal ideological partisan views are more likely
to accept evolution than individuals with conservative political views. We will return to a discussion
of the relative impact of ideological partisanship in our closing discussion.
Prior research indicates that higher levels of educational attainment and participation in college
science courses provides a set of cognitive tools that enable individuals to more effectively seek
science information in the Internet era (Miller, 2010b, 2010c; Miller et al., 2021; van Dijk, 2020).
The 2019 model is consistent with this view: individuals who report a higher level of overt infor-
mation-seeking on the Internet and related digital platforms were slightly more likely to accept
evolution than individuals who did not engage in this kind of science information-seeking (total
effect = 0.07), holding constant the preceding demographic and educational variables.
The total effect of minor children in the respondent’s household was –0.08 (see Figure 2). The
SEM indicates that the total effect of gender (with female coded as 1 and male as 0) on acceptance
of evolution is –0.05, lower than might be suggested by the raw percentages of men and women
who accept evolution, reflecting the differential gender impact on other intervening variables in the
model.
Older adults were less likely to accept evolution than younger adults (total effect = –0.19). The
paths in Figure 2 from age to the acceptance of evolution indicate that this result is the product of
having a lower exposure to college science courses, holding more fundamentalist religious beliefs,
and being less active in seeking current science information.
10 Public Understanding of Science 00(0)
Overall, the SEM is a good fit for the data, accounting for 50% of the total covariance in the
model (see Figure 2). This level of fit indicates that the structural assumptions of the model fit the
available data, and the path diagram shows a clear and understandable set of pathways of variables
relevant to the acceptance of evolution.
Discussion
The level of public acceptance of evolution in the United States has increased in the last decade—a
majority of American adults now accept biological evolution (see Figure 1). Given the changes over
time in this series, we infer that the increase in acceptance of evolution after 2008 is largely due to
changes in the “unsure” population, rather than the conversion of deniers of evolution, although there
was a small decline in the proportion of Americans who held fundamentalist views during this period.
Considering the comparison of the 1988 and 2019 data and the SEM for 2019, it is now possible
to address two important questions: What factors are related to the increased level of acceptance of
evolution in the United States? And what factors are related to the level of rejection of evolution in
the United States—still higher than that found in the vast majority of developed countries (Branch,
2009) and even in some countries with less industrialized and emerging economies (Branch 2009;
Kampourakis, 2020; Pew Research Center, 2020).
To answer the first question, education plays a significant role in the increased level of accept-
ance of evolution. Survey results from 1988 and 2019 found a stable positive association between
the level of educational attainment and the acceptance of evolution over the last three decades. The
proportion of American adults who earned at least a baccalaureate almost doubled over this 30-year
period—from 20% in 1988 to 35% in 2018 (United States Census Bureau, 2019). The influence of
participation in college science courses increased over the same period, demonstrating the impor-
tance of the uniquely American requirement for general education at the baccalaureate level. From
the 1980s, with John Moore’s seminal “Science as a Way of Knowing” series of essays on the
teaching of undergraduate biology to a recent Gordon Conference on understanding undergraduate
biology education, biologists have been debating how to improve undergraduate education (Austin,
2018; Dolan et al., 2020; Society for Integrative Comparative Biology, n.d.). Notably, teaching
evolution was central to Moore’s project, as well as to subsequent efforts by the AAAS, the National
Research Council, the Howard Hughes Medical Institute, and others. It is not a surprise to find an
increase in college-level science course associated with an increase in the acceptance of evolution.
The association plausibly may be both direct and mediated through secondary science education
since public high school biology teachers are more likely to teach evolution as a matter of scientific
consensus if they have studied evolution themselves at the college level (Plutzer et al., 2020).
The proportion of American adults qualifying as civic scientifically literate increased from 11%
in 1988 to 31% in 2019 (Miller, 2010a), and the relationship between CSL and acceptance of evo-
lution was steady with a gamma of approximately .50 throughout the last three decades. This
underscores the linkage between formal science education and the acceptance of evolution. The
weak positive relationship between the level of overt science information seeking and acceptance
of evolution suggests that information acquisition activities and informal science learning broadly
may provide important assistance to adults to sustain and enrich their level of CSL during the adult
years following post-secondary study. But there is no evidence that informal science learning or
focused information acquisition can substitute for formal education.
Changes in the American religious landscape may have influenced the change in the acceptance
of evolution shown here. Religious participation and adherence to organized religion are declining
(although still high in comparison to other developed countries) and confidence in the literal truth
of the Bible has declined (Chaves, 2017). A reduction in adherence to conservative Protestant
Miller et al. 11
beliefs might contribute to an increase in the acceptance of evolution. The percentage of Americans
who reject evolution remains high in comparison to other developed countries, which leads to the
second question regarding the continuing high level of rejection of evolution in the United States.
Religious fundamentalism plays a significant role in the rejection of evolution. The historical
explanation of the low rate of acceptance of evolution in the United States involves the central
place of the Bible in American Protestantism. In a country settled piecemeal by colonists of vary-
ing religious views and without a state church, it was natural for people of faith, especially
Protestants who already accepted the principle of sola Scriptura, to privilege the Bible—or their
interpretation of it—as the primary source of religious authority and an inerrant source of informa-
tion about history and science as well as faith and morals. In contrast, religion in European coun-
tries is strongly structured by ecclesiastic institutions and the public receptivity to creationism has
been limited as a result (Blancke et al., 2014; Branch, 2009).
It is thus a particular form of religion that is at the foundation of American anti-evolutionism of
the twentieth and twenty-first centuries, not religion in general (see Coyne, 2012, for a dissenting
view). Indeed, evolution is routinely taught in Catholic parochial schools in the United States, and
mainstream Protestant denominations similarly accept evolution (Martin, 2010). While not all anti-
evolutionism originates in Fundamentalism and its inerrantism about the Bible, it largely reflects a
conservative form of Protestantism with relatively inflexible and inerrantist religious views (Scott,
2009), which we have been calling fundamentalism.
Hill (2014) argues that religious beliefs are a stronger predictor of attitude toward evolution
than education in the United States. Using odds-ratio data in a set of logistic regression models,
Hill employs only a tripartite measure of educational attainment, including no measures of scien-
tific literacy or understanding, and finds that respondents’ view of the Bible is a stronger predictor.
He also finds that living in a homogeneous religious community with reinforcement from a reli-
gious social network increases the impact of Biblical beliefs. While agreeing with the primacy of
religious belief in predicting attitudes toward evolution, we contend that Hill’s use of ordinary least
squares regression models with limited measures of educational attainment or scientific literacy
produces an enhanced estimate of the impact of religiosity and a reduced estimate of the influence
of education and literacy. We believe that the SEM analysis model we use provides a more accurate
portrait of the factors associated with the acceptance or rejection of evolution. Applying binary
logistic regression models to data from a different national survey, Baker (2013) offers an assess-
ment of the influence of both religion and educational variables similar to ours, concluding that
higher education influences the acceptance of evolution only for individuals who do not hold a
literalist interpretation of the Bible.
Our 2019 SEM documents the continuing tension between religious fundamentalism and the
acceptance of evolution in the United States. Although scientific literacy has grown, and science
continues to have pervasive influence in American society, a tension between religious fundamen-
talism and evolution remains. We think that there are two primary reasons.
First, a significant portion of American adults continues to hold fundamentalist religious beliefs.
Although there is some disagreement over the definition of religious fundamentalism (Marsden,
1991; Williams, 2012), in broad terms, approximately 30% of American adults hold fundamentalist
religious beliefs largely rooted in Protestantism. Religious fundamentalists are not only more likely
to reject evolution themselves, but are also more likely to support efforts to undermine the teaching
of evolution in the public schools (Baker, 2013; Berkman and Plutzer, 2010; Branch et al., 2010).
Over the last century, such efforts have taken various forms—banning the teaching of evolution
outright, balancing evolution with biblical creationism, creation science, or intelligent design, or
belittling evolution as controversial. Even if most of the efforts have not been successful, the
intended effect is to encourage students to reject evolution.
12 Public Understanding of Science 00(0)
Second, the acceptance or rejection of evolution has been politicized in recent decades.
According to the Pew Research Center (2015), 54% of Republicans and 64% of Democrats
accepted human evolution in 2009; this 10-point gap grew to a 24-point gap 4 years later, when
43% of Republicans and 67% of Democrats accepted evolution. The Republican Party has culti-
vated conservative Christians who share “a Protestant-based moral order” (Williams, 2012) as part
of its base. In local elections, such as for school board, the teaching of evolution may be electorally
salient, especially in smaller communities with a large proportion of religiously conservative citi-
zens. While not as much a “litmus test” of Republican identity as abortion or the rejection of
anthropogenic climate change (Pew Research Center, 2019a), anti-evolutionism is part of the polit-
ical mix, and may augment fundamentalism in maintaining the core of anti-evolutionism in
American society.
To provide a more precise estimate of the role of ideological partisanship in the acceptance or
rejection of evolution, we included a measure of ideological partisanship in our 2019 SEM. In this
model, ideological partisanship has a total effect of 0.18, indicating that liberal Democrats were
more likely to accept evolution holding constant other preceding variables and, conversely, that
conservative Republicans were less likely to accept evolution holding constant other preceding
variables. Because this model is cross-sectional rather than longitudinal, we have no evidence
concerning the early development of attitudes toward evolution, but this model indicates that
among adults, ideological partisanship is not the driving force in influencing evolution attitude.
What does the future hold for the acceptance of evolution in the United States? On one hand,
our model indicates that post-secondary education and exposure to college science courses were
positively related to the acceptance of evolution throughout the last three decades, but the propor-
tion of American adults earning a baccalaureate and participating in college science courses has
increased markedly during this period. The continued growth of educational attainment among
American adults in the twenty-first century suggests that we might expect a moderate rate of
growth in the public acceptance of evolution in the United States in the decades ahead. On the other
hand, changes in the religious profile of Americans have been taking place; both the number of
atheists and the number of Americans who do not identify with organized religion (the “nones”)
have increased in the last decade. Some analyses have found increased political and social liberal-
ism (e.g. increased acceptance of homosexuality) in recent surveys of religious Americans (Pew
Research Center, 2015). If this is correct, a reduction in the influence of fundamentalism along
with growing civic scientific literacy is likely to continue these trends in increasing acceptance of
evolution.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publica-
tion of this article: The data reported in this article was collected with the support of the U.S. National Science
Foundation (awards: SRS85-17581, SRS88-07409, SRS90-02467, SRS92-17876, SRS99-06416); the
Foundation BBVA (Spain); the U.S. National Aeronautics and Space Administration (coopoerative agree-
ment: NNX16AC66A). We are grateful for this support, but all of the findings and conclusions reported in this
work are the responsibility of the authors and do not necessarily reflect the views of any funding agency.
ORCID iDs
Jon D. Miller https://orcid.org/0000-0001-8714-0126
Eugenie C. Scott https://orcid.org/0000-0003-3795-8207
Belén Laspra https://orcid.org/0000-0003-4553-4885
Glenn Branch https://orcid.org/0000-0002-4931-3935
Miller et al. 13
Carmelo Polino https://orcid.org/0000-0003-1789-8024
Jordan S. Huffaker https://orcid.org/0000-0003-2876-9842
Supplemental material
Supplemental material for this article is available online.
Notes
1. We use “developed nations” to denote countries that are wealthier and industrialized. The member coun-
tries of the Organisation for Economic Co-operation and Development (OECD) reflect our use of this
term (see oecd.org).
2. Using different wording, a 20-country survey from the Pew Research Center (2020) estimated the rate
of acceptance of evolution in the United States to be 64%, ranking the United States 15th, ahead of only
Poland, Singapore, India, Brazil, and Malaysia.
3. We have examined the mode, vendors, and procedures used during the 1985 through 2007 period and we
can find no methodological explanation of the higher rate of public acceptance in 1999. The magnitude
of this increase barely exceeded the levels required for statistical significance at the .05 level, but this
pattern was not repeated in subsequent years during the first decade of the twenty-first century.
4. Gamma is a proportional reduction of error (PRE) measure and should be interpreted in the same way as
an R2 in standard regression. Costner (1965) has provided the classic definitional discussion of gamma
and related PRE statistics.
5. The five items used are the following: (1) agreement that “There is a personal God that hears the prayers
of individuals,” (2) agreement that “The Bible is the actual word of God and is to be taken literally,” (3)
the self-reported number of times that each respondent attends a religious service in a typical week, (4)
the self-reported number of times that each respondent prays during a typical week, and (5) agreement
that “We depend too much on science and not enough on faith.”
6. A discussion of the measurement of civic scientific literacy (CSL) is provided in the Supplemental
Material.
References
Abramowitz A (2010) The Disappearing Center: Engaged Citizens, Polarization, and American Democracy.
New Haven, CT: Yale University Press.
Abramowitz A (2018) The Great Alignment: Race, Party Transformation, and the Rise of Donald Trump.
New Haven, CT: Yale University Press.
Axelrod R (1973) Schema theory: An information processing model of perception and cognition. American
Political Science Review 67: 1248–66.
Austin A (2018) Vision and Change in Undergraduate Biology Education: Unpacking a movement and shar-
ing lessons learned. Meeting Report. American Association for the Advancement of Science. Available
at: https://live-visionandchange.pantheonsite.io/wp-content/uploads/2018/09/VandC-2018-finrr.pdf
Baker JO (2013) Acceptance of evolution and support for teaching creationism in public schools: The con-
ditional impact of educational attainment. Journal for the Scientific Study of Religion 52(1): 216–228.
Bengtson VL, Biblarz TJ and Roberts REL (2002) How Families Still Matter: A Longitudinal Study of Youth
in Two Generations. Cambridge: Cambridge University Press.
Berkman M and Plutzer E (2010) Evolution, Creationism, and the Battle to Control America’s Classrooms.
New York, NY: Cambridge University Press.
Blancke S, Hjermitslev HH and Kjaergaard PC (eds) (2014) Creationism in Europe. Baltimore, MD: Johns
Hopkins University Press.
Branch G (2009) Creationism as a global phenomenon. In: Robbins RH and Cohen MN (eds) Darwin and the
Bible: The Cultural Confrontation. Boston, MA: Pearson Education, pp. 137–151.
Branch G, Scott EC and Rosenau J (2010) Dispatches from the evolution wars: Shifting tactics and expanding
battlefields. Annual Review of Genomics and Human Genetics 11: 317–338.
14 Public Understanding of Science 00(0)
Brenan M (2019) 40% Of Americans Believe in Creationism. Available at: https://news.gallup.com/
poll/261680/americans-believecreatioism.aspx
Case DO and Given LM (2016) Looking for Information: A Survey of Research on Information Seeking,
Needs, and Behavior. Bingley: Emerald.
Chaves M (2017) American Religion: Contemporary Trends. 2nd ed. Princeton, NJ: Princeton University
Press.
Costner HL (1965) Criteria for measures of association. American Sociological Review 30(3): 341–353.
Coyne J (2012) Science, religion, and society: The problem of evolution in America. Evolution 66(8): 2654–
2663.
Dolan E, Borrero M, Callis-Duehl K, Musgrove MMC, de Lima J, Ero-Tolliver, et al. (2020) Undergraduate
biology education research Gordon research conference: A meeting report. CBE – Life Sciences
Education 19(20): mr1
Emerson MO and Hartman D (2006) The rise of fundamentalism. Annual Review of Sociology 32: 127–144.
Hayduk LA (1987) Structural Equation Modeling with LISREL. Baltimore, MD: The Johns Hopkins
University Press.
Hill JP (2014) Rejecting evolution: The role of religion, education, and social networks. Journal for the
Scientific Study of Religion 53(3): 575–594 (Reprinted in Mirola WA, Emerson MO and Monahan SC
(eds) Sociology of Religion: A reader. New York, NY: Routledge. 2019. pp. 361–75).
Hyman HH, Wright CR and Reed JS (1975) The Enduring Effects of Education. Chicago, IL: University of
Chicago Press.
Ipsos (2011) Supreme Being, the Afterlife, and Evolution. Available at: https://www.ipsos.com.
Jöreskog K and Sörbom D (1993) LISREL 8. Chicago, IL: Scientific Software International.
Kampourakis K (2020) Understanding Evolution. Cambridge: Cambridge University Press.
Levendusky M (2009) The Partisan Sort: How Liberals Became Democrats and Conservatives Became
Republicans. Chicago, IL: University of Chicago Press.
Marsden G (1991) Understanding fundamentalism and evangelicalism. Grand Rapids, MI: Eerdmans
Publishing.
Martin JW (2010) Compatibility of major U.S. Christian denominations with evolution. Evolution: Education
and Outreach 3(3): 420–431.
Miller JD (1983) Scientific literacy: A conceptual and empirical review. Daedalus 112(2): 29–48.
Miller JD (1987) Scientific literacy in the United States. In: Evered D and O’Connor M (eds) Communicating
Science to the Public. London: Wiley, pp. 19–40.
Miller JD (1998) The measurement of civic scientific literacy. Public Understanding of Science 7: 1–21.
Miller JD (2000) The development of civic scientific literacy in the United States. In: Kumar DD and Chubin
D (eds) Science, Technology, and Society: A Sourcebook on Research and Practice. New York, NY:
Plenum Press, pp. 21–47.
Miller JD (2010a) The conceptualization and measurement of civic scientific literacy for the 21st century. In:
Hildebrand JG and Meinwald J (eds) Science and the Educated American: A Core Component of Liberal
Education. Cambridge, MA: American Academy of Arts and Sciences, pp. 241–255.
Miller JD (2010b) Civic scientific literacy: The role of the media in the electronic era. In: Kennedy D and
Overholser G (eds) Science, Technology, and the Media. Cambridge, MA: American Academy of Arts
and Sciences, pp. 44–63.
Miller JD (2010c) Adult science learning in the Internet era. Curator 53(2): 191–208.
Miller JD (2012) The sources and impact of civic scientific literacy. In: Bauer MW, Shukla R and Allum
N (eds) The Culture of Science: How the Public Relates to Science across the Globe. New York, NY:
Routledge, pp. 217–240.
Miller JD, Ackerman MA, Laspra B and Huffaker J (2021) The acquisition of health and science information
in the 21st century. The Information Society 37(2): 82–98.
Miller JD, Pardo R and Niwa F (1997) Public Perceptions of Science and Technology: A Comparative Study
of the European Union, the United States, Japan, and Canada. Madrid: BBV Foundation Press.
Miller JD, Scott EC and Okamoto S (2006) Public acceptance of evolution. Science 313: 765–766.
Miller et al. 15
Muller RA (2010) Physics for Future Presidents. In: Hildebrand JG and Meinwald J (eds) Science and the
Educated American: A Core Component of Liberal Education. Cambridge, MA: American Academy of
Arts and Sciences, pp. 112–129.
Pew Research Center (2013) Public’s Views on Human Evolution. Available at: https://www.pewforum.
org/2013/12/30/publics-views-on-human-evolution/
Pew Research Center (2015) America’s Changing Religious Landscape. Available at: https://www.pewfo-
rum.org/2015/05/12/americas-changing-religious-landscape/
Pew Research Center (2019a) U.S. Public Views on Climate and Energy. Available at: https://www.pewre-
search.org/science/2019/11/25/u-s-public-views-on-climate-and-energy/
Pew Research Center (2019b) The Evolution of Pew Research Center’s Survey Questions about the Origins
and Development of Life on Earth. Available at: https://www.pewforum.org/2019/02/06/the-evolution-
of-pew-research-centers-survey-questions-about-the-origins-and-development-of-life-on-earth/
Pew Research Center (2020) Biotechnology Research Viewed with Caution Globally, but Most Support Gene
Editing for Babies to Treat Disease. Available at: https://www.pewrearch.org/2020/12/10/biotechnol-
ogy-research-viewed-with-caution-globally-but-most-support-gene-editing-for-babies-to-treat-disease/
Plutzer E and Berkman M (2008) Evolution, creationism, and the teaching of human origins in schools. Public
Opinion Quarterly 72(3): 540–553.
Plutzer E, Branch G and Reid A (2020) Teaching evolution in U.S. public schools: a continuing challenge.
Evolution: Education and Outreach 13(4).
Putnam RD and Campbell DE (2010) American Grace: How Religion Divides and Unites Us. New York,
NY: Simon and Schuster.
Schank RC and RP Abelson (1977) Scripts, Plans, Goals, and Understanding: An Inquiry in Human
Knowledge Structures. Hillsdale, NJ: Lawrence Erlbaum.
Scott EC (2009) Evolution vs. Creationism: An Introduction. 2nd ed. Berkeley, CA: University of California
Press.
Shen BSJ (1975) Scientific literacy and the public understanding of science. In: Day S (ed.) Communication
of Scientific Information. Basel: Karger, pp. 44–52.
Shumow L and Miller JD (2001) Parents’ at-home and at-school academic involvement with adolescents.
Journal of Early Adolescence 21(1): 68–91.
Society for Integrative Comparative Biology (n.d.). Available at: https://sicb.org/rer/saawok.php3
Sternberg RJ and Ben-Zeev T (2001) Complex Cognition: The Psychology of Human Thought. New York:
Oxford University Press.
Swift A (2017) In U.S. belief in creationist view of humans at new low. Available at: https://news.gallup.com/
poll/210956/belief-creationist-view-humans-new-low.aspx
United States Census Bureau (2019) CPS historical time series. Table A-2. Available at: https://www.census.
gov/data/tables/time-series/demo/educational-attainment/cps-historical-time-series.html
van Dijk J (2020) The Network Society. 2nd ed. New York, NY: SAGE.
Wilcox C and Larson C (2006) Onward Christian Soldiers? The Religious Right in American Politics.
Boulder, CO: Westview Press.
Williams DK (2012) God’s Own Party: The Making of the Christian Right. New York, NY: Oxford University
Press.
Author biographies
Jon D. Miller is Director of the International Center for the Advancement of Scientific Literacy in the Institute
for Social Research at the University of Michigan, Ann Arbor. He has studied the public understanding of
science and technology and public engagement in the formulation of science and technology policy for the
last 50 years.
Eugenie C. Scott is the founding executive director of the National Center for Science Education, which pro-
motes and defends the teaching of evolution and climate change.
16 Public Understanding of Science 00(0)
Mark S. Ackerman is the George Herbert Mead Collegiate Professor of Human-Computer Interaction and a
Professor in the School of Information and in the Department of Electrical Engineering and Computer Science
at the University of Michigan, Ann Arbor. His major research area is social computing (also called CSCW).
Belén Laspra is an Assistant Professor at the Department of Philosophy of the University of Oviedo (Spain).
She is a member of the Social Studies of Science Research Team. Her current research interests are in the
meaning and measures of scientific culture.
Glenn Branch is deputy director of the National Center for Science Education. He received the Evolution
Education Award for 2020 from the National Association of Biology Teachers.
Carmelo Polino is an Assistant Professor in the Department of Philosophy at the University of Oviedo (Spain)
and a researcher at Centro Redes (Buenos Aires, Argentina). His research interests are public understanding
of science and sociology of science communication.
Jordan Huffaker is a PhD student in Computer Science and Engineering at the University of Michigan, Ann
Arbor. His area of research is social computing.
