No Missing Link: Knowledge Predicts Acceptance of Evolution in the United States

Abstract

Most Americans reject some or all parts of evolutionary theory, contrary to the scientific consensus. Americans' attitudes toward evolution at least partially depend on their religious or political values, and prior work has argued that knowledge of the theory plays a negligible role. But there have been no systematic, large-scale attempts to measure the public's knowledge of evolutionary theory, which means that claims about a lack of significant impact of evolution knowledge on evolution acceptance may be premature. Using a new demographically representative survey (N=1100) that includes a detailed measure of evolution knowledge, we find that knowledge predicts level of acceptance, even after accounting for the effects of religion and politics. These results demonstrate that Americans' views on evolution are significantly influenced by their knowledge about this theory and therefore might be amenable to change.

Full text

Thinking of Biology
212 BioScience •March 2018 / Vol. 68 No. 3 https://academic.oup.com/bioscience
No Missing Link: Knowledge Predicts
Acceptance of Evolution in the
United States
DEENA SKOLNICK WEISBERG, ASHELEY R. LANDRUM, S. EMLEN METZ, AND MICHAEL WEISBERG
Most Americans reject some or all parts of evolutionary theory, contrary to the scientific consensus. Americans’ attitudes toward evolution at
least partially depend on their religious or political values, and prior work has argued that knowledge of the theory plays a negligible role. But
there have been no systematic, large-scale attempts to measure the public’s knowledge of evolutionary theory, which means that claims about a
lack of significant impact of evolution knowledge on evolution acceptance may be premature. Using a new demographically representative survey
(N=1100) that includes a detailed measure of evolution knowledge, we find that knowledge predicts level of acceptance, even after accounting for
the effects of religion and politics. These results demonstrate that Americans’ views on evolution are significantly influenced by their knowledge
about this theory and therefore might be amenable to change.
Keywords: evolution, public understanding of science, acceptance, knowledge, public opinion poll
Americans have a fraught relationship with 
evolutionary theory. Despite widespread acceptance of
this theory in the scientific community (Funk and Rainie
2015), public-opinion surveys have demonstrated that 38%
of Americans identify as creationists (Swift 2017) and 52%
disagree that human beings developed from earlier species
of animals (National Science Board 2016). International
comparisons show that these views are extreme, particu-
larly among westernized countries: In a comparison of 34
countries, the United States ranks second to last in public
acceptance of evolution (Miller etal. 2006). This large-scale
rejection of the scientific consensus has troubling impli-
cations for science literacy and for public engagement in
science more generally.
The goal of the current study is to expand on work docu-
menting why ordinary Americans hold such views about
evolutionary theory. In particular, this study focuses on
the role of knowledge of evolutionary theory in supporting
acceptance of it.
Researchers examining the public perception of science
tend to argue that the public’s reported attitudes on evolution
are primarily expressions of their political or religious world-
views (Bishop and Anderson 1990, Kahan etal. 2011, Kahan
2016, Snow and Dibner 2016), a type of motivated reasoning
sometimes referred to as identity-protective cognition (Kunda
1990, Cohen et al. 2000). Rather than genuinely probing
individuals’ considered views, questions about evolution
acceptance trigger people’s identification with religious or
political groups, leading them to respond consistently with
their group’s values. These arguments are supported by work
finding that acceptance of evolution is either uncorrelated
or only weakly correlated with knowledge of it (Bishop and
Anderson 1990, Lawson and Worsnop 1992, Demastes etal.
1995, Sinatra etal. 2003, Ingram and Nelson 2006, Shtulman
2006, Nehm et al. 2009). Results such as these have led
many researchers in this area to reject the knowledge-deficit
hypothesis (see Suldovsky 2016), which suggests that evolu-
tion acceptance reflects level of knowledge.
However, other work shows that individuals with greater
knowledge of evolution are indeed more likely to accept
it (Lawson 1983, Johnson and Peeples 1987, Rutledge
and Warden 2000, McKeachie et al. 2002, Rutledge and
Mitchell 2002, Deniz etal. 2008, Nadelson and Sinatra 2009,
Shtulman and Calabi 2012), suggesting that these variables
are related. A major contribution of the current work, then,
is to clarify the relationship between knowledge and accep-
tance in the general population, a study that can inform not
only the literature on this issue but also educational and
political policy.
Crucially, to the best of our knowledge, studies that have
investigated the link between knowledge and acceptance
have either all recruited nonrepresentative samples—primar-
ily high school and college students—or have not included
an extensive or appropriately calibrated measurement of
BioScience 68: 212–222. © The Author(s) 2018. Published by Oxford University Press on behalf of the American Institute of Biological Sciences. All rights
reserved. For Permissions, please e-mail: journals.permissions@oup.com.
doi:10.1093/biosci/bix161 Advance Access publication 7 February 2018
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
https://academic.oup.com/bioscience March 2018 / Vol. 68 No. 3 •BioScience 213
the participants’ understanding of evolutionary theory. For
example, prior work has tended to test students’ knowledge
after they engaged in specific curricula focused on evolu-
tion (Demastes etal. 1995, Ingram and Nelson 2006); these
measures are too specifically focused to serve as assessments
for the general population. Therefore, the available evi-
dence does not allow us to draw conclusions about relations
between acceptance and knowledge in the population as a
whole, leaving open the question of how people’s reported
acceptance may be influenced by their knowledge. The pri-
mary goal of the current study was to address this issue (for
more details, see preregistration at https://osf.io/mvg4n).
Knowledge of evolution
To assess the potential influences of knowledge of evolu-
tionary theory on the acceptance of it, we developed a new
evolution knowledge battery specifically designed to capture
variance present in the general public. We administered this
test to a sample that matched the demographics of the over-
all American population, allowing us to test for the first time
whether evolution knowledge and acceptance are related
among members of the general public. If so, this would sug-
gest a need to include more than just identity factors in the
dialogue about evolution acceptance. Furthermore, such a
result could suggest avenues for interventions to increase
both knowledge and acceptance.
Our measure tests participants’ knowledge of several
of the key concepts of evolutionary theory (according to
National Research Council 2012), including inheritance,
variation, natural selection, and adaptation. Some of these
questions were based on previously published tests of
these concepts (e.g., Bishop and Anderson 1990, Anderson
etal. 2002, Blackwell etal. 2003, Shtulman 2006). In addi-
tion to establishing our scale’s face validity, we used item
response theory (IRT) to analyze pilot data collected from
Amazon’s Mechanical Turk system (MTurk) and from a
demographically representative phone survey conducted by
the Annenberg Public Policy Center. We used these analyses
to identify and refine or reject items with insufficient dis-
criminatory power. IRT analysis of the data from the current
study revealed that our items captured a wide range of diffi-
culty and discrimination, allowing us to discriminate among
the participants’ levels of knowledge across our demographi-
cally representative sample (see figure 1).
The final measure administered in the current survey
consisted of 17 multiple-choice or multiple-select items. All
of these items were prefaced with “according to scientists” or
“scientists would think,” which is a technique developed by
the National Science Foundation to ensure that participants
respond on the basis of their knowledge of the theory rather
than on their personal views about it (i.e., construct valid-
ity). In other words, this phrasing allows participants who
reject evolutionary theory to demonstrate knowledge about
it. Cronbach’s alpha for the scale was .77.
Scores on this task are the sum of the number of correct
responses (maximum of 26). Although most of the questions
are multiple choice, two of the questions have multiple cor-
rect answers that can be selected (i.e., the participants were
asked to “check all that apply”). We treated each item in
these questions independently, as if they were true or false
items, and counted both the correctly selected and correctly
not-selected items. For analysis, we created a score that
summed the number of correct answers.
Acceptance of evolution
In addition to administering this new measure of knowledge,
we also adjusted the way acceptance of evolutionary theory
was measured. As is noted in Miller and colleagues (2006),
small changes to the phrasing of the acceptance question
can lead to different response patterns. For example, simply
asking whether participants believe humans evolved or not
yields only 38% who deny evolution (PRRI 2011), whereas
asking participants to rate their agreement with a statement
such as “evolution is the best explanation for the origins
of human life on Earth” on a four-point scale yields 45%
disagreement (Pew 2007). Moreover, rephrasing the ques-
tion to ask about elephants instead of humans leads to only
25% disagreement with the evolutionist option instead of
52% (Maitland etal. 2014; see also Roos 2014). This latter
result in particular indicates that asking about acceptance of
human evolution is tapping into a different construct from
asking about acceptance of evolution in general. Extant
surveys thus may not fully capture Americans’ views about
evolutionary theory.
For the current study, we therefore developed a new mea-
sure of acceptance based on one of the most frequently cited
polls, Gallup. This poll asks which of three options comes
closest to respondents’ views on the origin and develop-
ment of human beings: (1) human beings have developed
over millions of years from less advanced forms of life, but
God guided this process; (2) human beings have developed
over millions of years from less advanced forms of life,
but God had no part in this process; and (3) God created
human beings pretty much in their current form in the last
10,000 years or so. Because this question offers three pos-
sible answers, it is better than most. However, these three
options do not allow for any nuance about God’s role in
evolution; the “guidance” could take many forms. We thus
added a fourth answer option, reflecting a deistic view: God
set up the laws of nature, which then unfolded on their own.
We also chose to ask about the origin of plants and animals
rather than of humans (acknowledging that “animals” in
colloquial usage is understood to exclude humans). Given
the issues identified above, we anticipated that our choice
of wording would affect the participants’ responses. For
example, our use of “plants and animals,” which does not
mention humans explicitly, might make the evolutionist
option more appealing in our survey than in others. On the
other hand, our removal of the reference to a young Earth
might make more people comfortable with choosing the
creationist option. Because any particular framing of the
acceptance question cannot yield a full view of Americans’
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
214 BioScience •March 2018 / Vol. 68 No. 3 https://academic.oup.com/bioscience
attitudes toward evolutionary theory, we chose the “plants
and animals” phrasing to avoid bringing to mind any poten-
tial personal consequences of accepting evolution (see Brem
etal. 2003) and to provide a more natural comparison with
the items in our knowledge battery, which focused on non-
human animals.
Specifically, we asked people to choose which of the
following options best described how they think animals
and plants came to exist on Earth: (a) animals and plants
were created by God in more or less their current form
(creationism); (b) animals and plants developed through
natural processes, which were guided by God the entire
time (theistic evolution); (c) animals and plants developed
through natural processes, which were set up by God but
continued on their own (deistic evolution); and (d) animals
and plants developed entirely through natural processes
(naturalistic evolution). Each level of acceptance varies the
extent of God’s involvement.
After answering this acceptance question, the participants
rated the extent to which various factors influenced their
views on this topic: the quality of the scientific evidence, the
theory’s conflict or consistency with their religious beliefs,
their education, and their family’s beliefs.
Other measures
In addition to assessing whether Americans’ knowledge
of evolutionary theory plays any role in their degree of
acceptance, we tested the impact of several other factors on
evolution acceptance, such as general science knowledge
and authoritarian tendencies. Prior work suggests that these
either directly predict acceptance (e.g., Lombrozo et al.
2008) or relate to religious or political beliefs, which in turn
predict acceptance (e.g., Jost etal. 2003).
Science knowledge and reasoning skills (OSI2). We included the
18-item ordinary science intelligence scale (OSI version
2; Kahan 2016), which measures general knowledge of
science and the scientific method (drawn from the National
Science Foundation Science and Engineering Indicators),
basic numeracy, and cognitive reflection (Frederick 2005).
Some of our participants had responded to these items in a
previous survey, so instead of repeating the test and possibly
biasing their responses, we used their responses from the
earlier survey.
Questions on this task are a combination of multiple choice
(e.g., “Which gas makes up most of the Earth’s atmosphere?”
with answer choices hydrogen, nitrogen, carbon dioxide, and
Figure 1. A histogram of our sample’s scores on the evolution knowledge scale (N = 1100).
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
https://academic.oup.com/bioscience March 2018 / Vol. 68 No. 3 •BioScience 215
oxygen) and open response (e.g., “Imagine that we roll a fair,
six-sided die 1000 times. Out of 1000 rolls, how many times
do you think the die would come up as an even number?”).
We constructed a scale by summing the correct scores
(maximum of 18).
Understanding the nature of theories. The participants saw
two items asking them to rate their agreement with state-
ments about the nature of scientific theories (adapted
from Lombrozo etal. 2008). The statements were, “Once a
scientific theory has been established, it is never changed”
and “Scientific theories are just scientists’ guesses.” Ratings
were made on a five-point scale (1, strongly agree; 3, unsure;
5, strongly disagree). We averaged together the participants’
responses to these two items. Final scores thus fell between
1 (strongly agree) and 5 (strongly disagree), with higher num-
bers indicating more correct responding.
Criteria for belief. The participants reported the extent to
which they think that two candidate criteria are good rea-
sons for belief. Specifically, the participants were asked to
consider the statements, “There is good scientific evidence
for it” and “I feel it is true in my gut” and to indicate whether
each is an excellent (coded 5), good (coded 4), okay (coded
3), bad (coded 2), or terrible (coded 1) reason to believe
something. These measures are based on prior work finding
that evolutionists and creationists accept different warrants
for belief (Metz et al. 2018). In our analyses, we considered
these two items separately because they measure two differ-
ent types of criteria that could be accepted independently.
Need for closure. We presented two subscales from the brief
need for closure scale (Roets and Van Hiel 2011). Three
items asked about tolerance of ambiguity (e.g., “I don’t like
situations that are uncertain”), and three items asked about
closed-mindedness (e.g., “I do not usually consult many dif-
ferent opinions before forming my view”). The participants
rated their agreement with these statements on a scale from 1
(completely disagree) to 6 (completely agree). Their scores for
these six items were averaged to create a single measure, with
higher numbers indicating higher need for closure.
Authoritarianism. We asked about 4 pairs of qualities that
children could have (e.g., independent versus respectful of
their elders). The participants chose which item in each pair
they thought was more important for children to display
(Feldman and Stenner 1997). The item that corresponded to
the authoritarian option was coded 1, and the other option
was coded 0. We took the average of how many authoritarian
choices each participant made, with scores ranging between
0 and 1.
Participants and procedures
We contracted with the survey firm YouGov to conduct a
nationally representative survey of Americans (N = 1100)
in July 2016. These individuals were drawn from standing
panels maintained by YouGov. The survey firm adminis-
tered online surveys to a total of 1268 respondents . YouGov
(blind to our hypotheses) pared down this sample to a final
sample of 1100 to match the demographic makeup of the
United States in terms of the dimensions of gender, age, race,
education, party identification, ideology, and political inter-
est, based on the 2010 American Community Survey and the
Current Population Survey.
Although our sample already matched the demographics
of the United States well, it was not possible for it to precisely
match the demographic breakdown of the entire US popula-
tion on every dimension of interest. As is standard in surveys
of this kind, YouGov also provided sample weights to obtain
an even closer fit between our data and the true demographic
makeup of the country. Following accepted practices in pub-
lic opinion research, we use the weighted data when report-
ing “top lines” or descriptive statistics (here, the percentages
of individuals who agreed with each acceptance category),
and we used the raw data for all inferential analyses.
In the survey, the participants completed seven measures
(see the supplemental material and https://osf.io/mvg4n):
evolution knowledge, evolution acceptance, basic scientific
literacy (OSI2), understanding the nature of scientific theo-
ries, criteria for belief, need for closure, and authoritarianism.
These measures were split into three blocks: the knowledge
block, which contained our evolution knowledge test and
the OSI2; the acceptance block, which contained only the
acceptance question and its accompanying questions about
influences on these views; and the opinion block, which con-
tained the remaining measures (nature of theories, criteria
for belief, need for closure, and authoritarianism).
The participants first saw the knowledge and opinion
blocks, order counterbalanced between participants. The
acceptance block was always presented at the end of the
survey to avoid priming participants’ responses to the other
measures by asking them to consider their personal beliefs
about evolutionary theory. Within the knowledge block, the
order of all of the questions from both of the measures was
randomized. Within the opinion block, we randomized the
order of each of the two questions about nature of theories,
each of the two questions about criteria for belief, the set of
need-for-closure questions, and the set of authoritarianism
questions. The order of the questions within the need-for-
closure measure and the authoritarianism measure was
randomized, but these questions were always presented as
a unit. Within the acceptance block, the order of questions
was fixed: The participants always responded to the main
acceptance question first and then the question about influ-
ences on their view.
Demographic variables. Because the participants were part of
an online panel, demographic data including political-party
affiliation, political ideology, religion, degree of religiosity,
age, and level of education were previously collected by
YouGov in a separate testing session and shared with us. We
relied on the three-question Pew Religious Life battery to
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
216 BioScience •March 2018 / Vol. 68 No. 3 https://academic.oup.com/bioscience
determine degree of religiosity, which more richly captures
the participants’ degree of religiosity rather than simply the
religion with which they identify. Although this is a standard
battery, we acknowledge that these three questions may not
fully capture all of the complex dimensions of religiosity (see
Cornwall etal. 1986). These three questions ask about atten-
dance at religious services aside from weddings and funerals
(more than once a week, once a week, once or twice a month,
a few times a year, seldom, and never), the importance of
religion (very important, somewhat important, not too impor-
tant, not at all important), and frequency of prayer (several
times a day, once a day, a few times a week, once a week, a few
times a month, seldom, or never). Because responses to these
three items were made on different scales, the scale for each
item was normalized, and then the normalized scores were
averaged and normalized again to yield a total religiosity
score. We divided our respondents into high (more than 1
standard deviation [SD] above the mean), average (between
1 SD above the mean and 1 SD below the mean), and low
religiosity (more than 1 SD below the mean).
Results
Our primary aims in this study were to determine Americans’
levels of acceptance of and knowledge about the theory of
evolution and to explore the relationship between these
factors. A full report of all of our preregistered hypotheses
and tests is available in the supplemental material and at our
OSF site; here, we focus only on the tests that address these
primary goals.
What are Americans’ views about evolutionar y theory? As we pre-
dicted, our acceptance question yielded strikingly different
results from those of Gallup, which asked about human evo-
lution and offered only three response options. Only 26% of
our respondents held creationist views (rather than Gallup’s
38%), whereas 32% accepted naturalistic evolution (rather
than Gallup’s 19%; figure 2).
In addition, as we predicted and as we expected from
prior work, each participant’s level of evolution acceptance
related to various identity factors. In regressions predicting
level of acceptance from each factor individually, the partici-
pants who had higher levels of religiosity were more likely
to reject evolution (β = –1.12, p < .001), as were those with
more conservative political ideology (β = –0.35, p < .001).
What do Americans know about evolutionary theory? We predicted
that people’s scores on our evolution knowledge battery would
be generally low. To test this prediction, we counted the total
number of participants who “failed” each test (i.e., scored less
than 60%, using a standard academic measure of failing) and
found that 68% of our participants fell below this threshold.
Although it is not possible to compare our test directly with
tests used in prior work because both the questions and the
sample populations are different, other studies of evolution-
ary knowledge find similarly low performance. For example,
about half of the students in Shtulman (2006) responded
incorrectly to questions about variation and inheritance, and
only 31% of the students studied by Bishop and Anderson
(1990) demonstrated understanding of topics in natural
selection.
Is there a relation between acceptance and knowledge? Our data
show meaningful, statistically significant differences in
understanding based on degree of acceptance (F(3) = 44.8,
p < .001, η2 = 0.11; figure 3). T-tests show that scores for
creationists (mean [M] = 11.8, SD = 3.4) and theistic evo-
lutionists (M = 12.3, SD = 4.3) did not differ significantly
(t(468) = 1.39, p = .17, Cohen’s d = 0.13). Theistic evolu-
tionists scored lower than deistic evolutionists (M = 14.3,
SD = 4.1, t(446) = 5.06, p < .001, Cohen’s d = 0.48), who in
turn scored lower than naturalistic evolutionists (M = 15.2,
SD = 4.5, t(628) = 2.3, p = .02, Cohen’s d = 0.19). A multino-
mial logistic regression predicting level of acceptance from
scores on the knowledge test shows that for each one-point
increase in knowledge score, individuals are 1.16 times more
likely to choose the deist option versus the creationist one
and 1.22 times more likely to choose the naturalistic evo-
lutionist option versus the creationist one. That is, as was
predicted, understanding evolutionary theory is related to
accepting it.
Figure 2. Views of the American public on the origins
of species from the Gallup poll (panel a) and our survey
(panel b).
creationist 38%
(a)
(b)
theistic 38%
naturalistic 19%
don’t know 5%
creationist 26%
theistic 18%
deistic 24%
naturalistic 32%
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
https://academic.oup.com/bioscience March 2018 / Vol. 68 No. 3 •BioScience 217
However, as we noted above, the tendency to accept evo-
lutionary theory is predicted by affiliations and prior beliefs,
such as religiosity. It is thus possible that those factors fully
explain levels of acceptance, leaving no role for under-
standing. To test this possibility, we conducted a logistic
regression predicting whether one leans evolutionist (i.e.,
accepts deistic or naturalistic evolution) from performance
on the knowledge battery, religiosity, having a conservative
ideology, age, gender, and level of education (table 1). This
test showed that the participants’ knowledge of evolution
remained predictive of their level of acceptance even after
taking into account the effect of demographic factors and
the interaction for knowledge and religion (although it is
important to note that the effect of religion loses significance
when the interaction term is included). Crucially, this model
fits the data significantly better than the same model that
does not include the participants’ knowledge scores (χ2(972)
= 51.31, p < .001), and this model explains more of the
variance (pseudo R2 = .43 compared with .38). This result
provides the first key piece of evidence that knowledge about
the theory of evolution plays a significant role in whether or
not members of the public accept this theory as the correct
explanation for the origin and development of life.
This result can be seen in figures 4 and 5, which show the
relationship between knowledge of evolution and level of
acceptance for three categories of religiosity (below average,
average, and above average) and five levels of political ideol-
ogy (very liberal, liberal, moderate, conservative, and very
conservative): Increasing knowledge is related to increasing
likelihood of acceptance of evolution. Crucially, this is the
case for each level of religiosity and each level of political
ideology, indicating that there is little polarizing effect of
increased knowledge.
What is the impact of the auxiliary measures? As was predicted, we
found positive relationships between acceptance of evolution
and performance on the OSI2 (β =0.18, p < .001), understand-
ing the nature of theories (β = 0.80, p < .001), acceptance of
scientific evidence as a criterion for belief (β = 0.66, p < .001),
and education (β = 0.22, p < .001). We also found negative
relationships between acceptance of evolution and political
conservatism (β = –0.35, p < .001), acceptance of gut feelings
as a criterion for belief (β = –0.25, p < .001), need for closure
(β = –0.29, p < .001), religiosity (β = –1.12, p < .001), and
authoritarianism (β = –1.68, p < .001). Contrary to our predic-
tions, there was no relationship with age (β = –0.002, p = .62).
0.000
0.025
0.050
0.075
0.100
0.125
510152
02
5
Knowledge Score
Population Density
Acc creationist theistic deistic naturalistic
Figure 3. The distribution of the participants’ scores on the evolution knowledge battery, split by their acceptance level.
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
218 BioScience •March 2018 / Vol. 68 No. 3 https://academic.oup.com/bioscience
predicting whether one leans evolutionist from performance
on the knowledge battery, performance on the OSI2, religi-
osity, having a conservative ideology, age, gender, and level
of education (table 2). We found that both domain-specific
evolution knowledge and OSI2 performance were indepen-
dently predictive of evolution acceptance and at approxi-
mately the same level (β = 0.082 and 0.087, respectively).
Table 1. The results of a logistic regression predicting level of acceptance from performance on the knowledge battery, religiosity,
having a conservative ideology, age, gender, and level of education.
BStandard error p value exp(B) 2.5% CI 97.5% CI
(Intercept) –2.389 0.471 <.001** 0.092 0.036 0.229
Knowledge 0.160 0.024 <.001** 1.174 1.121 1.232
Religiosity –0.226 0.336 .500 0.797 0.413 1.543
Conservative Ideology –0.152 0.076 .046* 0.859 0.740 0.997
Age 0.011 0.005 .031* 1.011 1.001 1.020
Female –0.046 0.161 .776 0.955 0.697 1.310
Education 0.226 0.059 <.001** 1.253 1.118 1.408
Knowledge x Religion –0.077 0.025 .002** 0.926 0.881 0.972
*p < .05. **p < .01.
Finally, we explored the contributions of domain-specific
knowledge about evolution (as was measured by our knowl-
edge battery) compared with those of domain-general
knowledge about science and scientific reasoning skills (as
was measured by the OSI2) to determine whether specific or
general knowledge might be more important to an individ-
ual’s level of acceptance. We conducted a logistic regression
0.00
0.25
0.50
0.75
1.00
510152
02
5
Number of Knowledge Questions Answered Correctly
Probability Leans Evolutionist
Religiosity
above
average
below
Figure 4. The participants’ probability of leaning evolutionist as a function of their scores on the evolution knowledge
battery, split by their degree of religiosity.
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
https://academic.oup.com/bioscience March 2018 / Vol. 68 No. 3 •BioScience 219
0.00
0.25
0.50
0.75
1.00
510152025
Number of Knowledge Questions Answered Correctly
Probability Leans Evolutionist
Ideology
very liberal
liberal
moderate
conservative
very conservative
Figure 5. The participants’ probability of leaning evolutionist as a function of their scores on the evolution knowledge
battery, split by their political ideology.
Conclusions
The current study was designed to gain a more accurate
view of how Americans think about evolutionary theory,
with the particular goal of investigating whether individuals’
knowledge about the theory plays a significant role in their
acceptance of it. To our knowledge, this is the first attempt
to explore these questions using a sample that reflects the
demographics of the entire country, as well as the first time
that such a sample has been tested on its knowledge of a
comprehensive set of topics within evolutionary theory. We
predicted that we would find a higher degree of acceptance
in our survey than in previous work and that knowledge
would predict acceptance. Both predictions were supported.
With respect to overall levels of acceptance, our data paint
a more optimistic picture of Americans’ attitudes toward
evolution than previous polls: Only a quarter of our respon-
dents were creationist, and more than half of our sample
accepted that evolution happened without specific divine
intervention. Although the current design does not allow
us to determine exactly why our results differ from those
of prior surveys, the fact that we found significantly fewer
creationists with these small wording changes suggests that
Americans may not be as resistant to evolutionary theory as
previously thought. In our future work, we intend to directly
compare levels of evolution acceptance when the question is
about humans as opposed to plants and animals, which will
provide greater insight into the role that perceived human
uniqueness plays in people’s beliefs.
In line with previous work in this area, our data show
that participants with high levels of religiosity and conser-
vative views are more likely to reject evolutionary theory.
On the basis of results such as these, some have argued that
people’s responses to evolution-acceptance questions can
be explained primarily by people expressing or protect-
ing their identities (“identity-protective cognition”; Kahan
et al. 2007). Importantly, however, we found that level of
acceptance is not solely a result of these identity factors;
knowledge of evolutionary theory also plays a significant
role. Even when taking into account the participants’ level
of religiosity and political conservatism, as well as their
age, gender, and level of education, performance on our
knowledge battery was still a significant predictor of their
level of acceptance. Crucially, and contrary to some similar
work on climate change (Kahan etal. 2012), we found a
relationship between increasing knowledge and increasing
acceptance at all levels of religiosity and political ideology.
We believe that we were able to find these relationships,
which have not been seen in previous work, because our test
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
220 BioScience •March 2018 / Vol. 68 No. 3 https://academic.oup.com/bioscience
was appropriately calibrated to capture the large variance
in knowledge in the general population, because we asked
about evolution in a context not involving humans, and
because we provided more options for categorizing people’s
views on evolution.
Indeed, our inclusion of four options in our acceptance
question allows us to construct a more nuanced view of
relations between Americans’ attitudes toward evolutionary
theory, their knowledge of the theory, and identity factors.
For example, highly religious participants and politically
conservative participants with high degrees of knowledge
tend to accept evolution, but they believe that God was the
prime mover. Specifically, for those with a high degree of
understanding of evolution (1 SD above the mean), 79% of
the most religious subjects and 41% of the political conser-
vatives are deistic evolutionists, as opposed to 24% of the
total population.
Our inclusion of several auxiliary measures can also
begin to tease out reasons for individuals’ stated level of
acceptance. The participants were more likely to accept
evolution if they scored higher on a test of general science
knowledge and reasoning skills, if they understood to a
greater extent the role of theories in science, and if they
believed that “it feels true in my gut” was a poor criterion
for belief. The participants were more likely to reject evolu-
tion if they reported higher levels of authoritarianism and
need for closure.
Because this study is correlational, it does not provide
definitive evidence about potential causal links among
these factors: Greater knowledge of evolution might lead
to views that more closely resemble the scientific con-
sensus, but it is also possible that individuals who accept
evolutionary theory are more likely to learn about it. Some
evidence for the former direction of causation comes
from interventions that aim to increase knowledge, which
have shown that developing a greater understanding of
evolution in the classroom can increase acceptance, at
least among college students (Lawson and Weser 1990,
Matthews 2001, Ingram and Nelson 2006, Shtulman and
Calabi 2012—but see Bishop and Anderson 1990, Lawson
and Worsnop 1992). Similar effects have been reported
for a brief educational intervention about climate change,
another publicly controversial scientific topic (Ranney and
Clark 2016). Optimistically, this suggests sustained atten-
tion to education might be able to improve the public’s
attitudes about evolutionary theory.
Although our data cannot currently determine whether
increasing knowledge may increase acceptance, we can offer
some suggestions on how to proceed should this be the case.
First and foremost, we believe that any efforts to bolster
people’s knowledge should be targeted not only toward
improving general knowledge about evolutionary theory but
also toward teaching a more realistic view of modern scien-
tific methods and how science creates knowledge. Research
in our labs and elsewhere suggests that the kind of scientific
and ecological literacy required for accepting evolutionary
theory crucially involves an appreciation of the nature of
scientific research (Lombrozo etal. 2006, 2008). In addition,
we believe that teaching evolution should not be the sole
province of the formal education system, especially given
that even biology teachers have many misconceptions about
this topic (Rutledge and Mitchell 2002, Cotner etal. 2016).
Informal educational experiences, such as television shows
or science museums, could help to bridge this gap by pro-
viding enriching and effective tools for conveying topics in
evolutionary theory. Indeed, recent work in developmental
psychology reveals that young children can learn the prin-
ciples of natural selection from storybooks (Kelemen etal.
2014, Shtulman etal. 2016). Finally, whether based in formal
or informal educational settings, any such effort should be
sensitive to our best science of science communication, aim-
ing to improve understanding without entrenching existing
biases (Jamieson 2017).
These recommendations underscore the fact that many
factors can play a role in shaping individuals’ particular
views on evolutionary theory. The most important con-
tribution of our data is to emphasize that a crucial part
of this complexity is due to individuals’ knowledge of
Table 2. The results of a logistic regression predicting level of acceptance from performance on the knowledge battery,
performance on the OSI2, religiosity, having a conservative ideology, age, gender, and level of education.
B SE p value exp(B) 2.5% CI 97.5% CI
(Intercept) –2.054 0.461 <.001** 0.128 0.051 0.313
Knowledge 0.082 0.027 .002** 1.086 1.030 1.145
OSI2 0.087 0.030 .003** 1.091 1.030 1.157
Religiosity –1.236 0.102 <.001** 0.291 0.237 0.353
Conservative ideology –0.199 0.077 .010** 0.820 0.705 0.953
Age 0.011 0.005 .024* 1.011 1.001 1.021
Female –0.064 0.164 .697 1.066 0.773 1.474
Education 0.199 0.060 <.001** 1.220 1.086 1.373
*p < .05. **p < .01.
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
https://academic.oup.com/bioscience March 2018 / Vol. 68 No. 3 •BioScience 221
evolution, the importance of which has previously been
underestimated.
Acknowledgments
Funding for this project was provided by the National
Science Foundation (STS 1455425 to MW and DSW). We
wish to thank Elysia Choi, Sabrina Elkassas, Dan Kahan,
Kelly Kennedy, Karen Kovaka, Jane Reznik, and all of
the members of the Penn Laboratory for Understanding
Science. The Annenberg Public Policy Center generously
provided support for piloting some of our survey questions
with a demographically representative phone sample.
Supplemental material
Supplementary data are available at BIOSCI online.
References cited
Anderson DL, Fisher KM, Norman GJ. 2002. Development and evaluation
of the conceptual inventory of natural selection. Journal of Research in
Science Teaching 39: 952–978. doi:10.1002/tea.10053
Bishop BA, Anderson CW. 1990. Student conceptions of natural selection
and its role in evolution. Journal of Research in Science Teaching 27:
415–427. doi:10.1002/tea.3660270503
Blackwell WH, Powell MJ, Dukes GH. 2003. The problem of student accep-
tance of evolution. Journal of Biological Education 37: 58–67. doi:10.10
80/00219266.2003.9655852
Brem SK, Ranney MA, Schindel J. 2003. Perceived consequences of evolu-
tion: College students perceive negative personal and social impact
in evolutionary theory. Science Education 87: 181–206. doi:10.1002/
sce.10105
Cohen GL, Aronson J, Steele CM. 2000. When beliefs yield to evidence:
Reducing biased evaluation by affirming the self. Personality and Social
Psychology Bulletin 26: 1151–1164. doi:10.1177/01461672002611011
Cornwall M, Albrecht SL, Cunningham PH, Pitcher, BL. 1986. The dimen-
sions of religiosity: A conceptual model with an empirical test. Review
of Religious Research 27: 226–244. doi:10.2307/3511418
Cotner S, Graczyk H, Rodríguez Garcia JL, Moore R. 2016. In Galápagos…
and uncomfortable with evolution. Journal of Biological Education 50:
115–119. doi:10.1080/00219266.2016.1175758
Demastes SS, Settlage J, Good R. 1995. Students’ conceptions of natural
selection and its role in evolution: Cases of replication and comparison.
Journal of Research in Science Teaching 32: 535–550. doi:10.1002/
tea.3660320509
Deniz H, Donnelly LA, Yilmaz I. 2008. Exploring the factors related to
acceptance of evolutionary theory among Turkish preservice biology
teachers: Toward a more informative conceptual ecology for bio-
logical evolution. Journal of Research in Science Teaching 45: 420–443.
doi:10.1002/tea.20223
Feldman S, Stenner K. 1997. Perceived threat and authoritarianism. Political
Psychology 18: 741–770. doi:10.1111/0162-895X.00077
Frederick S. 2005. Cognitive reflection and decision making. Journal of
Economic Perspectives 19: 25–42. doi:10.1257/089533005775196732
Funk C, Rainie L. 2015. Public and scientists’ views on science and soci-
ety. Pew Research Center. (18 December 2017; www.pewinternet.
org/2015/01/29/public-and-scientists-views-on-science-and-society)
Ingram EL, Nelson CE. 2006. Relationship between achievement and students’
acceptance of evolution or creation in an upper-level evolution course.
Journal of Research in Science Teaching 43: 7–24. doi:10.1002/tea.20093
Jamieson KH. 2017. The need for a science of science communication:
Communicating science’s values and norms. Pages 15–23 in Jamieson
KH, Kahan DM, Scheufele DA, eds. Oxford Handbook of the Science of
Science Communication. Oxford University Press.
Johnson RL, Peeples EE. 1987. The role of scientific understanding in college:
Student acceptance of evolution. American Biology Teacher 49: 93–98.
Jost JT, Glaser J, Kruglanski AW, Sulloway FJ. 2003. Political conservatism
as motivated social cognition. Psychological Bulletin 129: 339–375.
doi:10.1037/0033-2909.129.3.339
Kahan DM. 2016. “Ordinary science intelligence”: A science-compre-
hension measure for study of risk and science communication, with
notes on evolution and climate change. Journal of Risk Research 20:
995–1016. doi:10.1080/13669877.2016.1148067
Kahan DM, Braman D, Gastil J, Slovic P, Mertz CK. 2007. Culture
and identity-protective cognition: Explaining the white-male effect
in risk perception. Journal of Empirical Legal Studies 4: 465–505.
doi:10.1111/j.1740-1461.2007.00097.x
Kahan DM, Jenkins‐Smith H, Braman D. 2011. Cultural cognition of scien-
tific consensus. Journal of Risk Research 14: 147–174. doi:10.1080/136
69877.2010.511246
Kahan DM, Peters E, Wittlin M, Slovic P, Ouellette LL, Braman D, Mandel
G. 2012. The polarizing impact of science literacy and numeracy on
perceived climate change risks. Nature Climate Change 2: 732–735.
doi:10.1038/nclimate1547
Kelemen D, Emmons NA, Seston Schillaci R, Ganea PA. 2014. Young
children can be taught basic natural selection using a picture-
storybook intervention. Psychological Science 25: 893–902.
doi:10.1177/0956797613516009
Kunda Z. 1990. The case for motivated reasoning. Psychological Bulletin
108: 480–498. doi:10.1037/0033-2909.108.3.480
Lawson AE. 1983. Predicting science achievement: The role of develop-
mental level, disembedding ability, mental capacity, prior knowledge,
and beliefs. Journal of Research in Science Teaching 20: 117–129.
doi:10.1002/tea.3660200204
Lawson AE, Weser J. 1990. The rejection of nonscientific beliefs about
life: Effects of instruction and reasoning skills. Journal of Research in
Science Teaching 27: 589–606.
Lawson AE, Worsnop WA. 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: 143–166. doi:10.1002/tea.3660290205
Lombrozo T, Thanukos A, Weisberg M. 2008. The importance of under-
standing the nature of science for accepting evolution. Evolution:
Education and Outreach 1: 290–298. doi:10.1007/s12052-008-0061-8
Maitland A, Tourangeau R, Yan Y, Bell R, Muhlberger P. 2014. The effect
of question wording on measurement of knowledge about evolution:
An examination of survey experiment data collected for the National
Center for Science and Engineering Statistics. National Center for
Science and Engineering Statistics. Washington, DC.
Matthews D. 2001. Effect of a curriculum containing creation stories on
attitudes about evolution. American Biology Teacher 63: 404–409.
doi:10.1662/0002-7685(2001)063[0404:EOACCC]2.0.CO;2
McKeachie WJ, Lin Y-G, Strayer J. 2002. Creationist vs. evolutionary beliefs:
Effects on learning biology. American Biology Teacher 64: 189–192.
doi:10.1662/0002-7685(2002)064[0189:CVEBEO]2.0.CO;2
Metz SE, Weisberg DS, Weisberg M. 2018. Non-scientific epistemic criteria
sustain counter-scientific beliefs. Cognitive Science. Manuscript in press.
Miller JD, Scott EC, Okamoto S. 2006. Public acceptance of evolution.
Science 313: 765–766. doi:10.1126/science.1126746
Nadelson LS, Sinatra GM. 2009. Educational professionals’ knowledge and
acceptance of evolution. Evolutionary Psychology 7: 490–516.
National Research Council. 2012. A Framework for K–12 Science Education:
Practices, Crosscutting Concepts, and Core Ideas. National Academies Press.
National Science Board. 2016. Science and Engineering Indicators. National
Science Foundation.
Nehm RH, Kim SY, Sheppard K. 2009. Academic preparation in biology and
advocacy for teaching evolution: Biology versus non-biology teachers.
Science Education 93: 1122–1146. doi:10.1002/sce.20340
[Pew] Pew Forum on Religion and Public Life. 2007. Pew Research Center
for the People and the Press Poll. Pew Research Center.
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018

Thinking of Biology
222 BioScience •March 2018 / Vol. 68 No. 3 https://academic.oup.com/bioscience
[PRRI] Public Religion Research Institute. 2011. PRRI/RNS Religion News
Survey, September. Association of Religion Data Archives. (18 December
2017; www.thearda.com/Archive/Files/Descriptions/PRRIRNSP.asp)
Ranney MA, Clark D. 2016. Climate change conceptual change: Scientific
information can transform attitudes. Topics in Cognitive Science 8:
49–75. doi:10.1111/tops.12187
Roets A, Van Hiel A. 2011. Item selection and validation of a brief, 15-item
version of the need for closure scale. Personality and Individual
Differences 50: 90–94. doi:10.1016/j.paid.2010.09.004
Roos JM. 2014. Measuring science or religion? A measurement anal-
ysis of the National Science Foundation sponsored science liter-
acy scale 2006–2010. Public Understanding of Science 23: 797–813.
doi:10.1177/0963662512464318
Rutledge ML, Mitchell MA. 2002. High school biology teachers’ knowledge
structure, acceptance and teaching of evolution. American Biology
Teacher 64: 21–28. doi:10.1662/0002-7685(2002)064
Rutledge ML, Warden MA. 2000. Evolutionary theory, the nature of sci-
ence and high school biology teachers: Critical relationships. American
Biology Teacher 62: 23–31.
Shtulman A. 2006. Qualitative differences between naïve and scientific
theories of evolution. Cognitive Psychology 52: 170–194. doi:10.1016/j.
cogpsych.2005.10.001
Shtulman A, Calabi P. 2012. Cognitive constraints on the understand-
ing and acceptance of evolution. Pages 47–65 in Rosengren KS, ed.
Evolution Challenges: Integrating Research and Practice in Teaching
and Learning about Evolution. Oxford University Press.
Shtulman A, Neal C, Lindquist G. 2016. Children’s ability to learn evolu-
tionary explanations for biological adaptation. Early Education and
Development 27: 1222–1236. doi:10.1080/10409289.2016.1154418
Sinatra GM, Southerland SA, McConaughy F, Demastes JW. 2003.
Intentions and beliefs in students’ understanding and acceptance of bio-
logical evolution. Journal of Research in Science Teaching 40: 510–528.
doi:10.1002/tea.10087
Snow CE, Dibner KA, eds. 2016. Science Literacy: Concepts, Contexts, and
Consequences. National Academies Press. doi:10.17226/23595
Suldovsky B. 2016. In science communication, why does the
idea of the public deficit always return? Exploring key influ-
ences. Public Understanding of Science 25: 415–426.
doi:10.1177/0963662516629750
Swift A. 2017. In US, belief in creationist view of humans at new low.
Gallup News. (18 December 2017; www.gallup.com/poll/210956/belief-
creationist-view-humans-new-low.aspx)
Deena Skolnick Weisberg is affiliated with the Department of Psychology
and the Annenberg Public Policy Center at the University of Pennsylvania,
in Philadelphia. Asheley R. Landrum is with the College of Media and
Communication at Texas Tech University, in Lubbock. S. Emlen Metz is
affiliated with the Department of Physics at the University of California,
Berkeley. Michael Weisberg (weisberg@phil.upenn.edu) is affiliated with the
Department of Philosophy and the Annenberg Public Policy Center at the
University of Pennsylvania, in Philadelphia.
Downloaded from https://academic.oup.com/bioscience/article-abstract/68/3/212/4834949
by University of Pennsylvania Libraries user
on 21 March 2018