Tensions Between Science and Intuition Across the Lifespan

Abstract

The scientific knowledge needed to engage with policy issues like climate change, vaccination, and stem cell research often conflicts with our intuitive theories of the world. How resilient are our intuitive theories in the face of contradictory scientific knowledge? Here, we present evidence that intuitive theories in 10 domains of knowledge-astronomy, evolution, fractions, genetics, germs, matter, mechanics, physiology, thermodynamics, and waves-persist more than four decades beyond the acquisition of a mutually exclusive scientific theory. Participants (104 younger adults, Mage = 19.6, and 48 older adults, Mage = 65.1) were asked to verify two types of scientific statements as quickly as possible: those that are consistent with intuition (e.g., "the moon revolves around the Earth") and those that involve the same conceptual relations but are inconsistent with intuition (e.g., "the Earth revolves around the sun"). Older adults were as accurate as younger adults at verifying both types of statements, but the lag in response times between intuition-consistent and intuition-inconsistent statements was significantly larger for older adults than for younger adults. This lag persisted even among professional scientists. Overall, these results suggest that the scientific literacy needed to engage with topics of global importance may be constrained by patterns of reasoning that emerge in childhood but persist long thereafter.

Full text

Topics in Cognitive Science 8 (2016) 118–137
Copyright ©2015 Cognitive Science Society, Inc. All rights reserved.
ISSN:1756-8757 print / 1756-8765 online
DOI: 10.1111/tops.12174
Tensions Between Science and Intuition
Across the Lifespan
Andrew Shtulman, Kelsey Harrington
Department of Psychology, Occidental College
Received 30 June 2014; received in revised form 11 February 2015; accepted 20 February 2015
Abstract
The scientific knowledge needed to engage with policy issues like climate change, vaccination,
and stem cell research often conflicts with our intuitive theories of the world. How resilient are
our intuitive theories in the face of contradictory scientific knowledge? Here, we present evidence
that intuitive theories in 10 domains of knowledge—astronomy, evolution, fractions, genetics,
germs, matter, mechanics, physiology, thermodynamics, and waves—persist more than four dec-
ades beyond the acquisition of a mutually exclusive scientific theory. Participants (104 younger
adults, M
age
=19.6, and 48 older adults, M
age
=65.1) were asked to verify two types of scientific
statements as quickly as possible: those that are consistent with intuition (e.g., “the moon revolves
around the Earth”) and those that involve the same conceptual relations but are inconsistent with
intuition (e.g., “the Earth revolves around the sun”). Older adults were as accurate as younger
adults at verifying both types of statements, but the lag in response times between intuition-consis-
tent and intuition-inconsistent statements was significantly larger for older adults than for younger
adults. This lag persisted even among professional scientists. Overall, these results suggest that the
scientific literacy needed to engage with topics of global importance may be constrained by
patterns of reasoning that emerge in childhood but persist long thereafter.
Keywords: Conceptual development; Explanatory coexistence; Na€
ıve theories; Scientific
knowledge; Science education; Speeded reasoning
1. Introduction
Being an informed citizen in today’s highly technological world requires a fair amount
of scientific literacy. Engaging with debates on how to curtail climate change, for
instance, requires an understanding of geological systems and how those systems are
Correspondence should be sent to Andrew Shtulman, Department of Psychology, Occidental College,
1600 Campus Road, Los Angeles, CA 90041. E-mail: shtulman@oxy.edu

affected by human carbon emissions. Engaging with debates on how to regulate antibi-
otics requires an understanding of immune systems and how those systems are affected
by increased selection pressure on bacteria. And engaging with debates on how to man-
age invasive species requires an understanding of ecological systems and how those sys-
tems are affected by increased competition for resources. Many other issues of public
import—cloning, vaccination, pesticides, organ donation, nuclear power, genetically mod-
ified foods—require high levels of scientific knowledge as well.
Does the average adult possess that knowledge? Three decades of research in cognitive
development and science education suggest not. Researchers in these fields have discov-
ered that learning science is a two-fold process: Students must not only learn unfamiliar
concepts absent from everyday discourse, but they must also un-learn concepts acquired
earlier in development for making sense of those same phenomena. In other words, stu-
dents enter the science classroom with rich, pre-instructional theories—termed “folk theo-
ries,” “na€
ıve theories,” or “intuitive theories”—that typically interfere with learning a
more accurate, scientific theory of the same domain (Carey, 2009; Vosniadou, 1994). For
instance, students charged with learning a kinetic theory of heat must un-learn a sub-
stance-based theory in which heat is construed as an immaterial substance that flows in
and out of objects and can be trapped or contained (Reiner, Slotta, Chi, & Resnick,
2000). Students charged with learning a selection-based theory of evolution must un-learn
a need-based theory in which evolution is construed as a process that guarantees organ-
isms the traits they need in order to survive (Shtulman, 2006). And students charged with
learning an inertial theory of mechanics must un-learn an “impetus”-based theory in
which objects are assumed to move only if imparted an internal force, or impetus, and
will remain in motion until that impetus dissipates (McCloskey, 1983).
Further complicating matters, recent research suggests that intuitive theories not only
interfere with the acquisition of scientific theories but may also interfere with the opera-
tion of those theories many years past their acquisition. Recent research has shown that
adults exhibit cognitive conflict when retrieving scientific information that contradicts the
intuitive theories they had presumably abandoned as children. One of the best studied
cases of this phenomenon is conceptual development in the domain of living things.
Beginning with Piaget (1929), psychologists have long observed that young children con-
flate life with animacy. Not only do young children attribute life to animate, yet non-
living, entities like the sun and the wind, but they also deny life to living, yet seemingly
inanimate, entities like flowers and trees. By age 8, this pattern of attributions is typically
replaced by a more biologically informed pattern in which life is now identified with
metabolic processes (e.g., eating, breathing, growing) rather than mobility (Hatano & Ina-
gaki, 1994). Nevertheless, the adult-like pattern gives way to the child-like pattern when
adults are tested under speeded conditions (Babai, Sekal, & Stavy, 2010; Goldberg &
Thompson-Schill, 2009). That is, when adults asked to make living/non-living judgments
as quickly as possible, they make those judgments more slowly and less accurately for
plants (e.g., orchids, elms) as compared to animals (e.g., pigs, sharks) and for animate
non-living entities (e.g., comets, rivers) as compared to inanimate ones (e.g., brooms,
towels).
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 119

Animistic intuitions reemerge not only when adults are placed under speeded condi-
tions, but also when they sustain permanent cognitive impairments, like those produced
by Alzheimer’s disease. Zaitchik and Solomon (2008) asked Alzheimer’s patients what it
means for something to be alive and found that these patients were more likely to cite
motion as a prerequisite for life than to cite any truly biological properties (e.g., eating,
breathing, growing). Healthy age-matched controls, on the other hand, tended to cite only
the latter. When Alzheimer’s patients were asked to provide examples of living things,
they almost always mentioned animals but rarely mentioned plants; age-matched controls
tended to mention both animals and plants. And when Alzheimer’s patients were asked to
judge the life status of entities presented to them, they tended to err in the same ways as
children, judging the sun and the wind as alive, but judging flowers and trees as not alive;
age-matched controls continued to provide a biologically informed pattern of judgments.
Errors by Alzheimer’s patients do not appear to derive from a more general deficit in
cognitive functioning, as these patients show no impairments on tasks that impose the
same information-processing demands but assess physical reasoning rather than biological
reasoning (Zaitchik & Solomon, 2009).
Similar findings have been documented in the domain of teleology, or the perception
of design in nature. Previous research has shown that children are more “promiscuous”
with their teleological explanations than adults are (Kelemen, 1999). Although both chil-
dren and adults will provide teleological explanations for human artifacts (e.g., “pencils
are for writing”) and for biological parts (e.g., “ears are for hearing”), only children will
provide teleological explanations for whole organisms (e.g., “birds are for flying”) and
for naturally occurring events (e.g., “it rains so that flowers can drink”). Children become
more selective in their teleological explanations by early adolescence, but that selectivity
is tenuous. When college-educated adults are asked to judge the acceptability of teleologi-
cal explanations under speeded conditions, they tend to accept unwarranted explanations,
like “ferns grow in forests because they provide ground shade” and “the sun radiates heat
because warmth nurtures life,” which they tend not to accept under un-speeded conditions
and presumably have not accepted under such conditions for many years (Kelemen, Rott-
man, & Seston, 2013).
Furthermore, just as Alzheimer’s patients endorse animistic conceptions of life under
normal (non-speeded) conditions, they endorse teleological conceptions of nature under
normal conditions as well. Lombrozo, Kelemen, and Zaitchik (2007) provided Alzheimer’s
patients with both mechanistic and teleological explanations for a variety of natural phe-
nomena, some of which warranted a teleological explanation (e.g., eyes exist “so that peo-
ple and animals can see”) and some of which did not (e.g., rain exists “so that plants and
animals have water for drinking and growing”). Compared to healthy elderly adults, Alzhei-
mer’s patients were more likely to judge unwarranted teleological explanations as accept-
able. They were also more likely to judge those explanations as preferable to mechanistic
ones. These findings suggest that teleology, like animism, is a deep-seated form of intuition
that can be suppressed by a more scientific worldview but cannot be eradicated altogether.
Tensions between science and intuition have been documented not only at the level of
behavior, but also at the level of the brain. Dunbar, Fugelsang, and Stein (2007) used
120 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

fMRI to determine whether college-educated adults exhibit different patterns of brain
activity when watching motion displays that were either consistent or inconsistent with
the laws of physics. The physics-consistent displays depicted two balls of unequal size
falling to the ground at the same rate; the physics-inconsistent displays depicted the larger
ball falling to the ground more quickly than the smaller ball. Dunbar et al. found that,
among participants who judged the physics-consistent displays as natural and the physics-
inconsistent displays as unnatural, watching those displays increased activation in the
anterior cingulate cortex, an area of the brain associated with error detection and conflict
monitoring. That is, participants who exhibited no behavioral evidence of holding the
misconception “heavier objects fall faster than lighter ones” still exhibited neural evi-
dence of holding that misconception insofar that their brains appeared to be detecting and
inhibiting contradictory beliefs. Similar results have been documented in the domain of
electricity: Physics experts show increased activation in the anterior cingulate cortex,
among other areas associated with conflict monitoring, when evaluating electric circuits
that are intuitively correct but physically impossible (Masson, Potvin, Riopel, & Foisy,
2014; Potvin, Turmel, & Masson, 2014).
The studies reviewed thus far tracked the persistence of a single misconception—that
is, the misconception that life is synonymous with animacy, the misconception that
everything in nature exists for a purpose, and the misconception that heavier objects fall
faster than lighter objects. The focus on one, and only one, misconception was necessi-
tated by the type of judgment participants were asked to make, such as a living/non-liv-
ing judgment (Goldberg & Thompson-Schill, 2009) or a warranted/unwarranted
judgment (Kelemen et al., 2013). Shtulman and Valcarcel (2012) extended this para-
digm to many more misconceptions (50 in total) by asking participants to make true/
false judgments instead. More specifically, they asked participants to verify, as quickly
as possible, two types of statements: statements whose truth value is the same on both
intuitive and scientific theories of a domain (e.g., “the moon revolves around the Earth,”
“genes that code for eye color can be found in the eye”) and statements involving the
same predicates but whose truth value differs across intuitive and scientific theories
(e.g., “the Earth revolves around the sun,” “genes that code for eye color can be found
in the liver”). The logic behind this design is that if intuitive theories survive the acqui-
sition of a mutually incompatible scientific theory, then the latter type of statement
should cause greater cognitive conflict than the former, resulting in (a) slower verifica-
tions and (b) less accurate verifications.
Using this method, Shtulman and Valcarcel (2012) documented evidence of long-term
conflict between science and intuition in 10 domains of knowledge: astronomy, evolution,
fractions, genetics, germs, matter, mechanics, physiology, thermodynamics, and waves.
What is most notable about these findings is their robustness. Shtulman and Valcarcel
probed for conflict between science and intuition with respect to five concepts in each of
10 domains and observed such conflict for the vast majority of them (86%). They also
observed conflict both for statements that are scientifically true but intuitively false (e.g.,
“air is composed of matter,” “humans are descended from sea-dwelling creatures”) and
for statements that are scientifically false but intuitively true (e.g., “fire is composed of
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 121

matter,” “humans are descended from chimpanzees”), indicating that conflict arose both
for statements that underextended scientific principles (e.g., failing to classify air as mat-
ter) and for statements that overextended scientific principles (e.g., classifying fire as mat-
ter). Furthermore, the participants in Shtulman and Valcarcel’s study had taken more
science courses than the average American—around three college-level courses, plus 4 to
6 years of middle and high school courses—and virtually all showed the effect. The
robustness of this phenomenon across domains, concepts, statements, and participants
suggests that it reflects more than just a handful of stubborn misconceptions. Rather, it
appears to reflect a fundamental property of science learning, namely, that intuition can
be overridden but not overwritten.
Here, we adopt Shtulman and Valcarcel’s (2012) method to assess the resilience of
intuitive theories across two dimensions that were held constant in the original study:
age and expertise. Whereas participants in the original study were college undergraduates
of approximately the same age (18–22) and with approximately the same amount of
science expertise, participants in the present study were adults with at least 40 additional
years of life experience (M
age
=65.1), a quarter of whom were professional scientists.
The speed and accuracy with which older adults verified scientific statements were com-
pared to the speed and accuracy with which younger adults verified the same statements
to determine whether conflict between science and intuition diminishes with age and/or
education.
There are at least two reasons to expect that older adults should not show the effect of
interest (i.e., a lag in response times between intuition-consistent and intuition-inconsis-
tent statements). First, older adults would have learned the relevant scientific theories
much earlier in life, affording more time for their intuitive theories to fade in strength or
relevance. Second, older adults would have had considerably more opportunity to use
their scientific knowledge outside the classroom, allowing for greater integration and con-
solidation of that knowledge with preexisting beliefs. The latter consideration applies
even more forcefully to professional scientists, who would not only have consolidated
their knowledge, but would also have consulted that knowledge on a near-daily basis. To
be certain, scientists are more proficient than non-scientists at using scientific knowledge
to encode domain-relevant information (Feil & Mestre, 2010), solve domain-relevant
problems (Chi, Feltovich, & Glaser, 1981), and make domain-relevant decisions (Shan-
teau, 1992). The question under investigation, however, is whether scientists are more
efficient at retrieving that knowledge when it conflicts with an earlier-acquired intuitive
theory. The results presented below suggest they are not.
2. Method
2.1. Participants
Two groups of participants were recruited for this study: 104 younger adults, recruited
from introductory psychology courses at Occidental College and compensated with extra
122 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

credit, and 48 older adults, recruited from either the local community (n=27) or the
faculty at Occidental College (n=21) and compensated monetarily. The younger adults
reported having taken an average of 3.6 college-level math and science courses prior to
the study (SD =3.3, range =0–15) and averaged 19.6 years in age (SD =1.2,
range =18–22). The older adults reported having taken an average of 6.5 college-level
math and science courses (SD =5.9, range =0–21) and averaged 65.1 years in age
(SD =11.1, range =50–87). Approximately half of the older adults were women, and
approximately a quarter of the younger adults were women. Preliminary analyses revealed
that gender was not associated with either response accuracy or response latency and was
not therefore included as a variable in subsequent analyses.
All older adults were, to our knowledge, neurologically healthy. Those recruited from
the local community occupied a variety of professions—accountant, author, writer, thera-
pist—whereas those recruited from the Occidental College faculty were either humanities
professors (n=11) or science professors (n=10). The humanities professors came from
the departments of Asian Studies, Economics, English, History, Philosophy, and Spanish;
the science professors came from the departments of Biology, Chemistry, Geology, Kine-
siology, Psychology, and Physics. Not surprisingly, science professors reported having
taken significantly more college-level science and math courses than did humanities pro-
fessors (M=15.9 vs. M=4.2, t(19) =8.47, p<.001). Humanities professors, on the
other hand, reported having taken an equivalent number to the non-professors (M=4.2
vs. M=3.9, t(36) <1). All three groups were of approximately the same age: science
professors, M=64.7, SD =4.5; humanities professors, M=58.7, SD =11.9; non-
professors, M=67.9, SD =11.7; F(2, 45) =2.85, ns.
2.2. Materials
Participants verified, as quickly as possible, 200 statements about natural phenomena:
20 statements in each of 10 domains of knowledge, with each statement exemplifying
one of five concepts within that domain. All domains and concepts are displayed in
Table 1
The five concepts covered in each domain
Domain Concept
Astronomy Planet, star, solar system, lunar phase, season
Evolution Common ancestry, phylogeny, variation, selection, adaptation
Fractions Addition, division, conversion, ordering, infinite density
Genetics Heritability, chromosome, dominance, gene expression, mutation
Germs Contagion, contamination, infection, sterilization, microbe
Matter Mass, weight, density, divisibility, atom
Mechanics Force, velocity, acceleration, momentum, gravity
Physiology Life, death, reproduction, metabolism, kinship
Thermodynamics Heat, heat source, heat transfer, thermal expansion, temperature
Waves Light, color, sound, wave propagation, reflection
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 123

Topics in Cognitive Science 8 (2016) 118–137
Copyright ©2015 Cognitive Science Society, Inc. All rights reserved.
ISSN:1756-8757 print / 1756-8765 online
DOI: 10.1111/tops.12174
Tensions Between Science and Intuition
Across the Lifespan
Andrew Shtulman, Kelsey Harrington
Department of Psychology, Occidental College
Received 30 June 2014; received in revised form 11 February 2015; accepted 20 February 2015
Abstract
The scientific knowledge needed to engage with policy issues like climate change, vaccination,
and stem cell research often conflicts with our intuitive theories of the world. How resilient are
our intuitive theories in the face of contradictory scientific knowledge? Here, we present evidence
that intuitive theories in 10 domains of knowledge—astronomy, evolution, fractions, genetics,
germs, matter, mechanics, physiology, thermodynamics, and waves—persist more than four dec-
ades beyond the acquisition of a mutually exclusive scientific theory. Participants (104 younger
adults, M
age
=19.6, and 48 older adults, M
age
=65.1) were asked to verify two types of scientific
statements as quickly as possible: those that are consistent with intuition (e.g., “the moon revolves
around the Earth”) and those that involve the same conceptual relations but are inconsistent with
intuition (e.g., “the Earth revolves around the sun”). Older adults were as accurate as younger
adults at verifying both types of statements, but the lag in response times between intuition-consis-
tent and intuition-inconsistent statements was significantly larger for older adults than for younger
adults. This lag persisted even among professional scientists. Overall, these results suggest that the
scientific literacy needed to engage with topics of global importance may be constrained by
patterns of reasoning that emerge in childhood but persist long thereafter.
Keywords: Conceptual development; Explanatory coexistence; Na€
ıve theories; Scientific
knowledge; Science education; Speeded reasoning
1. Introduction
Being an informed citizen in today’s highly technological world requires a fair amount
of scientific literacy. Engaging with debates on how to curtail climate change, for
instance, requires an understanding of geological systems and how those systems are
Correspondence should be sent to Andrew Shtulman, Department of Psychology, Occidental College,
1600 Campus Road, Los Angeles, CA 90041. E-mail: shtulman@oxy.edu

2008); for matter, the change from a tactile theory of material substances to a particulate
theory (Smith, 2007); for mechanics, the change from an impetus theory of motion to an
inertial theory (McCloskey, 1983); for physiology, the change from a psychological the-
ory of bodily functions to a vitalist theory (Johnson & Carey, 1998); for thermodynamics,
the change from a substance-based theory of heat to a kinetic theory (Reiner et al.,
2000); and for waves, the change from a substance-based theory of light and sound to a
frequency-based theory (Mazens & Lautrey, 2003).
The materials were counterbalanced in three respects. First, there were an equal num-
ber of objectively true and objectively false statements per domain, discouraging partici-
pants from adopting a response bias. Second, the average number of words per statement
was held constant across statements and domains, give or take a few words. Third, the
linguistic complexity of the statements was held constant across the four statements
designed to probe any given concept so that simpler statements (e.g., “[entity] is com-
posed of matter”) were represented as often as more complex statements (e.g., “[mate-
rial
1
] is denser than [material
2
]”) among each stimulus category (TT statements, FF
statements, TF statements, and FT statements). Note that, by equating the logical com-
plexity of each statement in a four-statement grouping, those statements differed only in
content (i.e., the content of the objects to which the target conceptual relation was
applied). Our task was thus qualitatively different from those used to assess belief bias
and other types of deductive errors (e.g., Evans, Barston, & Pollard, 1983) as those tasks
purposely pit content against logic. The full list of 200 statements is available upon
request.
2.3. Procedure
Stimuli were presented to participants with MediaLab v1.21 software (Empirisoft, New
York, NY, USA), which recorded the speed and accuracy of their statement-verification
judgments. The mean response time across items and across subjects was 3.67 s for the
younger adults and 5.56 s for the older adults, and all response times that fell more than
2SD beyond the means for each group were eliminated from the data set. Statements
from the same domain were presented as a block, prefaced by the domain name, in order
to minimize abrupt changes in content, but their ordering was randomized within that
block, as was the ordering of the blocks themselves.
3. Results
Effects of age (younger vs. older adults) on the speed and accuracy of participants’
statement verifications are presented first, followed by effects of expertise (science profes-
sors vs. other older adults). Only correct responses were retained for the analysis of
response latency, as correct responses provide the cleanest test of the hypotheses of inter-
est. Nevertheless, the same effects were observed when incorrect responses were included
as well.
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 125

3.1. Effects of age
3.1.1. Response accuracy
Fig. 1A shows the mean proportion of intuition-consistent statements and intuition-
inconsistent statements correctly verified by younger and older participants, averaged
across domain and across the statements’ truth value. (Effects of truth value are analyzed
0.5
0.6
0.7
0.8
0.9
1.0
Younger Adults Older Adults
Mean Proportion Correct
Consistent Inconsistent
3.0
3.5
4.0
4.5
5.0
5.5
Younger Adults Older Adults
Mean Response Times (sec)
Consistent Inconsistent
(A)
(B)
Fig. 1. Mean proportion of correct verifications (A) and mean response times (B) by statement type (intu-
ition-consistent vs. intuition-inconsistent) and age group (younger vs. older adults); all SE <0.02 for (A) and
all SE <0.025 for (B).
126 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

below.) We submitted those data to a repeated measures analysis of variance (ANOVA)in
which statement type (intuition-consistent vs. intuition-inconsistent) was treated as a
within-participants variable and age group (younger adults vs. older adults) was treated as
a between-participants variable. This analysis revealed a significant effect of statement
type (F(1, 150) =786.58, p<.001), but no effect of age (F(1, 150) <1). That is, partici-
pants in both age groups verified intuition-consistent statements more accurately than they
verified intuition-inconsistent statements, but neither group was more accurate, on the
whole, than the other.
There was, however, a significant interaction between statement type and age (F(1,
150) =3.94, p<.05) such that the difference in correct verifications for intuition-
consistent and intuition-inconsistent statements was slightly more pronounced for younger
participants (0.18) than for older participants (0.16). That difference withstanding,
younger and older participants performed comparably with respect to the statements’ con-
sistency with intuition. They also performed comparably with respect to the statements’
truth value, judging objectively true statements as “true” (younger participants: M=0.81,
SD =0.10; older participants: M=0.80, SD =0.11) and objectively false statements as
“false” (younger participants: M=0.69, SD =0.13; older participants: M=0.70,
SD =0.14) at statistically equivalent rates.
3.1.2. Response latency
Mean response times for intuition-consistent and intuition-inconsistent statements are
displayed as a function of age group in Fig. 1B. These data were analyzed with repeated
measures ANOVAs of the same type as that described earlier. Participants in both age
groups verified inconsistent statements significantly slower than they verified consistent
statements (F(1, 150) =287.48, p<.001), and older adults verified both types of state-
ments significantly slower than younger adults did (F(1, 150) =48.33, p<.001). A sig-
nificant interaction between statement type and age group (F(1, 150) =28.20, p<.001)
indicated that the lag in response times between consistent and inconsistent statements
was not equivalent for younger and older adults. For younger adults that lag averaged
0.36 s, whereas for older adults that lag averaged 0.69 s.
To assess the consistency of these effects, we repeated our analyses separately for each
domain. Response times are displayed as a function of domain, age group, and statement
type in Table 3. In all domains, participants in both age groups verified inconsistent state-
ments more slowly than they verified consistent statements (all Fs>4.8, all ps<.05),
and older participants verified both types of statements significantly slower than younger
participants did (all Fs>23.5, all ps<.001). The interaction between statement type and
age was significant in only four domains (evolution: F(1, 150) =10.01, p<.01; frac-
tions: F(1, 150) >75.04, p<.001; matter: F(1, 150) >4.13, p<.05; physiology: F(1,
150) >6.32, p<.05). Nevertheless, older participants exhibited a larger lag in response
times between intuition-consistent and intuition-inconsistent statements than that exhibited
by younger participants in all domains but one (genetics).
Finally, we explored whether the effect of statement type on response latency held
both for statements that were objectively true (“fish are alive” vs. “coral is alive”) and
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 127

for statements that were objectively false (“rocks are alive” vs. “the sun is alive”). To
do so, we computed separate means for the two types of intuition-consistent statements
(TT and FF statements) and the two types of intuition-inconsistent statements (FT and
TF statements). Paired-samples ttests revealed that participants in both age groups veri-
fied intuition-inconsistent statements significantly slower than intuition-consistent state-
ments regardless of whether those statements were objectively true (younger
participants: t(103)=12.90, p<.001; older participants: t(47)=8.87, p<.001) or objec-
tively false (younger participants: t(103)=6.70, p<.001; older participants: t(47) =
6.39, p<.001).
That said, the difference in response times between the two types of statements was
larger for true statements (M=0.55 s) than for false statements (M=0.40 s), and a
repeated measures ANOVA confirmed that the interaction between statement type (intu-
ition-consistent vs. intuition-inconsistent) and statement truth value (true vs. false) was
significant (F(1, 151) =9.29, p<.01). False statements were apparently less reliable
than true statements at eliciting conflict between science and intuition, possibly because
false statements were verified more slowly overall (Mseconds for false state-
ments =4.17, Mseconds for true statements =3.88, F(1, 151) =58.65, p<.001). Nev-
ertheless, the size of the interaction between statement type and truth value (partial
g
2
=0.06) was substantially smaller than the size of the main effect of statement type
Table 3
Mean response times (seconds) for intuition-consistent and intuition-inconsistent statements in each domain
and age group, as well as response lags between the two types of statements (intuition-inconsistent minus
intuition-consistent) and differences in response lags between the two age groups (older minus younger)
Domain Age group Inconsistent Consistent Response Lag Difference
Astronomy Younger 3.86 3.68 0.18 0.11
Older 5.08 4.79 0.29
Evolution Younger 4.32 3.96 0.36 0.60
Older 5.68 4.72 0.96
Fractions Younger 4.54 3.70 0.84 1.27
Older 6.77 4.66 2.11
Genetics Younger 3.92 3.56 0.36 0.24
Older 4.75 4.63 0.12
Germs Younger 3.07 3.00 0.07 0.16
Older 3.91 3.68 0.23
Matter Younger 3.70 3.43 0.27 0.36
Older 4.96 4.33 0.63
Mechanics Younger 4.04 3.94 0.10 0.25
Older 5.26 4.91 0.35
Physiology Younger 3.11 2.62 0.49 0.35
Older 4.27 3.43 0.84
Thermodynamics Younger 4.30 3.78 0.52 0.22
Older 5.64 4.90 0.74
Waves Younger 3.90 3.44 0.46 0.21
Older 5.15 4.48 0.67
128 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

(partial g
2
=0.58), indicating that a statement’s truth value played a minor role in
how quickly it was verified compared to whether or not that statement was consistent
with intuition.
3.2. Effects of expertise
3.2.1. Response accuracy
The mean proportion of intuition-consistent and intuition-inconsistent statements cor-
rectly verified by the three groups of older adults is displayed in Fig. 2A. We submitted
those data to a repeated measures ANOVA in which statement type (intuition-consistent
vs. intuition-inconsistent) was treated as a within-participants factor and occupation
3.5
4.0
4.5
5.0
5.5
6.0
Non-Profs Humanities Profs Science Profs
Mean Response Times (sec)
Consistent Inconsistent
0.5
0.6
0.7
0.8
0.9
1.0
Non-Profs Humanities Profs Science Profs
Mean Proportion Correct
Consistent Inconsistent
(A)
(B)
Fig. 2. Mean proportion of correct verifications (A) and mean response times (B) as a function of statement
type (intuition-consistent vs. intuition-inconsistent) and occupation (non-professors, humanities professors,
science professors) for the older adults; all SE <0.03 for (A) and all SE <0.035 for (B).
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 129

(science professors vs. humanities professors vs. non-professors) was treated as a
between-participants factor. Correct verifications varied not only by statement type (F(1,
45) =149.38, p<.001), but also by occupation (F(2, 45) =13.76, p<.001) and by the
interaction between statement type and occupation (F(2, 45) =9.28, p<.001).
We explored the effect of occupation with linear contrast analyses in which scores
for non-professors were weighted “1,” scores for humanities professors “0,” and
scores for science professors “1.” These analyses revealed that correct verifications
increased monotonically from non-professors to humanities professors to science profes-
sors, both for consistent statements (F(1, 47) =8.92, p<.01) and for inconsistent
statements (F(1, 47) =30.58, p<.001). And Bonferroni comparisons revealed that
science professors performed significantly better than the other two groups for both
types of statements (p<.05). The superior performance of science professors helps
validate the task itself, particularly the difference in performance between science pro-
fessors and humanities professors. Both types of participants had attained similar levels
of education and had engaged in similar kinds of professional activities throughout
their careers, yet only science professors possessed the requisite content knowledge for
verifying the vast majority of statements correctly. Science professors also exhibited a
smaller discrepancy in accuracy between intuition-consistent and intuition-inconsistent
statements than did the other two groups (science professors, 9%; humanities profes-
sors, 13%; non-professors, 20%), as revealed by a significant interaction between state-
ment type and occupation noted above. This finding indicates that a statement’s
consistency with intuition matters less the more one knows about science, at least with
respect to accuracy.
3.2.2. Response latency
Mean response times for intuition-consistent and intuition-inconsistent statements are
displayed as a function of occupation in Fig. 2B. Similar to the accuracy data, the latency
data varied significantly by both statement type (F(1 ,45) =76.96, p<.001) and
occupation (F(2, 45) =3.72, p<.05). That is, older adults, on the whole, verified intu-
ition-inconsistent statements significantly slower than they verified intuition-consistent
statements, and older adults in some occupations verified both types of statements signifi-
cantly slower than did those in other occupations.
To explore the latter effect, we submitted response times to linear contrasts of the
same form described earlier. These analyses revealed that response times decreased
monotonically from non-professors to humanities professors to science professors, both
for intuition-consistent statements (F(1, 47) =4.66, p<.05) and for intuition-inconsistent
statements (F(1, 47) =5.47, p<.05). Nevertheless, there was no interaction between
statement type and occupation (F(2, 45) =1.57, ns), as the lag in response times between
intuition-consistent and intuition-inconsistent statements was comparable across groups:
non-professors (0.72 s), humanities professors (0.82 s), science professors (0.46 s). Thus,
although science professors outperformed their age-matched peers in terms of accuracy,
they performed similarly to their peers in terms of speed.
130 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

4. Discussion
Scientific knowledge is critical for making informed decisions about many globally
important issues, but scientific knowledge does not come easily. Learning science
requires revising and restructuring earlier-acquired, intuitive theories that typically con-
tradict scientific theories of the same phenomena. Moreover, recent research suggests
that intuitive theories are not replaced by scientific theories but coexist with them,
obstructing the retrieval of scientific theories by providing an alternative explanation for
the phenomena at hand that must first be inhibited. In the present study, we explored
the robustness of this finding across age and expertise. We found that, regardless of
how old our participants were or how much science expertise they had acquired, they
verified scientific statements that were inconsistent with intuition significantly slower
than they verified scientific statements that were consistent with intuition. This pattern
is remarkable given that, in many domains (e.g., germs, matter, physiology), partici-
pants acquired their relevant scientific knowledge as children. Yet the lag in response
times between intuition-consistent and intuition-inconsistent statements did not diminish
with age; if anything, it increased. And this lag was evident not only among adults in
non-scientific occupations, but also among professional scientists with three to four
decades of career experience.
Our finding that older adults are no more immune to the conflict between science
and intuition than are younger adults accords well with prior research on scientific rea-
soning in Alzheimer’s patients (Lombrozo et al., 2007; Zaitchik & Solomon, 2008).
Although the cognitive impairments wrought by Alzheimer’s disease appear to liberate
pre-scientific intuitions at an explicit level, our findings indicate that the same intuitions
persist at an implicit level among neurologically healthy adults of a similar age. More-
over, our finding that scientists are no more immune to the conflict between science
and intuition than are non-scientists accords well with prior research involving other
expert populations. Under speeded conditions, professional biologists reveal animistic
intuitions of the same sort revealed by non-biologists, for example, that comets are
alive and that orchids are not alive (Goldberg & Thompson-Schill, 2009), and profes-
sional physicists endorse unwarranted teleological explanations of the same sort
endorsed by non-physicists, for example, that “moss forms around rocks in order to
stop soil erosion” and that “the sun makes light so that plants can photosynthesize”
(Kelemen et al., 2013). Our results suggest that these concept-specific instances of cog-
nitive conflict are symptomatic of a more general pattern, one that encompasses multi-
ple concepts in multiple domains.
Thus far, we have interpreted the observed effects as evidence of conflict between intu-
itive theories and scientific theories, but it is also possible that these effects reflect a more
localized conflict—that is, a conflict between isolated beliefs. In other words, the conflict
observed for statements like “ice has heat” (which is scientifically true but intuitively
false) or “coats produce heat” (which is scientifically false but intuitively true) may
reflect discrepancies not in our theories of heat but in our beliefs about heat, some of
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 131

which are consistent with science and some of which are not. This interpretation is con-
sistent with a “knowledge-in-pieces” view of folk beliefs that treats such beliefs as frag-
mented and incoherent (DiSessa, 1993; Hammer, 1996). Whether our folk beliefs are best
characterized as self-consistent theories or as piecemeal conglomerations is outside the
scope of the present study (see instead Vosniadou, 2010). Still, there are at least two find-
ings that favor a theory-based interpretation of the data over a knowledge-in-pieces inter-
pretation.
First, participants demonstrated the effect of interest across most to all concepts within
any given domain of knowledge. That is, they exhibited slower response times for intu-
ition-inconsistent statements relative to intuition-consistent statements for four of the five
concepts in the domains of astronomy, germs, genetics, matter, mechanics, thermodynam-
ics, and waves, and for all five concepts in the domains of evolution, fractions, and physi-
ology. The consistency of this effect across multiple concepts within the same domain
suggests that those concepts are interrelated (as has been shown in many other studies as
well, e.g., Au et al., 2008; Mazens & Lautrey, 2003; McCloskey, 1983; Shtulman, 2006;
Smith, 2007). Second, even scientists exhibited the effect, and it would be a stretch to
claim that scientists’ beliefs about natural phenomena are fragmented and incoherent.
Indeed, a corollary of the knowledge-in-pieces view of folk beliefs is that science learn-
ing serves to connect and unify those beliefs, yielding a knowledge base increasingly
devoid of internal inconsistencies. Yet such inconsistencies do not appear to go away,
and only a theory-based view of the observed effects provides a means of explaining
them.
In sum, the present results demonstrate that intuitive theories are highly resilient across
age, occupation, and domain. They do not tell us, however, why intuitive theories are so
resilient or how intuitive theories affect our reasoning outside the laboratory. Below we
address each question in turn, speculating on how the observed effects alter our under-
standing of the acquisition and representation of scientific knowledge.
5. Why are intuitive theories so resilient?
One explanation for why intuitive theories seem to persist across the lifespan is that
they may be represented in the brain in a cognitively impenetrable format, similar to the
seemingly impenetrable representations of language (Coltheart, 1999) and vision (Pyly-
shyn, 1999). In vision, for instance, we can be well aware that our eyes deceive us when
viewing the Muller-Lyer illusion or the Ponzo illusion, but we perceive the illusion
nonetheless. The visual biases that give rise to such illusions constitute a stable backdrop
against which all new visual information is interpreted, and those biases operate even
when we are aware of their fallibility. Intuitive theories might be represented in the brain
in a similar fashion, though this explanation begs the question as to what constitutes an
intuitive theory and why such representations are impervious to revision. Part of the
appeal of describing folk beliefs as “theories,” after all, is that such beliefs are presum-
ably open to revision (Gopnik & Wellman, 2012).
132 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

Another explanation for why intuitive theories seem to persist across the lifespan is
that intuitive theories are actively reinforced by how we talk about natural phenomena
in everyday discourse and how we perceive natural phenomena in everyday situations.
Much of our colloquial language seems to be predicated on intuitive conceptions. The
terms “sunrise” and “sunset,” for instance, imply that day and night are caused by
movements of the sun rather than movements of the earth. More accurate terms would
be “sun accretion” and “sun occlusion.” Likewise, the terms “warm coat” and “cold
wind” imply that heat is an intrinsic property of objects rather than something that is
transferred across physical systems. More accurate terms would be “insulating coat” and
“disequilibrating wind.”
Our perceptual experience is no less misleading. Coats feel as if they produce heat,
and the sun looks as if it moves across the sky. Recent models of theory change have
begun to emphasize the role of a theory’s cognitive utility in catalyzing that process (e.g.,
Gopnik & Wellman, 2012; Ohlsson, 2009), and these models predict that an intuitive the-
ory should be maintained alongside a scientific theory so long as that theory provides a
sufficiently useful interpretation of the data at hand—that is, the data of everyday linguis-
tic practices and the data of everyday perceptual experiences. The challenge to those who
would explain the resilience of intuitive theories in terms of their cognitive utility is to
clarify the dimensions along which cognitive utility is calculated and the situations in
which intuitive theories trump scientific theories along those particular dimensions.
6. How do intuitive theories affect scientific reasoning?
In the present study, participants of all levels of science expertise exhibited cognitive
conflict when verifying intuition-inconsistent statements. This conflict demonstrates that
all participants continued to hold onto their intuitive theories (at some level of representa-
tion), but it does not tell us how those theories influence scientific reasoning “in the
wild.” We suspect that intuitive theories influence scientific reasoning in different ways at
different points in the development of scientific expertise. Early on, intuitive theories
undoubtedly interfere with the acquisition of scientific theories. Such interference has
been documented previously, in many domains using many methods (see Vosniadou,
2010), but the discovery that intuitive theories are never truly overwritten by scientific
theories suggests that this interference is more pervasive and more pernicious than origi-
nally thought. Students must contend with their intuitive theories not only at the outset of
learning, but also throughout the process of learning.
The resilience of intuitive theories may also help to explain why knowledge of science
has often been found to be unrelated to acceptance of science—for example, acceptance
that humans have caused climate change (Kahan et al., 2012) or that humans evolved from
non-human ancestors (Sinatra, Southerland, McConaughy, & Demastes, 2003). Intuitive
theories likely cause conflict or confusion when attempting to engage with scientific issues
that run counter to those theories, creating opportunities for religious views (Jelen & Lock-
ett, 2014; Shtulman, 2013) or political views (Kahan, Jenkins-Smith, & Braman, 2011) to
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 133

sway one’s opinions on the matter. On the other hand, knowledge of science and accep-
tance of science are not always unrelated; some studies have found that the former does
indeed predict the latter (Ingram & Nelson, 2006; Ranney & Clark, 2015; Rutledge &
Warden, 1999; Shtulman & Calabi, 2012). The discrepancy between these studies may be
due, in part, to the types of knowledge that they have assessed. Studies that have found a
relation between science knowledge and science acceptance have typically assessed partic-
ipants’ mechanistic understanding of the relevant science, whereas those that have not
found such a relation have typically assessed participants’ factual understanding. The bet-
ter we understand the mechanisms of a phenomenon, the more likely we accept that phe-
nomenon as true. Still, future research is needed to determine whether mechanistic
knowledge produces this effect on its own or whether it mediates the effect by shielding
participants from other, non-scientific considerations, such as religious views, political
views, or—most pertinent to the present study—one’s own intuitive theories.
While intuitive theories may affect the credence that science novices place on scientific
findings, they are unlikely to affect the credence that science experts place on such find-
ings. Science experts, after all, have likely established strong boundaries between their
intuitive theories of a domain and their scientific theories of a domain, particularly for
their own domains of expertise. Given the right contextual cues, even non-scientists read-
ily partition mutually incompatible pieces of knowledge into separate mental parcels
(Lewandowsky, Kalish, & Ngang, 2002; Yang & Lewandowsky, 2004). Those partitions
may break down, however, in the absence of the contextual cues upon which they were
first established. For scientists, those cues are likely embedded in the practices of their
trade—that is, running experiments, analyzing data, writing up findings, conversing with
colleagues. Outside those contexts, scientists may be no more likely to honor the bound-
ary between science and intuition than non-scientists. Indeed, there is evidence that scien-
tists are prone to error when applying familiar principles to unfamiliar problems (Dunbar,
1995), when evaluating familiar information in unfamiliar formats (Eddy, 1982), and
when making empirical projections that run counter to prevailing beliefs (Brysse,
Oreskes, O’Reilly, & Oppenheimer, 2013). Whether those errors are caused by a break-
down in the partitioning of science and intuition, as opposed to some other cognitive lim-
itation, is a question in need of further research.
7. Conclusion
Open questions aside, our results indicate that intuitive theories persist across the
lifespan, influencing scientific reasoning for decades beyond the acquisition of a mutu-
ally incompatible scientific theory. The resilience of intuitive theories may be partially
responsible for public skepticism toward science, though future research is needed to
determine whether and how the implicit representation of intuitive theories affects our
explicit attitudes toward science. Still, developing an awareness of those theories may
provide some immunity to their sway over attitudes and decisions better informed by
science alone.
134 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

Acknowledgments
We thank the National Science Foundation for supporting this research with grant
DRL-0953384 awarded to Andrew Shtulman. We also thank Victoria Halote, Alexander
Levin, Cara Neal, and Linneen Warren for their assistance with data collection.
References
Au, T., Chan, C., Chan, T., Cheung, M., Ho, J., & Ip, G. (2008). Folkbiology meets microbiology: A study
of conceptual and behavioral change. Cognitive Psychology,57,1–19.
Babai, R., Sekal, R., & Stavy, R. (2010). Persistence of the intuitive conception of living things in
adolescence. Journal of Science Education and Technology,19,20–26.
Brysse, K., Oreskes, N., O’Reilly, J., & Oppenheimer, M. (2013). Climate change prediction: erring on the
side of least drama? Global Environmental Change,23, 327–337.
Carey, S. (2009). The origin of concepts. New York: Oxford University Press.
Chi, M. T., Feltovich, P. J., & Glaser, R. (1981). Categorization and representation of physics problems by
experts and novices. Cognitive Science,5, 121–152.
Coltheart, M. (1999). Modularity and cognition. Trends in Cognitive Sciences,3, 115–120.
DiSessa, A. A. (1993). Toward an epistemology of physics. Cognition and Instruction,10, 105–225.
Dunbar, K. (1995). How scientists really reason: scientific reasoning in real-world laboratories. In J. E.
Davidson & R. J. Sternberg (Eds.), The nature of insight (pp. 365–395). Cambridge, MA: The MIT Press.
Dunbar, K., Fugelsang, J., & Stein, C. (2007). Do na€
ıve theories ever go away? Using brain and behavior to
understand changes in concepts. In M. Lovett & P. Shah (Eds.), Thinking with data (pp. 193–206). New
York: Lawrence Erlbaum Associates.
Duncan, R. G., Rogat, A. D., & Yarden, A. (2009). A learning progression for deepening students’
understandings of modern genetics across the 5th-10th grades. Journal of Research in Science Teaching,
46, 655–674.
Eddy, D. M. (1982). Probabilistic reasoning in clinical medicine: problems and opportunities. In D.
Kahneman, P. Slovic, & A. Tversky (Eds.), Judgment under uncertainty: Heuristics and biases (pp. 249–
267). Cambridge, UK: Cambridge University Press.
Evans, J. St. B. T., Barston, J., & Pollard, P. (1983). On the conflict between logic and belief in syllogistic
reasoning. Memory and Cognition,11, 295–306.
Feil, A., & Mestre, J. P. (2010). Change blindness as a means of studying expertise in physics. Journal of the
Learning Sciences,19, 480–505.
Goldberg, R. F., & Thompson-Schill, S. L. (2009). Developmental “roots” in mature biological knowledge.
Psychological Science,20, 480–487.
Gopnik, A., & Wellman, H. M. (2012). Reconstructing constructivism: causal models, Bayesian learning
mechanisms, and the theory theory. Psychological Bulletin,138, 1085–1108.
Hammer, D. (1996). Misconceptions or p-prims: How may alternative perspectives of cognitive structure
influence instructional perceptions and intentions? The Journal of the Learning Sciences,5,97–127.
Hatano, G., & Inagaki, K. (1994). Young children’s na€
ıve theory of biology. Cognition,50, 171–188.
Ingram, E. L., & Nelson, C. E. (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.
Jelen, T. G., & Lockett, L. A. (2014). Religion, partisanship, and attitudes toward science policy. SAGE
Open,4,1–8. doi:10.1177/2158244013518932.
Johnson, S. C., & Carey, S. (1998). Knowledge enrichment and conceptual change in folkbiology: Evidence
from Williams syndrome. Cognitive Psychology,37, 156–200.
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 135

Kahan, D. M., Jenkins-Smith, H., & Braman, D. (2011). Cultural cognition of scientific consensus. Journal of
Risk Research,14, 147–174.
Kahan, D. M., Peters, E., Wittlin, M., Slovic, P., Ouellette, L. L., Braman, D., & Mandel, G. (2012). The
polarizing impact of science literacy and numeracy on perceived climate change risks. Nature Climate
Change,2, 732–735.
Kelemen, D. (1999). Function, goals and intention: Children’s teleological reasoning about objects. Trends in
Cognitive Sciences,3, 461–468.
Kelemen, D., Rottman, J., & Seston, R. (2013). Professional physical scientists display tenacious teleological
tendencies: Purpose-based reasoning as a cognitive default. Journal of Experimental Psychology: General,
142, 1074–1083.
Lewandowsky, S., Kalish, M., & Ngang, S. K. (2002). Simplified learning in complex situations: Knowledge
partitioning in function learning. Journal of Experimental Psychology: General,131, 163–193.
Lombrozo, T., Kelemen, D., & Zaitchik, D. (2007). Inferring design: Evidence of a preference for
teleological explanations for patients with Alzheimer’s disease. Psychological Science,18, 999–1006.
Masson, S., Potvin, P., Riopel, M., & Foisy, L. M. B. (2014). Differences in brain activation between novices
and experts in science during a task involving a common misconception in electricity. Mind, Brain, and
Education,8,44–55.
Mazens, K., & Lautrey, J. (2003). Conceptual change in physics: children’s na€
ıve representations of sound.
Cognitive Development,18, 159–176.
McCloskey, M. (1983). Na€
ıve theories of motion. In D. Gentner & A. L. Stevens (Eds.), Mental models (pp.
299–324). Hillsdale, NJ: Erlbaum.
Ohlsson, S. (2009). Resubsumption: a possible mechanism for conceptual change and belief revision.
Educational Psychologist,44,20–40.
Piaget, J. (1929). The child’s conception of physical causality. London: Routledge.
Potvin, P., Turmel, E., & Masson, S. (2014). Linking neuroscientific research on decision making to the
educational context of novice students assigned to a multiple-choice scientific task involving common
misconceptions about electrical circuits. Frontiers in Human Neuroscience,8,1–13.
Pylyshyn, Z. (1999). Is vision continuous with cognition? The case for cognitive impenetrability of visual
perception. Behavioral and Brain Sciences,22, 341–365.
Ranney, M. A., & Clark, D. (2015). Climate change conceptual change: scientific information can transform
attitudes. Topics in Cognitive Science,8,49–75.
Reiner, M., Slotta, J. D., Chi, M. T., & Resnick, L. B. (2000). Na€
ıve physics reasoning: a commitment to
substance-based conceptions. Cognition and Instruction,18,1–34.
Rutledge, M. L., & Warden, M. A. (1999). The development and validation of the measure of acceptance of
the theory of evolution instrument. School Science and Mathematics,99,13–18.
Shanteau, J. (1992). How much information does an expert use? Is it relevant? Acta Psychologica,81,
75–86.
Shtulman, A. (2006). Qualitative differences between na€
ıve and scientific theories of evolution. Cognitive
Psychology,52, 170–194.
Shtulman, A. (2013). Epistemic similarities between students’ scientific and supernatural beliefs. Journal of
Educational Psychology,105, 199–212.
Shtulman, A., & Calabi, P. (2012). Cognitive constraints on the understanding and acceptance of evolution.
In K. S. Rosengren, S. Brem, E. M. Evans, & G. Sinatra (Eds.), Evolution challenges: Integrating
research and practice in teaching and learning about evolution (pp. 47–65). Cambridge, UK: Oxford
University Press.
Shtulman, A., & Valcarcel, J. (2012). Scientific knowledge suppresses but does not supplant earlier intuitions.
Cognition,124, 209–215.
Siegler, R. S., Fazio, L. K., Bailey, D. H., & Zhou, X. (2013). Fractions: the new frontier for theories of
numerical development. Trends in Cognitive Sciences,17,13–19.
136 A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016)

Sinatra, G. M., Southerland, S. A., McConaughy, F., & Demastes, J. W. (2003). Intentions and beliefs in
students’ understanding and acceptance of biological evolution. Journal of Research in Science Teaching,
40, 510–528.
Smith, C. L. (2007). Bootstrapping processes in the development of students’ commonsense matter theories:
using analogical mappings, thought experiments, and learning to measure to promote conceptual
restructuring. Cognition and Instruction,25, 337–398.
Vosniadou, S. (1994). Capturing and modeling the process of conceptual change. Learning and Instruction,4,
45–69.
Vosniadou, S. (Ed.) (2010). International handbook of research on conceptual change. New York:
Routledge.
Vosniadou, S., & Brewer, W. F. (1994). Mental models of the day/night cycle. Cognitive Science,18, 123–
183.
Yang, L. X., & Lewandowsky, S. (2004). Knowledge partitioning in categorization: Constraints on exemplar
models. Journal of Experimental Psychology: Learning, Memory, and Cognition,30, 1045–1064.
Zaitchik, D., & Solomon, G. E. A. (2008). Animist thinking in the elderly and in patients with Alzheimer’s
disease. Cognitive Neuropsychology,25,27–37.
Zaitchik, D., & Solomon, G. E. A. (2009). Conservation of species, volume, and belief in patients with
Alzheimer’s disease: The issue of domain specificity and conceptual impairment. Cognitive
Neuropsychology,26, 511–526.
A. Shtulman, K. Harrington / Topics in Cognitive Science 8 (2016) 137