Intuitive biological thought: Developmental changes and effects
of biology education in late adolescence
John D. Coley
, Melanie Arenson
, Yian Xu
, Kimberly D. Tanner
Department of Psychology, Northeastern University, United States
Department of Medicine, University of California, San Francisco, United States
San Francisco VA Medical Center, United States
Department of Biology, San Francisco State University, United States
Accepted 1 November 2016
A large body of cognitive research has shown that people intuitively and effortlessly reason
about the biological world in complex and systematic ways. We addressed two questions
about the nature of intuitive biological reasoning: How does intuitive biological thinking
change during adolescence and early adulthood? How does increasing biology education
inﬂuence intuitive biological thinking? To do so, we developed a battery of measures to
systematically test three components of intuitive biological thought: anthropocentric think-
ing, teleological thinking and essentialist thinking, and tested 8th graders and university stu-
dents (both biology majors, and non-biology majors). Results reveal clear evidence of
persistent intuitive reasoning among all populations studied, consistent but surprisingly
small differences between 8th graders and college students on measures of intuitive bio-
logical thought, and consistent but again surprisingly small inﬂuence of increasing biology
education on intuitive biological reasoning. Results speak to the persistence of intuitive
reasoning, the importance of taking intuitive knowledge into account in science class-
rooms, and the necessity of interdisciplinary research to advance biology education.
Further studies are necessary to investigate how cultural context and continued acquisition
of expertise impact intuitive biology thinking.
Ó2016 Elsevier Inc. All rights reserved.
Cognitive scientists and educators alike acknowledge that students do not arrive at the science classroom as blank slates,
but rather have developed complex and adaptive intuitive conceptual systems for understanding the world around them. As
such, science education results from the interplay between students’ intuitive ways of knowing and scientiﬁc concepts intro-
duced by expert instructors, across a range of STEM disciplines, including physics (e.g., Chi, 1992; DiSessa, 1993; Vosniadou
& Brewer, 1992), chemistry (Maeyer & Talanquer, 2010), and biology (Coley & Tanner, 2012, 2015; Kelemen, Rottman, &
Seston, 2013; Shtulman, 2006). As such, it is critically important to understand the nature and content of intuitive
understandings to inform science education. In this paper, we investigate the development of intuitive biological thought
0010-0285/Ó2016 Elsevier Inc. All rights reserved.
Corresponding author at: Department of Psychology, 125 Nightingale Hall, Northeastern University, 360 Huntington Avenue, Boston, MA 02115, United
E-mail address: firstname.lastname@example.org (J.D. Coley).
Cognitive Psychology 92 (2017) 1–21
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/cogpsych
in adolescence and young adulthood, and begin to explore potential impacts of increasing science education on this intuitive
1.1. Cognitive construals in intuitive biological thought
Humans naturally, intuitively, and effortlessly reason about biological entities, structures, processes, and phenomena in
predictable ways (e.g., Atran & Medin, 2008; Berlin, 1992; Brown, 1984; Carey, 1985; Coley, Solomon, & Shafto, 2002; Inagaki
& Hatano, 2006; Medin & Atran, 2004). Elsewhere, we have dubbed these regularities cognitive construals (Coley & Tanner,
2012, 2015). A cognitive construal is an intuitive, often implicit, way of thinking about the world. It might be a set of assump-
tions, a type of explanation, or a predisposition to a particular type of reasoning. Three such cognitive construals—teleological
thinking, essentialist thinking, and anthropocentric thinking—recur in research on the development of intuitive biological
thought. Framing intuitive biological thought in terms of these three construals is a useful way to represent a large and dis-
parate literature. We do not claim that this list is exhaustive, nor do we claim that these construals are exclusively relevant
to thinking about biology. Rather, we focus on these three construals because they have received the bulk of attention in the
cognitive and developmental literature, and we have shown that they are linked to scientiﬁc misconceptions in previous
work (Coley & Tanner, 2015). In the following, we brieﬂy describe each construal, its role in intuitive biological thought,
and its developmental trajectory.
1.1.1. Anthropocentric thinking
Anthropocentric thinking can involve the tendency to attribute human characteristics to non-human or inanimate objects
(e.g., Piaget, 1929; Richards & Siegler, 1986), to use humans as a default analogical base for reasoning about biological species
or processes (e.g., Carey, 1985; Inagaki & Hatano, 1991), or to see humans as unique and biologically discontinuous with the
rest of the animal world. Although undoubtedly useful for adaptive reasoning and social cognition, anthropocentric thinking
can result in misrepresentation of the place of human beings in the natural world. Such ‘‘human exceptionalism” (Gee, 2013),
involves the way in which human beings are incorporated into the intuitive taxonomy of living things. According to geneti-
cists, humans are African great apes; we share a common ancestor who lived c. 5–8 million years ago with our closest living
relative: chimpanzees. However, intuitive biological taxonomies—particularly those found in industrialized western soci-
eties—tend to see humans as essentially separate from other species (Coley, 2007; Johnson, Mervis, & Boster, 1992). Likewise,
undergraduate students are also slower and less accurate at classifying plants—as compared to animals—as living things
(Goldberg & Thompson-Schill, 2009); this is consistent with anthropocentric thinking because it suggests that students
are less likely to apply universal biological properties to organisms that are highly dissimilar to humans.
Developmental psychologists have paid little attention to the development of anthropocentric thinking past the age of 10.
Although some studies document a shift from human-based analogical reasoning to category-based attribution of biological
properties (e.g., Carey, 1985; Inagaki & Sugiyama, 1988), research with children reveals a persistent reluctance to classify
humans with other animals (e.g., Coley, 2007; Johnson et al., 1992; Leddon, Waxman, Medin, Bang, & Washinawatok,
2012) or to attribute core biological properties, which are familiar in humans, to nonhuman organisms dissimilar to humans,
particularly plants (Arenson & Coley, 2016; Richards & Siegler, 1986). As such, the degree to which anthropocentric thinking
persists into young adulthood remains an open question.
1.1.2. Teleological thinking
Teleological thinking is causal reasoning in which a goal, purpose, function, or outcome of an event is taken as the cause of
that event (Keil, 2006; Talanquer, 2009, 2013). Kelemen (1999) argues that teleological thinking is a central component of
adults’ everyday thought. For example, people appropriately make the teleological assumptions that human actions are
directed toward certain goals, and that human artifacts, such as chairs and coats, are designed by their creators to fulﬁll some
intended purpose. In intuitive biology, people likewise apply teleological thinking to explain biological entities, structures
and processes, as if biological phenomena are deliberately designed to serve a purpose just as human actions and artifacts
do. As Kelemen emphasizes, teleological thinking provides an important component of adults’ intuitive interpretations of
why events occur or why objects have the properties that they do. Although the causes, origins, and nature of teleological
thinking are the subjects of considerable debate (e.g., Kelemen, 2004; Lombrozo & Carey, 2006; ojalehto, Waxman, &
Medin, 2013) this construal seems to be an integral part of intuitive thinking about biology.
The developmental arc of teleological thinking involves a pattern of ‘‘pruning.” Kelemen has shown that teleological
thinking is widespread (or in her terms, ‘‘promiscuous”) among young children and becomes increasingly selective
(Kelemen, 1999, 2012). In one study, 6-year-olds favored teleological explanations for a broad range of phenomena, includ-
ing properties of nonliving objects (e.g., ‘‘The rocks were pointy so that animals wouldn’t sit on them and smash them”) and
animals (e.g., birds exist ‘‘for ﬂying,” lions exist ‘‘to go in the zoo”). College students were more selective, but still utilized
teleological thinking in a biological context. Indeed, Kelemen and Rossett (2009) found that undergraduates endorsed
unwarranted teleological statements about biological phenomena (e.g., ‘‘Earthworms tunnel underground to aerate the soil”)
35% of the time, and that under time pressure, this ﬁgure increased to 51%. Thus, teleological thinking appears to become
more narrowly applied, but does not disappear in adults.
2J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
1.1.3. Essentialist thinking
Essentialist thinking captures the idea that in addition to summarizing knowledge, concepts also involve a possibly impli-
cit assumption that some unobservable essential property (an ‘‘underlying reality” or ‘‘true nature”) conveys category iden-
tity and causes observable similarities among category members (Ahn et al., 2001; Gelman, 2003; Medin & Ortony, 1989).
Although essentialist thinking provides us with an important tool to reduce the complexity of incoming information, and
allows us to organize what we know and make inferences about what we don’t know, it can also lead to overgeneralization
or unwarranted assumptions of homogeneity. In intuitive biology, essentialist thinking may encourage reasoning that
assumes a core underlying property or feature of a biological structure, species, or system determines its overt features
and identity. For example, essentialist thinking leads to the idea that superﬁcial transformations (e.g., changes in appear-
ance) should not affect category membership, which is ultimately based on the presence or lack of an underlying essential
property, rather than superﬁcial features (Keil, 1989; Rips, 1989; but see Hampton, Estes, & Simmons, 2007). Another con-
sequence of essentialist thinking is that members of a category are relatively uniform with respect to shared properties,
because a shared essence should give rise to similar properties in all category members (e.g., Shtulman & Schulz, 2008), lead-
ing us to underestimate variability. A third consequence of essentialist thinking is that category membership conveys innate
potential: because of an underlying essence, category members not only share the same properties, but also share the
propensity to develop the same characteristics over time (e.g., Gelman & Wellman, 1991; Solomon, Johnson, Zaitchik, &
Carey, 1996). In sum, essentialist thinking yields assumptions—about the naturalness and uniformity of category mem-
bers—which reduce the complexity of incoming information.
Like teleological thinking, the development of essentialist thinking involves increasing selectivity. There is substantial
evidence suggesting that essentialist thinking is an early and pervasive cognitive bias (see Gelman, 2003). Preschool children
are often overzealous in their essentialist thinking; for example, preschoolers tend to predict that offspring will resemble
birth parents with respect to characteristics ranging from biological properties like eye color to behavioral properties like
beliefs. Undergraduates are more selective, and tend to predict that offspring will resemble birth parents with respect to
physical characteristics, but adoptive parents with respect to beliefs and preferences (e.g., Eidson & Coley, 2014; Taylor,
Rhodes, & Gelman, 2009). Although scant research has explicitly examined changes in essentialist thinking between elemen-
tary school and young adulthood, the literature is somewhat equivocal about essentialist thinking in adults. For instance,
Medin and Atran (2004) argue that essentialist thinking is a pervasive aspect of intuitive biological thought across diverse
cultures, Diesendruck and Gelman (1999) argue that membership in biological kinds is more likely to be seen as absolute,
and numerous studies of category-based inference in adults (e.g., Coley & Vasilyeva, 2010; Feeney & Heit, 2007) are consis-
tent with essentialist thinking. In contrast, Kalish (2002) argues that membership in biological kinds may not always be per-
ceived as absolute, and Hampton et al. (2007) argue that membership in biological categories is surprisingly malleable. Thus,
the developmental trajectory—and indeed, the adult ‘‘endpoint”—of essentialist thinking is complex and not entirely under-
stood at present.
1.2. The inﬂuence of increasing biology education on intuitive biological reasoning
Anthropocentric, teleological, and essentialist thinking represent powerful, useful, and adaptive principles for organizing
what we know about the biological world, and allowing us to make predictions about what we don’t know. Although under-
standing the development of these construals into adulthood is important in its own right (Coley, 2000), documenting the
nature of informal intuitive biology in young adults becomes even more important in light of the fact that cognitive constru-
als may interact with scientiﬁc reasoning in unanticipated ways and may have speciﬁc implications for science education
(e.g., Coley & Tanner, 2012, 2015; Evans, 2008, 2013; Evans, Rosengren, Lane, & Price, 2012; Nehm, 2010; Nehm &
Ridgway, 2011; Opfer, Nehm, & Ha, 2012; Rosengren, Brem, Evans, & Sinatra, 2012; Talanquer, 2006). For example, measures
of both teleological and essentialist thinking have been shown to predict students’ understanding of evolution by natural
selection (e.g., Kelemen & Rossett, 2009; Shtulman & Schulz, 2008). Other evidence shows that even trained and practicing
scientists show implicit evidence of residual construal-based thinking (e.g., Goldberg & Thompson-Schill, 2009; Kelemen
et al., 2013) and that construal-based language present in biology majors’ explanations is associated with acceptance of
related scientiﬁc misconceptions (Coley & Tanner, 2015). However, to our knowledge no study has systematically explored
the impact of increasing levels of life science education on intuitive biological thought.
This question is important because it addresses the potential depth and generality of emerging disciplinary expertise
resulting from increasing amounts of formal biology education. On one hand, formal education in biology might replace intu-
itive construals with more accurate, normative thinking about biological structure and function. If so, we might expect intu-
itive biological thought to weaken or disappear among biology students. Alternatively, formal biology education might
provide students with an alternative explanatory framework that coexists alongside intuitive biology, but does not replace
it (e.g., Legare, Evans, Rosengren, & Harris, 2012; Shtulman & Valcarcel, 2012). If so, we might expect intuitive biological
thought to persist among biology students.
1.3. Current study
The current study aims to investigate two questions: (1) What is the developmental trajectory of construal-based intu-
itive biological thinking beyond childhood, into adolescence and early adulthood? (2) How do increasing amounts of biology
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 3
education inﬂuence the developmental trajectory of intuitive biological thinking? In contrast to previous studies which have
mostly focused on a single construal and utilized a single measure to test for its presence (with few exceptions, e.g., Johnson
& Carey, 1998), we developed a battery of multiple measures to systematically assess anthropocentric, teleological, and
essentialist thinking. We administered this battery to three groups of participants with increasing amounts of biology edu-
cation: 8th graders (who have experienced a standard middle school life science course), undergraduate non-science majors
(who have experienced both a standard middle school and a standard high school life science course), and undergraduate
biology majors (who have experienced both a standard middle school life science course, as well as an intensive, college-
level life science course). By comparing 8th graders to non-majors, we can examine developmental differences between pop-
ulations that have experienced standard life science education through typical secondary science curricula; such differences
could reﬂect developmental change between early adolescence and early adulthood, as well as increasing amounts of biology
education. By comparing biology majors to non-science majors, we can gauge the effects of even more extensive biology edu-
cation on intuitive biological reasoning between populations at a similar developmental stage.
A total of 211 individuals participated in the study. These were drawn from three groups. Biology majors (N = 69) were
ﬁrst-year undergraduate students at Northeastern University who were enrolled in three sections of a seminar for ﬁrst-
year biology, biochemistry, and behavioral neuroscience majors with Advanced Placement (AP) biology credit. As such, these
students were at the beginning of their university-level studies in biology, but had all taken AP biology courses in high school
and achieved a score of 4 or 5 on the AP Biology exam. Importantly, while these students had declared their intention to
major in biology and received AP biology credit, they had not yet experienced a university biology course. Non-majors
(N = 68) were students enrolled in an introductory psychology course. Non-majors were pre-screened to ensure that they
met two criteria: (1) that they were not majoring in biology, biochemistry, or behavioral neuroscience (non-majors repre-
sented a wide variety of majors, most of which were in the humanities and social sciences) and (2) that they had earned
AP credit in some subject other than biology. This was to roughly equate general academic ability for the two undergraduate
groups. Finally, 8th graders (N = 74) were students enrolled in a standard 8th grade science class at a suburban public middle
school in a medium-sized town outside of Boston, Massachusetts. Biology majors and 8th graders participated as part of their
normal classroom activities; non-majors participated to partially fulﬁll a course research requirement.
We designed a written survey comprised of measures of three different cognitive construals (Coley & Tanner, 2012, 2015)
relevant to intuitive biological reasoning: essentialist thinking, anthropocentric thinking, and teleological thinking. Most mea-
sures were borrowed and adapted from published reports. Two forms of the survey were constructed. Measures appeared on
the survey in a randomly determined order, with the constraint that different components of the same measure did not
appear consecutively. The order on Form B was the reverse of Form A. Half of the participants in each group received Form
A, and half received Form B. Some measures were identical on both forms, but most were comparable in structure and intent
but differed in speciﬁc exemplars, to maximize the number of items measuring any given construct across all participants.
Measures are summarized in Table 1, and described in detail in Appendix A.
Biology majors were tested in class for a full class period at the beginning of the academic year, and 8th graders were
tested in class several months into the academic year. In both cases, the survey was distributed and students were told
to take their time and answer carefully. Non-majors were tested individually or in small groups in a laboratory setting,
and were given the same instructions. In all cases, the survey on intuitive biological thought was administered along with
a survey on biological misconceptions (see Coley & Tanner, 2015). Completion of the surveys typically took 45–50 min.
We present the results in four sections. First, for each of the three target cognitive construals, we present a detailed anal-
ysis of performance on corresponding measures. For each measure, we will present a comparison of the three groups on a
summary statistic, followed by a more detailed analysis of performance on that speciﬁc measure. In the fourth section, we
compare 8th graders, non-majors, and biology majors on composite scores for each of our three target cognitive construals.
This provides an overview of our ﬁndings and a summary of group comparisons in terms of essentialist, anthropocentric, and
4J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
3.1. Anthropocentric thinking
The component of anthropocentric thinking we focused on was ‘‘human exceptionalism,” the idea that humans as a spe-
cies are biologically unique, separate, and distinct from the rest of the natural world.
3.1.1. Common ancestor
For this task, participants were given a list of biological categories ranging from great apes to plants, and asked which of
them shared a common ancestor with humans. Any indication that humans do not share common ancestry with another
species was taken to demonstrate anthropocentric thinking. To index anthropocentric thinking in the three groups, we
summed the number of positive responses (out of a possible score of seven) and compared means across the three groups
(see Fig. 1A). Biology majors reported more species shared common ancestry with humans than did non-majors or 8th gra-
ders, who did not differ from each other, F(2, 202) = 19.14, p< 0.001,
= 0.16, Tukey HSD p< 0.001.
To examine responses in more detail, we computed the percentage of positive responses for each speciﬁc target category
for each group (see Fig. 1B). We examined these data in two ways. First, we compared each cell to chance performance (50%)
via binomial test; above-chance performance suggests consensual endorsement of common ancestry, whereas below-chance
performance indicates systematic denial of common ancestry and chance performance reﬂects lack of consensus. On this
Measures of anthropocentric, teleological, and essentialist thinking.
Construal Measure Description Source
Participants were presented with a list of nine organisms and were asked
to choose ‘which share a common ancestor with humans.’ Organisms
include examples of an ape, a non-primate mammal, a dinosaur, a bird,
an amphibian, an insect and a plant. Lower numbers of organisms chosen
indicates higher anthropocentric thinking.
Participants were taught a novel intrinsic, physiological property about
humans and asked to rate the likelihood on a 1–7 scale that a range of
organisms (gorilla, squirrel, deer, sparrow, frog, bee, grasshopper, and daisy)
would share the same property. Lower likelihood ratings indicate higher
Carey (1985), Inagaki and Sugiyama
(1988), Ross et al. (2003)
Participants were presented with triads of pictures, each including a
human and two nonhuman species. In each triad, the human and one of
the non-human species were biologically more closely related (they
share a more recent common ancestor) than the two non-human species.
Participants were asked to choose in each triad which two were ‘most
similar biologically.’ Pairing the two non-human species indicates
Participants were presented a set of statements providing goal-oriented
explanations for human-made artifacts and natural phenomena (5
appropriate, 5 inappropriate, and 9 natural phenomena). Participants
rated agreement with each statement on a scale of 1 (strongly disagree)
to 6 (strongly agree). We are particularly interested in participants’
responses to natural phenomena statements. Higher scores indicate
stronger teleological thinking.
Kelemen et al. (2013)
Participants were told that an individual creature (ant or kangaroo)
possessed an intrinsic (Gene DH97/Cell HN45) or extrinsic property
(Parasite WJ61/Infection GU38), and were asked what percentage of the
entire category would share the same property with the exemplar.
Higher percentage responses indicate stronger essentialist thinking.
Shtulman and Schulz (2008)
Participants were told that an infant born in one family was raised by
another family, and were asked whether the child would share traits
(physical, preference, personality, ability and belief) with birth or
adoptive parents upon reaching young adulthood. Choosing birth parents
indicates essentialist thinking.
Gelman and Wellman (1991),
Solomon et al. (1996), Eidson and
Participants were given two ﬁctitious scenarios in which an individual
animal started out looking like a member of one category, but after a
transformation (due to either natural developmental process or toxic
environmental contamination) ended up looking like a member of
another category. Participants were asked to rate on a 1–9 scale whether
the animal was biologically a member of the original category or the end-
state category both before and after the transformation. Essentialist
thinking is reﬂected in the preservation of category membership.
Keil (1989), Rips (1989), Hampton
et al. (2007)
Participants were presented with a range of living things, and asked
whether each of them was an absolute member, absolute non-member,
or ‘sort of member’ of the categories tree,bird,orﬁsh. Exemplars varied in
their typicality. Essentialist thinking is reﬂected in absolute yes/no
Kalish (2002),Diesendruck and
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 5
metric, all three groups endorsed common ancestry with apes (binomial p< 0.001). Eighth graders and non-majors both
denied common ancestry with humans for all remaining categories (binomial p< 0.001). In contrast, biology majors
endorsed common ancestry with mammals (binomial p< 0.001) showed no consensus on common ancestry for dinosaurs,
birds, reptiles or insects (binomial p> 0.336), and denied common ancestry with plants (binomial p= 0.008).
Second, we ran 3 (Group) 2 (Response) Chi Square analyses for each target category to compare responses among the
three groups (see Fig. 1B). For apes, groups did not differ (X
(2, 208) = 2.07, p= 0.355) because virtually all participants
acknowledged that humans and apes shared a common ancestor. For all other targets, biology majors endorsed shared com-
mon ancestry more frequently than non-majors or 8th graders (X
(2, 208) > 14.55, p60.001). When biology majors were
excluded from the analysis, non-majors and 8th graders did not differ in their endorsement of common ancestry for mam-
mals or plants (X
(1, 139) < 1.94, p> 0.165). Although afﬁrming common ancestry for other targets (dinosaur, bird, amphib-
ian, insect) was quite rare, non-majors nevertheless did so more frequently than 8th graders (X
(1, 139) > 4.22, p< 0.040).
3.1.2. Property projection
For this measure, participants rated the likelihood that a novel gene found in humans would also be found in a range of
organisms varying in their degree of perceived similarity to humans (gorilla, deer, squirrel, sparrow, frog, grasshopper, daisy).
Participants indicated their estimate of the likelihood that each target species would share a gene with humans on a scale of
1 (very unlikely) to 7 (very likely). Reasoning grounded in human uniqueness indicating anthropocentric thinking would be
manifest in an unwillingness to attribute a human biological property to other organisms. To index overall anthropocentric
thinking on this measure, we examined mean likelihood ratings for the three groups via one-way ANOVA, which revealed
that groups differed in their mean likelihood ratings, F(2, 209) = 10.79, p< 0.001,
= 0.09, see Fig. 2A. Speciﬁcally, biology
majors gave higher ratings on average than non-majors (Tukey HSD p= 0.043) or 8th graders (Tukey HSD p< 0.001). Mean
ratings for non-majors were marginally higher than for 8th graders (Tukey HSD p= 0.078).
To examine response patterns in more detail, we compared mean likelihood ratings via 3 (Group) 7 (Target) mixed
ANOVA. Several ﬁndings emerged in addition to the overall group differences discussed above. First, mean ratings varied
by target species, F(6,1218) = 443.74, p< 0.001,
= 0.67. To examine speciﬁc response patterns, we arranged targets in
order of perceived similarity to humans and compared mean likelihood ratings for adjacent targets via t-tests.
for six comparisons, we used a critical p-value of 0.008 for each comparison. Using this method, mean ratings for each target
differed from those for adjacent targets, except for sparrow and frog, which did not differ.
Finally, the effects of group and target were qualiﬁed by a group target interaction, (F(12, 1218) = 7.57, p< 0.001,
= 0.02). To explore this interaction, we conducted t-tests comparing mean likelihood ratings for each species with the
adjacent species separately for each group, using the modiﬁcation described above to control for multiple comparisons.
As depicted in Fig. 2B, inference patterns were highly similar, with the exception that for 8th graders, inferences were some-
what less differentiated than for undergraduates.
In sum, even though biology majors were more willing to project a novel gene from humans to nonhumans than non-
majors or 8th graders, patterns of projection to different targets were remarkably similar across the three groups, as indi-
cated by a linear decrease from gorilla through daisy.
Fig. 1. Anthropocentric thinking, common ancestor task. (A) Mean number of organisms said to share a common ancestor with humans. (Note: Error bars
correspond to 95% conﬁdence intervals). (B) Percentage of positive responses for each target species on the common ancestor task for 8th graders, non-
majors, and biology majors. Note: bars with diagonally-lined ﬁll did not differ from chance (50%) via binomial test.
Although squirrels are phylogenetically more closely related to humans than deer are (e.g., Asher, Bennett, & Lehmann, 2009; Madsen et al., 2001), pilot
testing revealed that participants generally perceive deer to be more similar to humans than squirrels.
6J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
3.1.3. Biological similarity
This measure consisted of ﬁve triplets consisting of pictures of a human and two other species. Participants indicated
which two were most biologically similar. In each case, the normative answer was to choose the human and one non-
human. Responses were scored as correct (pairing the human with the more closely related non-human species), anthro-
pocentric (pairing the two non-humans together), or other. To index overall anthropocentric thinking on this measure, we
simply summed the number of anthropocentric responses for each participant, out of a possible ﬁve items (see Fig. 3A),
and compared groups via one-way ANOVA. On this analysis, biology majors made fewer anthropocentric responses than
8th graders, and marginally fewer than non-majors, F(2, 208) = 5.30, p= 0.006,
= 0.05, Tukey HSD p= 0.005 and
p= 0.062, respectively. Non-majors and 8th graders did not differ on this measure.
To look at responses in more detail, we classiﬁed each participant with respect to a consistent response pattern across all
ﬁve items, based on whether the majority of responses (three or more out of ﬁve) were correct or anthropocentric. The few
participants who showed neither consistent pattern were classiﬁed as ‘‘other.” A 3 (Group) 3 (Response Pattern) Chi
Square analysis revealed group differences in the distribution of response patterns, X
(4, 211) = 12.52, p= 0.014. The distri-
bution of response patterns is presented in Fig. 3B, which shows that anthropocentric responding was the modal pattern for
non-majors and 8th graders, whereas correct responding was the modal pattern for biology majors. As with the previous
task, this result suggests a more normative understanding of biological similarity among biology majors, but also a strong
residual inﬂuence of anthropocentric thinking; less than half of biology majors showed a correct response pattern, and
38% demonstrated an anthropocentric response pattern.
3.1.4. Relations among measures of anthropocentric thinking
To examine relations among different measures of anthropocentric thinking, we computed standardized scores for sum of
common ancestors, mean likelihood ratings for projections from humans, and number of anthropocentric responses for each
participant. We reverse-scored the ﬁrst two measures so that larger scores on all measures would ostensibly reﬂect more
anthropocentric thinking (i.e., unwillingness to acknowledge common ancestry, unwillingness to project a human property
to nonhumans, and anthropocentric construals of biological similarity). These measures were consistently intercorrelated;
speciﬁcally, anthropocentric responding on the common ancestor task predicted anthropocentric responding on both the
property projection task (r(207) = 0.40, p< 0.001) and on the biological similarity task (r(208) = 0.26, p< 0.001), and anthro-
pocentric responding on the latter two tasks was also correlated, r(210) = 0.22, p= 0.001. This suggests that anthropocentric
thinking may be a single underlying construct.
3.2. Teleological thinking
To examine teleological thinking, we asked participants to rate their agreement with appropriate teleological explanations
(e.g., ‘‘Pencils exist so that people can write with them”), inappropriate teleological explanations (e.g., ‘‘Houses have doorbells
in order to make dogs bark”) and natural kind teleological explanations (e.g., ‘‘Rain falls in order to allow plants to grow”).
Each participant’s ratings for appropriate, inappropriate, and natural kind explanations were averaged resulting in a score
ranging from 1 (disagree) to 6 (agree) for each kind of statement. To examine the general degree of teleological thinking
on this measure, we averaged agreement ratings across all three types of items, and compared groups on these scores via
one-way ANOVA (see Fig. 4A). On this analysis, biology majors were less likely to agree with statements than non-majors,
who were less likely to agree than 8th graders, F(2,208) = 21.03, p< 0.001,
= 0.17, all differences p< 0.017 via Tukey
Fig. 2. Anthropocentric thinking, property projection task. (A) Mean likelihood, averaged across targets, that a target species would share a hypothetical
gene with humans for 8th graders, non-majors, and biology majors. (B) Mean likelihood that each target species would share a hypothetical gene with
humans for 8th graders, non-majors, and biology majors. Note: error bars represent 95% conﬁdence intervals; points connected by dotted lines differed
reliably (LSD p< 0.008); points connected by solid lines did not.
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 7
To examine response patterns in more detail, we compared mean agreement ratings via 3 (Group) 3 (Statement Type)
mixed ANOVA. Several ﬁndings emerged in addition to the overall group differences discussed above. First, participants
strongly agreed with appropriate statements and strongly disagreed with inappropriate statements; agreement for natural
kind statements fell between the two (F(2, 416) = 1646.01, p< 0.001,
= 0.88, all differences signiﬁcant at p< 0.001 via
Bonferroni-corrected t-test). Second, the effects of group and statement were qualiﬁed by an interaction, F(4, 416) = 10.01,
= 0.01 (see Fig. 4B). To examine this interaction, we conducted separate one-way ANOVAs comparing group per-
formance separately for each type of question. Based on these tests, there were no group differences in agreement with
appropriate items (F(2,208) = 1.88, p= 0.155). In contrast, although mean ratings were all quite low, 8th graders agreed more
with inappropriate items than biology majors or non-majors, who did not differ (F(2, 208) = 11.18, p< 0.001,
= 0.10, Tukey
HSD p< 0.027). Non-majors and 8th graders did not differ in their agreement with natural kind items, and both exhibited
higher agreement than biology majors (F(2,208) = 23.80, p< 0.001,
= 0.19, Tukey HSD p< 0.001). Most notably, for all
three groups, agreement for natural kind items was higher than agreement with inappropriate items (tP9.41, p< 0.001,
dP1.13), indicating some endorsement for teleological thinking about natural phenomena.
To compare responses on different measures of teleological thinking, we conducted Pearson correlations on these scores.
Results show consistent interrelations among the three measures. The strongest correlation was between agreement with
teleological explanations for natural kinds and inappropriate teleological explanations, r(211) = 0.50, p< 0.001. Agreement
with appropriate teleological explanations was also positively correlated with agreement with inappropriate teleological
explanations (r(211) = 0.19, p= 0.005) and agreement with teleological explanations for natural kinds (r(211) = 0.22,
p= 0.001). The relatively strong positive correlation between agreement with inappropriate and natural kind teleological
explanations is particularly informative, in that both are incorrect from a scientiﬁc point of view, but may indicate a stable
underlying propensity toward teleological thinking.
Fig. 3. Anthropocentric thinking, biological similarity task. (A) Mean number of anthropocentric choices on the biological similarity task. Note: error bars
represent 95% conﬁdence intervals. (B) Percentage of 8th graders, non-majors and biology majors showing correct, anthropocentric, and other response
patterns across all ﬁve items for the biological similarity task.
Fig. 4. Teleological thinking, agreement with explanations task. (A) Overall mean agreement with teleological statements for 8th graders, non-majors and
biology majors. (B) Mean agreement with appropriate, inappropriate and natural kind teleological statements for 8th graders, non-majors and biology
majors. Note: error bars represent 95% conﬁdence intervals.
8J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
3.3. Essentialist thinking
3.3.1. Category homogeneity
Participants were told that an individual member of a species (ant, grasshopper, giraffe, kangaroo) had a hypothetical
intrinsic (X cells, gene X) or extrinsic (parasite X, bacteria X) property, and asked what percentage of all members of that
species would share the property. To index essentialist thinking, we calculated the mean estimate for percentage of category
members sharing an intrinsic property (novel cells or genes). One-way ANOVA revealed that groups differed in their homo-
geneity estimates, F(2,190) = 4.22, p= 0.016,
= 0.04, see Fig. 5A. Speciﬁcally, non-majors estimated that a higher percent-
age of species members would share an intrinsic property than 8th graders (Tukey HSD p= 0.014), whereas biology majors
did not differ from either group (Tukey HSD p> 0.155).
To examine responses in more detail, and because homogeneity estimates in isolation are hard to interpret in an absolute
sense, we compared estimates for intrinsic versus extrinsic properties using a 3 (Group: 8th graders, non-majors, biology
majors) 2 (property type: intrinsic, extrinsic) 2 (Form: A, B) mixed ANOVA. We included form as a factor because the
exemplars of intrinsic and extrinsic properties differed for forms A (cells and parasites, respectively) and B (genes and infec-
tion). Overall, as expected, homogeneity estimates were higher for intrinsic properties than extrinsic properties, F(1,179) =
110.18, p< 0.001,
= 0.36. However, we also observed a property group interaction (F(2,179) = 5.31, p= 0.006,
a main effect of form (F(1,179) = 14.07, p< 0.001,
= 0.07), and group form interaction (F(2, 179) = 3.25, p= 0.041,
= 0.03). Together, these suggest different responses for speciﬁc properties among the three groups. We explored this pos-
sibility via separate Form Property Type ANOVAs for each group (see Fig. 5B).
For biology majors and non-majors, patterns were identical: homogeneity estimates for genes and cells were equivalent,
and higher than for infections and parasites, which were also equivalent. This was afﬁrmed by large main effects of property
(non-majors: F(1,62) = 57.76, p< 0.001,
= 0.48; biology majors: F(1, 50) = 38.06, p< 0.001,
= 0.43), no differences
between forms, and no interaction. In contrast, for 8th graders, estimates for cells were higher than for the other three prop-
erties, which were equivalent (see Fig. 5B). This was afﬁrmed by main effects of property (F(1, 67) = 18.22, p< 0.001,
= 0.20), form (F(1, 67) = 19.41, p< 0.001,
= 0.22), and an interaction between the two (F(1,67) = 4.49, p= 0.038,
= 0.05). Lastly we compared homogeneity scores for all three groups on each property. No group differences were
observed for cells,infection,orparasites (F< 1.20, p> 0.307). In contrast, for gene, 8th graders’ scores were lower than both
undergraduate groups, which did not differ (F(2, 90) = 7.55, p= 0.001,
= 0.14, Tukey HSD p< 0.005).
To summarize, consistent with essentialist thinking, all groups perceived the target living kinds as more homogeneous
with respect to intrinsic properties like cells than extrinsic properties like parasites or infections. Groups did not differ in per-
ceived homogeneity for cells, but, unexpectedly, 8th graders estimated much lower homogeneity for genes than for cells.In
contrast, homogeneity estimates for gene and cell did not differ for biology majors or non-majors.
3.3.2. Innate potential
To probe individuals’ thinking about innate potential, participants were given a scenario about a human infant born to
one set of parents but raised by another, and asked which set of parents the child would more closely resemble. Responses
indicating that the child would share characteristics with the birth parents, indicate essentialist thinking. To index essential-
ist thinking, we calculated the mean number of times (out a possible ﬁve) that participants said the child would more closely
resemble the birth parents. One-way ANOVA revealed that groups did not differ in their responses, F(2,208) = 1.98, p= 0.141,
= 0.02, see Fig. 6A.
To examine the data in more detail, we calculated the percentage of responses indicating resemblance to birth parents for
each type of trait. Overall, participants’ responses about the likelihood of resemblance to birth parents varied markedly by
trait, ranging from 93% for physical traits to 6% for beliefs (see Fig. 6B). We examined these data in two ways. First, we com-
pared each cell to chance performance (50%) via binomial test; above-chance performance suggests consensus that a trait is
inherited—and thus subject to innate potential—whereas below-chance performance indicates that a trait is environmentally
determined—and thus not subject to innate potential. Chance performance reﬂects lack of consensus. On this metric, all three
groups believed that children were likely to resemble birth parents with respect to physical traits (p < 0.001) and abilities
(p 60.008) and to resemble adoptive parents with respect to preferences (p 60.001) and beliefs (p < 0.001). For personality
traits, beliefs were more equivocal. 8th graders and biology majors believed that children were likely to resemble adoptive
parents on personality traits (p < 0.05), whereas non-majors responses as a whole didn’t differ from chance (p = 0.904).
Second, to compare groups on speciﬁc types of traits, we conducted, a series of 3 (Group) 2 (Response) Chi Square anal-
yses for each type of trait. Only physical traits were close to signiﬁcantly different, X
(2, 211) = 5.75, p= 0.057.
large majority of all participants thought that a child would resemblance its birth parents on physical traits, the proportion was
smaller for 8th graders (88%) than for biology majors (97%) or non-majors (96%). Group differences for other traits did not
approach signiﬁcance, X
(2) < 4.27, p> 0.118. In sum, participants shared very similar reasoning about inherent potential for
this set of traits.
This result should be interpreted with caution because 50% of the cells had an expected cell frequency of less than 5.
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 9
3.3.3. Identity over transformation
To assess individuals’ thinking about the effects of transformations on category identity, participants were given two ﬁc-
titious scenarios in which an animal underwent a transformation in appearance and behavior. In one, the transformation was
a result of natural developmental process (akin to metamorphosis). In the other, the transformation was an artiﬁcial one trig-
gered by exposure to toxic contamination. For each item, participants indicated on a scale whether the animal was biolog-
ically a member of the pre-transformation category or the post-transformation category both before the transformation and
after the transformation.
To index essentialist thinking, we computed change scores for each participant by subtracting their pre-transformation
category rating from their post-transformation category ratings. Thus, higher scores represent larger perceived effects of the
transformation on category membership, whereas lower scores reﬂect more essentialist thinking. We conducted a 2 (Trans-
formation: natural, artiﬁcial) 3 (Group) mixed ANOVA on these change scores. Results revealed that 8th graders’ change
scores were larger than those of biology majors (F(2,203) = 5.16, p= 0.007,
= 0.05, Tukey HSD p = 0.004; non-majors scores
were intermediate between these, and differed from neither (see Fig. 7A). No effects of transformation type approached sig-
niﬁcance. One sample t-tests revealed that mean change scores for all three groups differed from zero, suggesting that all
three groups readily accepted that transformations had an effect on category membership, (8th graders: t(73) = 13.24,
p< 0.001, d= 1.54; non-majors: t(67) = 11.50, p< 0.001, d= 1.39; biology majors: t(67) = 10.49, p< 0.001, d= 1.27).
We further examined the degree of change between pre- and post-transformation ratings by dividing the response scale
into three categorical responses indicating reasoning that the creature is a member of the pre-transformation category (1–3),
reasoning that the creature is intermediate between the two categories (4–6), and reasoning that the creature is a member of
the post-transformation category (7–9). We then characterized each participant’s response to both pre- and post-
transformation questions together as demonstrating category preservation (i.e., falling into the same response category
Fig. 5. Essentialist thinking, category homogeneity task. (A) Mean homogeneity estimates for intrinsic properties (novel cells or genes) for 8th graders, non-
majors and biology majors. (B) Mean estimates of percentage of a species sharing hypothetical intrinsic properties (cells, genes) and extrinsic properties
(parasite, infection) for 8th graders, non-majors and biology majors. Note: Error bars represent 95% conﬁdence intervals.
Fig. 6. Essentialist thinking, innate potential task. (A) Mean number of traits that 8th graders, non-majors and biology majors predicted that a child would
share with birth parents. (Note: Error bars represent 95% conﬁdence intervals.) (B) Percentage of 8th graders, non-majors and biology majors indicating
resemblance to birth parents for each type of trait. Note: bars with diagonallylined ﬁll did not differ from chance (50%) via binomial test.
10 J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
pre- and post-transformation), category shift (i.e., moderate change, from pre-transformation category to intermediate, or
intermediate to post-transformation category), or outright category change (from pre- to post-transformation category).
For natural transformations, 38% of responses were consistent with essentialist thinking and involved category preservation,
19% demonstrated a shift in perceived category membership, and 43% reﬂected actual category change. A 3 (Group) 3
(Response Pattern) Chi Square analysis revealed no group differences in response patterns, X
(4, 208) = 3.31, p= 0.508;
see Fig. 7B. For artiﬁcial transformations, 40% of responses involved category preservation, 20% demonstrated a shift in per-
ceived category membership, and 40% reﬂected category change. A 3 (Group) 3 (Response Pattern) Chi Square analysis
revealed group differences in response patterns, (X
(4, 207) = 11.52, p= 0.021. As depicted in Fig. 7B, each group showed
a distinctive response pattern. For 8th graders, category change was the modal response, for non-majors, preservation
and change were equally frequent, and for biology majors, category preservation was the modal response.
In sum, reasoning about transformations presented a mixed picture of essentialist thinking. Mean ratings suggest that
participants readily accepted that both natural and artiﬁcial transformations changed category membership in these scenar-
ios. Response patterns suggests that preservation of category membership was about as common as change of category
membership, and that category shifts were more rare. Group differences were minimal, but suggested that 8th graders were
more willing to accept change in category membership, and hence can be characterized as less essentialist, than
3.3.4. Absolute category membership
This task examined reasoning about whether membership in intuitive biological categories (bird, ﬁsh, tree) is absolute
(e.g., a given species must necessarily either be a bird or not be a bird) or graded (something could ‘‘sort of” be a bird). Each
participant was given a score corresponding to the percentage of known exemplars to which they gave absolute responses
(yes or no). As expected, scores were relatively high (see Fig. 8A). Nevertheless, both undergraduate groups were more abso-
lute in their category judgments than the 8th graders, F(2, 208) = 17.16, p< 0.001,
= 0.14, Tukey HSD p< 0.001. Biology
majors and non-majors did not differ. This suggests stronger essentialist reasoning among undergraduates than among
8th graders with respect to category boundaries.
To examine responses in more detail, we conducted a 3 (Group) 3 (Category: Birds, Fish, Trees) mixed ANOVA on per-
centage of absolute category membership judgments. In addition to the overall group differences mentioned above, we also
observed a group category interaction, F(4,416) = 3.04, p= 0.017,
= 0.02, see Fig. 8B. We explored this interaction via
one-way ANOVAs comparing the three groups’ responses for each category. For ﬁsh and trees, as in the overall analyses, biol-
ogy majors and non-majors did not differ, and both gave a higher proportion of absolute responses than 8th graders (ﬁsh:
F(2,208) = 15.17, p< 0.001,
= 0.13, Tukey HSD p< 0.001; trees: F(2, 208) = 6.38, p= 0.002,
= 0.06, Tukey HSD p< 0.022).
For birds, biology majors gave a higher proportion of absolute responses than 8th graders (F(2,208) = 3.28, p= 0.040,
= 0.03, Tukey HSD p= 0.031) but non-majors did not differ from either group (Tukey HSD p> 0.303).
3.3.5. Relations among measures of essentialist thinking
As a ﬁnal analysis, we looked for interrelations among the ﬁve measures of essentialist thinking. To do so we ran corre-
lation analyses on the indices of essentialist thinking for each measure: standardized scores for homogeneity estimates for
intrinsic properties, sum of ‘‘birth parent” responses, percentage of absolute category judgments, and mean transformation
difference scores (reverse-scored so that larger scores on all measures would reﬂect more essentialist thinking). For the most
Fig. 7. Essentialist thinking, transformation task. (A) Mean pre-post transformation change scores averaged over natural and artiﬁcial transformations.
(Note: Error bars correspond to 95% conﬁdence intervals). (B) Percentage of participants in each group showing category preservation, shift, and change
patterns for natural and artiﬁcial transformation scenarios.
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 11
part, scores on these measures were unrelated, suggesting that essentialist thinking may not be a monolithic construct;
results revealed a single signiﬁcant correlation between essentialist responding on the transformation task and essentialist
responding on the absolute category membership task, r(210) = 0.158, p= 0.022.
3.4. Composite analyses
To construct composite measures for essentialist, anthropocentric, and teleological thinking, we computed standardized
scores for all participants on each individual measure. We then averaged the scores for all measures corresponding to each
construal, to obtain a single score for each participant on each construal-based set of measures. This score represents the
degree to which an individual’s responses to a set of measures reﬂects the corresponding cognitive construal, relative to
all other participants. Finally, we compared 8th graders, non-majors, and biology majors on each composite score in order
to assess the relative strength of each type of thinking for each group. Scores are depicted in Fig. 9.
The anthropocentric thinking composite measure was composed of an average of standardized scores representing (1)
number of common ancestors (reverse scored), (2) mean likelihood rating for projection from humans (reverse scored), and
(3) number of anthropocentric responses to triad items. One-way ANOVA showed that groups differed in their degree of
anthropocentric responding, F(2,208) = 22.56, p< 0.001,
= 0.18. Speciﬁcally, biology majors show less anthropocentric
responding than either non-majors or 8th graders (Tukey HSD p> 0.001). Non-majors and 8th graders differed only margin-
ally (Tukey HSD p= 0.068).
The teleological thinking composite measure was composed of an average of standardized scores representing mean
agreement ratings for (1) appropriate teleological explanations, (2) inappropriate teleological explanations, and (3) teleolog-
ical explanations for natural phenomena. One-way ANOVA showed that groups differed in their degree of teleological
responding, F(2,208) = 17.99, p< 0.001,
= 0.15. Speciﬁcally, 8th graders responses were more teleological than those of
non-majors (Tukey HSD p= 0.014), whose responses were more teleological than biology majors (Tukey HSD p= 0.007).
The essentialist thinking composite measure was composed of an average of standardized scores representing (1) esti-
mates of homogeneity for intrinsic properties, (2) number of ‘‘birth parent” responses for switched-at-birth items, (3) abso-
lute difference score for pre-post transformation ratings (averaged over natural and artiﬁcial transformations and reverse
scored), and (4) percentage of absolute category judgments. One-way ANOVA showed that groups differed in their degree
of essentialist responding, F(2,208) = 17.01, p< 0.001,
= 0.14. Strikingly, 8th graders’ responses were less essentialist than
those of biology majors or non-majors (Tukey HSD p< 0.001). Biology majors and non-majors didn’t differ in essentialist
responding (see Fig. 9).
In addition to examining group differences, we also wanted to ascertain whether these construals reﬂect independent
components of intuitive biological thought (as in Coley & Tanner, 2015), or whether they were interrelated. To do so, we
computed correlations between the three composite measures described above. These were computed separately for each
age group, so that systematic group differences did not artiﬁcially inﬂuence correlations. Results are depicted in Table 2,
and suggest that indeed, anthropocentric, teleological, and essentialist thinking are independent components of intuitive
biological thought. The only signiﬁcant correlation was a negative one between essentialist thinking and anthropocentric
thinking for 8th graders; no other correlations approached signiﬁcance.
In sum, composite analyses reveal unique patterns of group differences for each intuitive cognitive construal. Speciﬁcally,
8th graders and non-majors showed comparable levels of anthropocentric thinking, whereas biology majors showed much
less. In contrast, all three groups differed in their degree of teleological thinking, with 8th graders showing the most and biol-
ogy majors showing the least. Interestingly, although differences in essentialist thinking were least pronounced, biology
Fig. 8. Essentialist thinking, absolute category membership task. (A) Overall percentage of absolute category judgments for each group. (B) Percentage of
absolute category judgments for bird, ﬁsh, and tree categories for each group. Note: Error bars correspond to 95% conﬁdence intervals.
12 J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
majors and non-majors showed comparable levels of essentialist thinking, and both showed stronger essentialist tendencies
than 8th graders. Finally, the lack of correlations among composite measures suggests that the three construals represent
independent components of intuitive biological thought.
In this study we examined the nature of intuitive biological thinking in adolescence and early adulthood, and investigated
the extent to which increasing amounts of biology education might inﬂuence intuitive biological thinking. To do so, we
employed multiple measures to assess the prevalence of anthropocentric, teleological, and essentialist thinking in 8th grade
middle school students, ﬁrst-year undergraduate biology majors who had earned Advanced Placement credit in biology, and
undergraduates who were not biology majors and had earned Advanced Placement credit in a ﬁeld other than biology. To our
knowledge, this work is unique with respect to both the scope of measures utilized and the populations compared.
Several ﬁndings are of particular note. First, we ﬁnd clear evidence of intuitive reasoning among all populations studied.
Thus, our results ﬁrmly establish the persistence of intuitive biological reasoning—in the form of multiple cognitive constru-
als—from 8th grade into young adulthood. Given that these ﬁndings emerged across a broad range of measures, it suggests
that intuitive reasoning is not simply a characteristic of a particular task or probe. For those who have primarily investigated
cognitive development among younger age groups, these data may at ﬁrst appear surprising. However, in recent years, a
variety of research both in cognitive psychology (e.g. Kelemen et al., 2013; Shtulman & Harrington, 2016; Shtulman &
Schulz, 2008; Shtulman & Valcarcel, 2012) and in science education (e.g. Coley & Tanner, 2015; Nehm & Reilly, 2007) have
begun to provide evidence that intuitive understandings may persist into adulthood.
Second, although we observed consistent differences between biology majors, non-majors, and 8th graders on measures
of intuitive biological thought, these differences were surprisingly small given the differences in age, general cognitive devel-
opment, and the increasing amounts of science education among our participant groups. For the most part, all three groups
showed the same qualitative response patterns on tasks, and differences tended to be markedly smaller than effects of
within-measure manipulations. This suggests that more extensive biology education may have relatively minimal inﬂuence
Fig. 9. Composite scores summarizing relative performance on measures of anthropocentric, teleological, and essentialist thinking for 8th graders, non-
majors, and biology majors. Error bars represent 95% CI.
Correlations among composite scores of anthropocentric, teleological, and essentialist
Group Measures Pearson’s r, p-value
8th Graders Anthropocentric—teleological r(74) = 0.14, p= 0.223
Teleological—essentialist r(74) = 0.05, p= 0.679
Essentialist—anthropocentric r(74) = 0.31, p= 0.007
Non-majors Anthropocentric—teleological r(68) = 0.18, p= 0.140
Teleological—essentialist r(68) = 0.12, p= 0.312
Essentialist—anthropocentric r(68) = 0.00, p= 0.981
Biology majors Anthropocentric—teleological r(69) = 0.14, p= 0.241
Teleological—essentialist r(69) = 0.06, p= 0.634
Essentialist—anthropocentric r(69) = 0.14, p= 0.241
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 13
on intuitive biological thinking. Below we discuss these ﬁndings in more detail for each of our target cognitive construals. We
then brieﬂy discuss implications and conclusions.
4.1. Anthropocentric thinking
Results from our measures of anthropocentric thinking show that both undergraduate groups, as well as 8th graders,
responded in ways that suggest an underlying reasoning grounded in human exceptionalism. Consistent with anthropocen-
tric thinking, participants in all three groups rarely attributed common ancestry with humans to any species apart from great
apes, attributed novel genes from humans to non-humans only if the non-humans were perceived as sufﬁciently similar to
humans, and overestimated human taxonomic uniqueness.
Although all three groups demonstrated anthropocentric thinking, group differences were also evident; whereas 8th gra-
ders and non-majors showed few differences in anthropocentric thinking, both groups evinced more than biology majors. How-
ever, effect sizes for group differences were relatively small compared to effect size for within-task manipulations. For example,
in the property projection measure of anthropocentric thinking, the effect size for the difference between target categories was
0.67, whereas the effect sizes for the difference between groups and the interaction were 0.09 and 0.02, respectively. This
reveals that the common reasoning among all participants that species differ in their likelihood of sharing genes with humans
was much stronger than difference among groups with respect to degree of likelihood or pattern of judgments.
Biology majors appeared to employ anthropocentric reasoning, in spite of their more extensive formal training in biology.
For example, while 8th graders and non-majors asserted common ancestry between just humans and primates, biology
majors went somewhat further in asserting common ancestry with both primates and mammals. However, biology majors
were still far from asserting common ancestry of humans with all living creatures (see Fig. 2). Less than half of biology majors
asserted that humans shared a common ancestor with dinosaurs, birds, amphibians, insects, or plants, and qualitative drops
in acknowledgement of shared ancestry with phylogenic distance were similar among all groups. These results are surprising
given that a core tenet of evolution, which is central in formal biology education, is the common ancestry of all living organ-
isms. However, these results on the common ancestry task suggest that anthropocentric reasoning persists despite intensive,
college-level biology training.
Taken together, these results support prior research indicating that anthropocentric thinking persists into young adult-
hood (e.g., Coley, 2007; Goldberg & Thompson-Schill, 2009; Johnson et al., 1992) and extend it by showing that increasing
amounts of biology education may have only a minimal effect on such thinking.
4.2. Teleological thinking
Results of our measure of teleological thinking revealed evidence of this cognitive construal in all three groups. Speciﬁ-
cally, all groups agreed more with teleological statements about natural kinds than with inappropriate statements, suggest-
ing some willingness to accept technically incorrect teleological explanations for natural phenomena.
Again, group differences were evident; 8th graders showed more teleological thinking than non-majors, who showed
more than biology majors. Differences between 8th graders and undergraduates suggest a general reduction in the scope
of what are viewed as appropriate teleological explanations with development (i.e., undergraduates are less teleologically
promiscuous than 8th graders). Differences between non-majors and biology majors suggests that more intensive biology
education may also play an additional role in reducing teleological thinking. These ﬁndings support Kelemen’s argument that
teleological thinking persists but becomes less ‘‘promiscuous” with development (e.g., Kelemen & Rossett, 2009; Kelemen
et al., 2013). Nevertheless, group differences were again relatively small. The effect size for differences in agreement with
different classes of teleological explanation was 0.88, whereas the effect size for the difference between groups and the inter-
action were 0.17 and 0.01, respectively, showing that the common thinking among all participants about the differences in
acceptability of different teleological explanations was much stronger than differences among groups in agreement with
Biology majors also demonstrated use of teleological reasoning to explain natural phenomena. Although they were less
likely than non-majors to agree with scientiﬁcally inaccurate ‘‘natural kind” teleological statements, biology majors were
more accepting of these statements than they were of inaccurate teleological statements that were about general, non-
biological topics, suggesting some persistence of teleological reasoning about natural phenomena. These results suggest that
increasing amounts of biology education may decrease—but not eliminate—teleological reasoning. This is consistent with the
hypothesis that persistence of teleological thinking into the college years may underlie systems of biological misconceptions,
namely inaccurate biological reasoning that may cross disparate topic areas such as physiology, evolution, and molecular
biology, yet have in common threads of teleological reasoning (Coley & Tanner, 2015; Kelemen & Rossett, 2009).
4.3. Essentialist thinking
Essentialist thinking was also evident across age and biology education; participants in all three groups expected biolog-
ical kinds to be more homogeneous with respect to intrinsic properties than extrinsic properties, believed that biological
properties and abilities were likely to be driven by nature rather than nurture, and thought that membership in most bird,
ﬁsh, and tree categories we asked about was absolute.
14 J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
Surprisingly, 8th graders demonstrated lower levels of essentialist thinking than biology majors or non-majors, who did
not differ. This suggests that essentialist reasoning about biology may actually increase with development through adoles-
cence, and may do so relatively independently of increasing amounts of biology education. Previous research suggests that
essentialist thinking may become increasingly focused on the domain of biology with development (e.g., Taylor et al., 2009);
because we focused exclusively on intuitive biological reasoning, our results may reﬂect that tendency.
As with the other construals, group differences were relatively small. For example, for the category homogeneity measure
of essentialist thinking, the effect size for the difference between student groups was 0.04, yet the effect size for the differ-
ences between intrinsic and extrinsic properties was 0.36, an order of magnitude larger. This shows that the common rea-
soning among all participants that intrinsic properties like cells were likely to be more homogenous than extrinsic properties
like parasites and infections far outweighed the difference among groups with respect to the degree of homogeneity of genes.
On multiple measures of essentialist thinking, biology majors were statistically indistinguishable from their college-age
peers with less intensive and advanced biology education experiences. For example, the innate potential measure (switched-
at-birth task) probed students’ predictions about whether a child’s characteristics – physical characteristics, abilities, and
beliefs – would be more like their birth parents or their adoptive parents. Results on this essentialist measure were strikingly
similar between the two college-age populations under study. This is surprising given that advanced, college-level biology
education would likely include explorations of epigenetics—the impact of the environment on gene expression and inherited
traits. Indeed, one might have predicted that biology majors—presumably with some prior exposure to the concept of gene
expression and the interplay of the environment with genetics—might be less willing than non-majors to endorse a strong
‘‘nature” view of a child’s properties. For example, over two-thirds of all college-age students, including biology majors,
asserted that a child’s aptitude for particular sports (soccer v swimming) or academic subject (English v math) would resem-
ble that of their birth parents, which reﬂects minimal attention to environmental effects on growth, development, and learn-
ing. Even for physical traits, a child’s height would be dependent not only on genetics, but also on environmental variables
such as available nutrition. Yet, over 90% of all college-age students asserted that a child’s height would be more similar to
their birth parents than their adoptive parents (see also Ware & Gelman, 2014). These results suggest that intensive, college-
level biology education has minimal inﬂuence on essentialist reasoning, resulting in biology majors that are highly similar in
their essentialist reasoning to non-majors, even in biological contexts.
These ﬁndings raise questions about relations between increasing amounts of biology education and essentialist thinking.
Unlike anthropocentric and teleological thinking, essentialist thinking was higher among college students than 8th graders,
and did not differ for biology majors versus non-majors. One possible interpretation of this ﬁnding is that—unlike anthro-
pocentric and teleological thinking—in some respects essentialist thinking may be consistent with normative scientiﬁc bio-
logical understanding. For example, one might argue that essentialist responses such as maintaining category identity over
transformations, or believing that a hypothetical gene is likely to be shared widely among a species, or believing that mem-
bership in biological categories is absolute, are biologically correct. Alternatively, essentialist thinking may lead to percep-
tions of category homogeneity, and corresponding underestimates of natural variability, contrary to what Mayr (1982) calls
‘‘population thinking.” In other words, the (correct) understanding that biological species have a genetic basis can—when
coupled with a penchant for essentialist thinking—may give rise to the incorrect supposition that a particular gene is likely
to be widely shared within a species. Likewise, essentialist thinking may lead to perceptions of sharp, discrete category
boundaries among classes of living things, which is inconsistent with the idea that all living things are related via descent
from a common ancestor (e.g., O’Hara, 1998; for further discussion of these issues, see Coley & Muratore, 2012; Gelman &
Rhodes, 2012). In sum, relations between essentialist thinking and scientiﬁc understanding may be quite complex.
Finally, we found that measures of essentialist thinking were less cohesive than measures of anthropocentric or teleolog-
ical thinking; individual measures were largely independent of each other. We see two implications of this ﬁnding. First, it
raises questions about conceptualizing essentialist thinking as a unitary construct. The work of Haslam and colleagues (e.g.,
Haslam, Rothschild, & Ernst, 2000) in the realm of social categorization suggests that essentialist thinking may be proﬁtably
analyzed into distinct subcomponents of naturalness—the degree to which concepts are seen as objective, discreet and stable
– and cohesiveness (‘‘entitativity”)—the degree to which concepts are seen as homogeneous and informative. Perhaps essen-
tialist thinking in intuitive biology might be proﬁtably analyzed into similar orthogonal subcomponents. Although our mea-
sures were not designed to examine this possibility directly, our ﬁnding that innate potential (arguably related to category
naturalness) is independent of perceived category homogeneity (a measure of cohesiveness) is consistent with this idea. Sec-
ond, this ﬁnding raises procedural questions about employing a single measure of essentialist thinking in studies of intuitive
biology. Typically, such studies have relied on a single measure of performance; however, our results demonstrate that
essentialist thinking may not be a unidimensional construct. At the very least, relations among measures of essentialist
thinking—and by proxy, among potential components of the construct—are ripe for further investigation.
4.4. Limitations and extensions
Although the results presented here are intriguing, the investigations need to be replicated among additional populations
at similar developmental and educational levels. The results presented were collected from single populations of students
from two educational institutions. Future studies should attempt to replicate these ﬁndings with students at other institu-
tions that are geographically and culturally diverse.
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 15
Relatedly, we do not claim that our list of construals is exhaustive; other cognitive construals relevant to intuitive bio-
logical thought most certainly exist. One excellent candidate is vitalist thinking (e.g., Hatano & Inagaki, 1994; Inagaki &
Hatano, 2002; Morris, Taplin, & Gelman, 2000). Nor do we claim that these three construals are universal. Although some
of the construals we examined might be relatively robust to cultural and contextual differences (Atran & Medin, 2008;
Casler & Kelemen, 2008; Rottman et al., 2016; Sousa, Atran, & Medin, 2002), others might well vary with culture or other
contextual factors (i.e., anthropocentric thinking: Herrmann, Waxman, & Medin, 2010; Medin, Waxman, Woodring, &
Washinawatok, 2010; Ross, Medin, Coley, & Atran, 2003; Waxman, Medin, & Ross, 2007); teleological thinking: ojalehto
et al., 2013). Indeed, both the content of intuitive biology and the patterns of developmental change might vary for different
populations. For now, these remain open, and important, questions.
It is also important to emphasize that biology majors in this study were in their ﬁrst semester of college biology. Although
they had all previously demonstrated achievement in biology through success on the AP biology exam—a commonly
accepted metric of understanding of basic biological principles which is often substituted for completion of a college-
level biology course—the present results should be taken as an examination of the impact of advanced, intensive biological
studies in high school on intuitive thinking. It may well be that more college-level coursework in the life sciences in a higher-
education context would lead to a continued reduction in the use of intuitive cognitive construals among more advanced
biology majors, and a corresponding reduction in similarity between majors and non-majors. In fact, analysis of performance
of disciplinary experts in biology – arguably an endpoint of cognitive development within biology as a conceptual area – may
provide an important touchstone for considering relations between intuitive biological reasoning and formal training in the
biological sciences. Although some recent work has revealed evidence of implicit effects of intuitive knowledge even among
highly trained practicing scientists (Goldberg & Thompson-Schill, 2009; Kelemen et al., 2013; Shtulman & Harrington, 2016),
it remains to be seen how the acquisition of formal disciplinary expertise in biological science affects performance on more
explicit tasks like those employed here.
An important related question raised by these ﬁndings is the degree to which biology majors hold multiple frameworks of
biology knowledge (e.g., Legare et al., 2012; Shtulman & Valcarcel, 2012). Perhaps biology majors possess the formal biology
knowledge to respond in a way that is less indicative of intuitive cognitive construals, but simply did not deploy such knowl-
edge in this context on our tasks. On the face of it, this seems somewhat unlikely, because many of the items were transpar-
ently biological in nature, and because biology majors were tested during their biology classes. However, future
investigations are necessary to differentiate these possibilities.
By systematically utilizing a broad array of measures of intuitive biological reasoning, we have shown clear evidence of
persistent, explicit intuitive reasoning among college students as well as 8th graders. Differences among student groups,
although consistently signiﬁcant, were small and tended to be markedly smaller than effects of within-measure manipula-
tions for each task. Perhaps more surprisingly, increasing amounts of formal biology education had a much smaller impact
on intuitive reasoning than one might have anticipated. Despite showing a quantitative reduction in construal-based reason-
ing, biology majors often showed the same qualitative pattern of responses as non-majors and 8th graders.
Elsewhere we have argued that cognitive construals involved in intuitive biological thought may provide the foundation
for systems of misconceptions common to students in the biology classroom (Coley & Tanner, 2012) and presented evidence
for systematic linkages between construal-based thinking and biological misconceptions in both biology majors and non-
majors (Coley & Tanner, 2015). Taken together with those results, the current study underscores the impact of intuitive
knowledge in the science classroom, and raises important questions about the effects of science education on conceptual
understanding more generally. Intuitive biological reasoning is systematic and pervasive; consideration of such informal
conceptual systems that students bring with them into the biology classroom might greatly enhance our understanding
of how we learn formal biological science, and in turn how best to teach it. Addressing these and other related research ques-
tions will require increasingly interdisciplinary approaches that integrate the theories, methods, and traditions of both cog-
nitive science and science education research.
This research was funded by National Science Foundation EHR Core Research Grant #1535496 to J.D. Coley and K.D. Tan-
ner. We thank Heather Krepelka, Rebeca Rosengaus and Fred Davis for their support making this research possible, Kathryn
Hardin, Samantha Daoust and Nicole Betz for help in data collection, and Sarah Bissonnette for comments on the manuscript.
Appendix A. Detailed description of measures of intuitive biological thinking
A.1. Measures of anthropocentric thinking
These measures focused on the ‘‘human exceptionalism” aspect of anthropocentric thinking, and examined the degree to
which humans are seen as biologically unique versus as one species among many.
16 J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
A.1.1. Common ancestry
This measure was adapted from Shtulman (2006). Participants were given a list of nine organisms and asked to indicate
‘‘which share a common ancestor with humans.” Organisms were chosen to represent a broad sample of living things. Form
A included chimpanzee,squirrel,robin,triceratops,frog,goldﬁsh,spider,pine tree, and seaweed. Form B included gorilla,ele-
phant,sparrow,tyrannosaurus rex,salamander, jellyﬁsh, bee, daffodil and algae.
Participants checked a box next to each organ-
ism that they believed to share a common ancestor with humans. The correct (i.e., non-anthropocentric) response is to check all
boxes; lower numbers of checked boxes represent increasingly anthropocentric responses.
A.1.2. Property projection
This task was developed by Susan Carey in her classic work on conceptual change in biology (1985), and has been used
primarily in studies of cognitive development (e.g., Inagaki & Sugiyama, 1988; Ross et al., 2003). The task involves teaching a
novel property (which is usually physiological in nature) about a given biological category (the ‘‘base” organism), and asking
whether the property is likely to be shared by a range of ‘‘target” organisms that vary in their phylogenic distance from the
For this task, participants were taught a novel intrinsic, physiological property about humans (‘‘Suppose that gene X15 is
found in humans...”) and asked to rate the likelihood (on a scale from 1 ‘‘very unlikely” to 7 ‘‘very likely”) that a range of
target organisms (gorilla, squirrel, deer, sparrow, frog, bee, grasshopper, and daisy) would also have the gene. Targets were pre-
sented in a single randomized order. Participants circled a number corresponding to their likelihood rating for each
The task taps into thinking about biological similarity; the degree to which organisms are thought to share novel
intrinsic properties has been taken as an index of their perceived similarity. As such, projections from humans can be
seen as a measure of anthropocentric thinking; willingness to acknowledge the likelihood that other organisms share
intrinsic properties with humans—as indicated by relatively high likelihood scores—indicates a view of humans an
‘‘one animal among many” and therefore reﬂects a relative lack of anthropocentric thinking. Conversely, low likelihood
ratings indicate reasoning grounded in a stance that humans are biologically unique. One possible objection to this mea-
sure is that, because participants weren’t told what the gene in question does, it was reasonable for them to infer that it
was speciﬁc to humans. This objection is misguided, however, because it fails to take into account the overall genetic
similarity between humans and the target organisms. Statistically, it’s more like that humans share the gene with other
organisms than they do not.
A.1.3. Biological similarity
To measure reasoning about the phylogenic relatedness of speciﬁc organisms, we presented participants with ﬁve dif-
ferent triplets, each of which consisted of realistic black and white line drawings of a human and two other species, for
example, human, monkey, elephant. In each triplet, the human and one nonhuman species were more closely related than
the two non-humans were (i.e., they shared a more recent common ancestor according to current scientiﬁc consensus).
As such, the task pits biological knowledge against the anthropocentric construal of humans as unique. For each triad,
participants indicated which two were ‘‘most similar biologically.” Pairing the two non-humans was considered an
A.2. Measures of teleological thinking
To measure the extent of teleological thinking—the tendency to explain certain phenomena by appealing to goals, func-
tions, or purposes—in our participants, we borrowed a set of statements from Kelemen et al. (2013). Each statement pre-
sented a one-sentence explanation for a speciﬁc phenomenon, and statements fell into three categories. Appropriate
explanations (Kelemen et al.’s ‘‘true teleological” explanations) were reasonable explanations of human actions or properties
of human-made artifacts, e.g., ‘‘Pencils exist so that people can write with them.” Inappropriate explanations (Kelemen et al.’s
‘‘false teleological” explanations) linked related phenomena in causally spurious ways, e.g., ‘‘Houses have doorbells in order
to make dogs bark.” Finally, natural kind explanations (‘‘test” statements in Kelemen et al., 2013) provided a goal-oriented
explanation for natural phenomena, e.g., ‘‘Rain falls in order to allow plants to grow.”
We chose 10 appropriate, 10 inappropriate, and 18 natural kind explanations from Kelemen et al. (2013). We then ran-
domly assigned half of each type to Form A and half to Form B, so that each form included 19 statements (5 appropriate, 5
inappropriate, and 9 natural kind). Statements were presented in random order; participants indicated their agreement with
each statement on a scale of 1 (disagree) to 6 (agree). We expected participants to generally agree with appropriate state-
ments, and disagree with inappropriate statements. We were most interested in responses to natural kind statements as a
reﬂection of the degree of teleological thinking, and differences therein, among our participants.
Two items were excluded from analysis because, due to experimenter error, the categories that appeared on the two versions of the measures were
biological non-equivalent (goldﬁsh/jellyﬁsh and seaweed/algae).
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 17
A.3. Measures of essentialist thinking
A.3.1. Category homogeneity
One major consequence of essentialist thinking in biology is the assumption of homogeneity among category members
with respect to intrinsic properties (e.g., Gelman & Coley, 1990; Gelman & Markman, 1986; Shtulman & Schulz, 2008). Specif-
ically, essentialist thinking predicts that intrinsic properties such as physiological, structural, and stable behavioral traits are
typically seen as being shared by many or all members of a biological kind because they are caused by an underlying essence.
In contrast, biological kinds may be perceived as less homogeneous with respect to extrinsic properties—those true of an
individual due to unique events or transient experiences (e.g., Coley, 2012; Coley & Vasilyeva, 2010). To examine perceptions
of category homogeneity, we adapted a measure used by Shtulman and Schulz (2008; see also Emmons & Kelemen, 2015).
Participants were told that an individual creature (ant or kangaroo) possessed a novel intrinsic property (‘‘HN45 cells” or
‘‘gene DH97”) and separately, that another individual creature (grasshopper or giraffe) possessed a novel extrinsic property
(‘‘parasite WJ61” or ‘‘GU38 infection”), and asked what percentage of the entire category would share the property. For
Imagine that an individual ant has been found to have HN45 Cells. Given the presence of HN45 Cells in this individual ant,
what percentage of all ants would have HN45 Cells?
Thus, each participant was asked about a mammal and an insect, and about an intrinsic property (cells or genes) and an
extrinsic property (parasite or infection); pairings of animal categories and properties were counterbalanced across Forms A
and B. We asked about intrinsic versus extrinsic properties – rather than anatomical versus behavioral properties (e.g.,
Emmons & Kelemen, 2015; Shtulman & Schulz, 2008) – in order to maximally contrast a potentially essentialized property
with one much less likely to be prone to essentialist thinking, to expand evidence concerning reasoning about homogeneity,
and to provide a potential bridge to related work on inductive inference (e.g., Coley, 2012; Coley & Vasilyeva, 2010).
Essentialist thinking would be manifested in reasoning that assumes members of a species are relatively uniform with
respect to intrinsic properties, but not necessarily so with respect to extrinsic properties. This would be indicated by higher
homogeneity estimates for intrinsic than for extrinsic properties. Categories were varied to improve generalizability of
results; we had no predictions about category differences.
A.3.2. Innate potential
Another hallmark of essentialist thinking is the idea of innate potential; because of an underlying essence, category mem-
bers not only share properties, but also share the propensity to develop certain characteristics over time. To examine indi-
viduals’ thinking about innate potential, experimenters have often relied on ‘‘adoption” or ‘‘switched-at-birth” tasks (e.g.,
Eidson & Coley, 2014; Gelman & Wellman, 1991; Solomon et al., 1996; Taylor et al., 2009). In this task, participants are given
a scenario in which the offspring of one family/species are said to have been raised by another family/species, and are then
queried as to whether the target offspring would resemble birth or adoptive parents on various dimensions. We adopted this
‘‘nature-nurture” task to assess reasoning about innate potential. We chose to utilize a human version of this task for two
reasons. First, it allowed us to ask about a wider range of types of properties (personality, beliefs, preferences, abilities) than
an animal-based task would. Second, we thought than an animal-based adoption scenario, such as that used by Taylor et al.
(2009), although appropriate for younger children, was actually highly anthropomorphic, and probably inappropriate for a
study involving only older children and adults.
In our task, participants were given a scenario describing an infant born to one set of parents but raised by another, fol-
lowed by ﬁve pairs of contrasting properties for the birth parents and the parents that raised the child (e.g., ‘‘Mr. and Mrs. B
are shy. Mr. and Mrs. R are outgoing.) For each pair, they were asked which property the child would exhibit upon reaching
the participant’s age. We asked about ﬁve different kinds of properties: personality traits (e.g., shy vs outgoing), beliefs (e.g.,
belief that eating meat is healthy vs unhealthy), preferences (e.g., liking candy vs chips), abilities (e.g., being better at soccer
vs swimming) and physical traits (e.g., having brown vs blue eyes). Participants circled which property they thought the
child would display upon reaching young adulthood; different items for each kind of property were used in Form A and
Essentialist thinking would be manifest in reasoning that assumes that individuals are endowed at birth with innate
potential to develop certain characteristics. As such, responses that predict resemblance to birth parents indicate essentialist
A.3.3. Identity over transformations
Another consequence of essentialist thinking is that although category members typically resemble each other in many
ways, category membership is ultimately determined by underlying causal properties rather than superﬁcial ones. This type
of reasoning has been examined using what are known as ‘‘transformation tasks” (e.g., Hampton et al., 2007; Keil, 1989; Rips,
1989) in which reasoning about the category membership of an individual who had undergone an appearance-changing
transformation are assessed. We adapted this task and gave participants two ﬁctitious scenarios in which an animal under-
went a transformation in appearance and behavior. In one type of scenario, the transformation was a result of natural devel-
opmental process (akin to metamorphosis). For example,
18 J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
An animal had a segmented body with no arms or legs, and it burrowed into the soil sometimes. The animal looked and
acted just like a worm. One day, as the result of a natural developmental process, the animal began to change. It devel-
oped a small shell that it carried on its back, grew two short antennae from its head, and left a slimy trail wherever it
went. The animal ended up looking and acting just like a snail.
In the other type of scenario, the transformation was triggered by exposure to toxic contamination. For example,
An animal had sharp front fangs, scaly skin, and a forked tongue. It looked and acted just like snake. One day, as the result
of toxic contamination in its environment, the animal began to change. It grew four legs, shed its fangs, and its tongue
became sticky. The animal ended up looking and acting just like a lizard.
After reading each scenario, participants indicated on a 9-point scale whether the animal was biologically a member of
the pre-transformation category (e.g., snake) or the post-transformation category (e.g., lizard) both before the change and
after the change. Form A and Form B each included a developmental and a mutation transformation, but the content of
the scenarios was different. These scenarios were constructed in the same way as those used by Hampton et al. (2007),
Experiment 3, Reduced Condition. We used this version of the task because we wanted a relatively conservative test of essen-
Essentialist thinking should give rise to responses that preserve category membership for both scenarios. For develop-
mental transformations, this might reﬂect reasoning based on the assumption that the creature at time 1 was a juvenile ver-
sion of the creature at time 2. For the mutation scenario, this might reﬂect reasoning that assumes changes brought about by
the toxic chemical affected superﬁcial features, but such a change should not change the underlying nature of the animal,
and therefore not affect category membership.
A.3.4. Absolute category membership
Another consequence of essentialist thinking is that membership in essentialized categories should be seen as absolute,
rather than graded. If indeed there is some underlying causal principle that makes a robin a robin, then any given individual
either has that essence or does not, and consequently, either is or is not a robin. Essentialist thinking does not permit some-
thing to be partially a robin (Diesendruck & Gelman, 1999; Kalish, 2002).
To assess individuals’ reasoning about the absolute nature of membership in biological categories, we presented potential
exemplars of biological categories (bird, ﬁsh, tree) and asked participants whether they were absolute members, absolute
non-members, or ‘‘sorta” members. Exemplars were chosen to vary in their typicality; some were prototypical category
members, some were atypical category members, and some were non-members that were similar to category members
(e.g., for bird,robin was a typical member, penguin was an atypical member, and bat was a similar non-member).
Participants were presented with ﬁve exemplars associated with each category for a total of 15 exemplars. Exemplars
were presented in a randomized order, and different exemplars were used for Form A and Form B. For each item, participants
saw the name of the exemplar paired with the name of the category, and checked one of four associated boxes: YES (the
exemplar is absolutely a member of the category), SORTA (the exemplar is partially a member and partially a nonmember),
NO (it’s absolutely not a member), or DON’T KNOW. Essentialist thinking should lead to high rates of absolute response
(either ‘‘yes” or ‘‘no”) and low rates of ‘‘sorta” responses.
Ahn, W., Kalish, C., Gelman, S. A., Medin, D. L., Luhmann, C., Atran, S., ... Shafto, P. (2001). Why essences are essential in the psychology of concepts. Cognition,
Arenson, M. B., & Coley, J. D. (2016). Anthropocentric thinking about familiar and novel properties. Poster presented at the 28th annual convention of the
Association for Psychological Science, Chicago, IL.
Asher, R. J., Bennett, N., & Lehmann, T. (2009). The new framework for understanding placental mammal evolution. BioEssays, 31, 853–864.
Atran, S. A., & Medin, D. L. (2008). The native mind and the cultural construction of nature. Boston: MIT Press.
Berlin, B. (1992). Ethnobiological classiﬁcation: Principles of categorization of plants and animals in traditional societies. Princeton, NJ: Princeton University
Brown, C. H. (1984). Language and living things: Uniformities in folk classiﬁcation and naming. New Brunswick, NJ: Rutgers University Press.
Carey, S. (1985). Conceptual change in childhood. Cambridge, MA: MIT Press.
Casler, K., & Kelemen, D. (2008). Developmental continuity in the teleo-functional bias: Reasoning about nature among Romanian Roma adults (Gypsies).
Journal of Cognition and Development, 9, 340–362.
Chi, M. T. H. (1992). Conceptual change in and across ontological categories: Examples from learning and discovery in science. In R. Giere (Ed.), Cognitive
models of science (pp. 129–160). Minneapolis, MN: University of Minnesota Press.
Coley, J. D. (2000). On the importance of comparative research: The case of folkbiology. Child Development, 71, 82–90.
Coley, J. D. (2007). The human animal: Developmental changes in judgments of taxonomic and psychological similarity among humans and other animals.
Cognitive Brain Behavior, 11, 733–756.
Coley, J. D. (2012). Where the wild things are: Informal experience and ecological reasoning. Child Development, 83(3), 992–1006.
Coley, J. D., & Muratore, T. M. (2012). Trees, ﬁsh, and other ﬁctions: Folk biological thought and its implications for understanding evolutionary biology. 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. 22–46). Cambridge: Oxford University Press.
Coley, J. D., Solomon, G. E. A., & Shafto, P. (2002). The development of folkbiology: A cognitive science perspective on children’s understanding of the
biological world. In P. Kahn & S. Kellert (Eds.), Children and nature: Psychological, sociocultural and evolutionary investigations (pp. 65–91). Cambridge,
MA: MIT Press.
Coley, J. D., & Tanner, K. (2012). Common origins of diverse misconceptions: Cognitive principles and the development of biology thinking. Cell Biology
Education, 11(3), 209–215.
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 19
Coley, J. D., & Tanner, K. (2015). Relations between intuitive biological thinking and biological misconceptions in biology majors and nonmajors. Cell Biology
Education, 14, 1–19.
Coley, J. D., & Vasilyeva, N. Y. (2010). Generating inductive inferences: Premise relations and property effects. In B. H. Ross (Ed.). The psychology of learning
and motivation (Vol. 53, pp. 183–226). Burlington, MA: Academic Press.
Diesendruck, G., & Gelman, S. A. (1999). Domain differences in absolute judgments of category membership: Evidence for an essentialist account of
categorization. Psychonomic Bulletin and Review, 6, 338–346.
DiSessa, A. A. (1993). Toward an epistemology of physics. Cognition and Instruction, 10(2 & 3), 105–225.
Eidson, R. C., & Coley, J. D. (2014). Not so fast: Gender essentialism in young adults. Journal of Cognitive Development, 15, 382–392.
Emmons, N. A., & Kelemen, D. (2015). Young children’s acceptance of within-species variation: Implications for essentialism and teaching evolution. Journal
of Experimental Child Psychology, 139, 148–160.
Evans, E. M., Rosengren, K., Lane, J. D., & Price, K. S. (2012). Encountering counterintuitive ideas: Constructing a developmental learning progression for
biological 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. 174–199). New York: Oxford University Press.
Evans, E. M. (2008). Conceptual change and evolutionary biology: A developmental analysis. In S. Vosniadou (Ed.), International handbook of research on
conceptual change (pp. 263–294). New York: Routledge.
Evans, E. M. (2013). Conceptual change and evolutionary biology: Taking a developmental perspective. In S. Vosniadou (Ed.), International handbook of
research on conceptual change (2nd ed., pp. 220–239). New York, NY: Routledge.
Feeney, A., & Heit, E. (Eds.). (2007). Inductive reasoning. Cambridge: Cambridge University Press.
Gee, H. (2013). The accidental species: Misunderstandings of human evolution. Chicago, IL: University of Chicago Press.
Gelman, S. A. (2003). The essential child: Origins of essentialism in everyday thought. New York: Oxford University Press.
Gelman, S., & Coley, J. D. (1990). The importance of knowing a dodo is a bird: Categories and inferences in 2-year-old children. Developmental Psychology, 26
Gelman, S. A., & Markman, E. M. (1986). Categories and induction in young children. Cognition, 23, 183–209.
Gelman, S. A., & Rhodes, M. (2012). ‘‘Two-thousand years of stasis”: How psychological essentialism impedes evolutionary understanding. In K. R.
Rosengren, S. Brem, E. M. Evans, & G. Sinartra (Eds.), Evolution challenges: Integrating research and practice in teaching and learning about evolution
(pp. 3–21). New York: Oxford University Press.
Gelman, S., & Wellman, H. (1991). Insides and essences: Early understandings of the non-obvious. Cognition, 38(3), 213–244.
Goldberg, R., & Thompson-Schill, S. (2009). Developmental ‘‘roots” in mature biological knowledge. Psychological Science, 20(4), 480–487.
Hampton, J., Estes, Z., & Simmons, S. (2007). Metamorphosis: Essence, appearance, and behavior in the categorization of natural kinds. Memory & Cognition,
Haslam, N., Rothschild, L., & Ernst, D. (2000). Essentialist beliefs about social categories. British Journal of Social Psychology, 39(1), 113–127.
Hatano, G., & Inagaki, K. (1994). Young children’s naive theory of biology. Cognition, 50, 171–188.
Herrmann, P., Waxman, S. R., & Medin, D. L. (2010). Anthropocentrism is not the ﬁrst step in children’s reasoning about the natural world. Proceedings of the
National Academy of Sciences, 107(22), 9979–9984.
Inagaki, K., & Hatano, G. (1991). Constrained person analogy in young children’s biological inference. Cognitive Development, 6, 219–231.
Inagaki, K., & Hatano, G. (2002). Young children’s naïve thinking about the biological world. New York: Psychology Press.
Inagaki, K., & Hatano, G. (2006). Young children’s conception of the biological world. Current Directions in Psychological Science, 15(4), 177–181.
Inagaki, K., & Sugiyama, K. (1988). Attributing human characteristics: Developmental changes in over- and under-attribution. Cognitive Development, 3,
Johnson, S. C., & Carey, S. (1998). Knowledge enrichment and conceptual change in folk biology: Evidence from people with Williams syndrome. Cognitive
Psychology, 37, 156–200.
Johnson, K. E., Mervis, C. B., & Boster, J. S. (1992). Developmental changes within the structure of the mammal domain. Developmental Psychology, 28, 74–83.
Kalish, C. W. (2002). Essentialist to some degree: Beliefs about the structure of natural kind categories. Memory & Cognition, 30, 340–352.
Keil, F. C. (1989). Concepts, kinds, and cognitive development. Cambridge, MA: MIT Press.
Keil, F. C. (2006). Explanations and understanding. Annual Review Psychology, 57, 227–254.
Kelemen, D. (1999). The scope of teleological thinking in preschool children. Cognition, 70, 241–272.
Kelemen, D. (2004). Are children ‘‘intuitive theists”? Reasoning about purpose and design in nature. Psychological Science, 15, 295–301.
Kelemen, D. (2012). Teleological minds: How natural intuitions about agency and purpose inﬂuence learning about evolution. In K. S. Rosengren, S. K. Brem,
E. M. Evans, & G. M. Sinatra (Eds.), Evolution challenges: Integrating research and practice in teaching and learning about evolution. New York: Oxford
Kelemen, D., & Rossett, E. (2009). The human function compunction: Teleological explanation in adults. Cognition, 111, 138–143.
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(4), 1074–1083.
Leddon, E., Waxman, S. R., Medin, D. L., Bang, M., & Washinawatok, K. (2012). One animal among many? Children’s understanding of the relation between
humans and nonhuman animals. In G. Hayes & M. Bryant (Eds.), Psychology of culture.Psychology of emotions, motivations and actions: Focus on
civilizations and cultures series. Hauppauge, NY: Nova Science Publishers.
Legare, C., Evans, E., Rosengren, K., & Harris, P. (2012). The coexistence of natural and supernatural explanations across cultures and development. Child
Development, 83(3), 779–793.
Lombrozo, T., & Carey, S. (2006). Functional explanation and the function of explanation. Cognition, 99, 167–204.
Madsen, O., Scally, M., Douady, C. J., Kao, D. J., DeBry, R. W., Adkins, R., ... Springer, M. S. (2001). Parallel adaptive radiations in two major clades of placental
mammals. Nature, 409, 610–614.
Maeyer, J., & Talanquer, V. (2010). The role of intuitive heuristics in students’ thinking: Ranking chemical substances. Science Education, 94(6), 963–984.
Mayr, E. (1982). The growth of biological thought: Diversity, evolution, and inheritance. Cambridge, MA: Harvard University Press.
Medin, D., & Atran, S. (2004). The native mind: Biological categorization and reasoning in development and across cultures. Psychological Review, 111(4),
Medin, D. L., & Ortony, A. (1989). Psychological essentialism. In S. Vosniadou & A. Ortony (Eds.), Similarity and analogical reasoning (pp. 179–195). Cambridge,
New York: Cambridge University Press.
Medin, D., Waxman, S., Woodring, J., & Washinawatok, K. (2010). Human-centeredness is not a universal feature of young children’s reasoning: Culture and
experience matter when reasoning about biological entities. Cognitive Development, 25(3), 197–207.
Morris, S. C., Taplin, J. E., & Gelman, S. A. (2000). Vitalism in naïve biological thinking. Developmental Psychology, 36(5), 582–595.
Nehm, R. H. (2010). Understanding undergraduates’ problem-solving processes. Journal of Microbiology & Biology Education, 11(2), 119.
Nehm, R., & Reilly, L. (2007). Biology majors’ knowledge and misconceptions of natural selection. BioScience, 57(3), 263.
Nehm, R. H., & Ridgway, J. (2011). What do experts and novices ‘‘see” in evolutionary problems? Evolution: Education and Outreach, 4(4), 666–679.
O’Hara, R. J. (1998). Population thinking and tree thinking in systematics. Zoologica Scripta, 26, 323–329.
ojalehto, b., Waxman, S., & Medin, D. (2013). Teleological reasoning about nature: Intentional design or relational perspectives? Trends in Cognitive Sciences,
Opfer, J. E., Nehm, R. H., & Ha, M. (2012). Cognitive foundations for science assessment design: Knowing what students know about evolution. Journal of
Research in Science Teaching, 49(6), 744–777.
Piaget, J. (1929). The child’s conception of the world. London: Routledge & Kegan Paul.
20 J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21
Richards, D., & Siegler, R. S. (1986). Children’s understandings of the attributes of life. Journal of Experimental Child Psychology, 42, 1–22.
Rips, L. J. (1989). Similarity, typicality, and categorization. In S. Vosniadou & N. Ortony (Eds.), Similarity and analogical reasoning (pp. 21–59). Cambridge:
Cambridge University Press.
Rosengren, K. R., Brem, S., Evans, E. M., & Sinatra, G. (Eds.). (2012). Evolution challenges: Integrating research and practice in teaching and learning about
evolution. New York: Oxford University Press.
Ross, N., Medin, D., Coley, J., & Atran, S. (2003). Cultural and experiential differences in the development of folkbiological induction. Cognitive Development,
Rottman, J., Zhu, L., Wang, W., Seston Schillaci, R., Clark, K. J., & Kelemen, D. (2016). Cultural inﬂuences on the teleological stance. Evidence from China.
Religion, Brain & Behavior.http://dx.doi.org/10.1080/2153599X.2015.1118402.
Shtulman, A. (2006). Qualitative differences between naïve and scientiﬁc theories of evolution. Cognitive Psychology, 52(2), 170–194.
Shtulman, A., & Harrington, K. (2016). Tensions between science and intuition across the lifespan. Topics in Cognitive Science, 8(1), 118–137.
Shtulman, A., & Schulz, L. (2008). The relation between essentialist beliefs and evolutionary reasoning. Cognitive Science, 32(6), 1049–1062.
Shtulman, A., & Valcarcel, J. (2012). Scientiﬁc knowledge suppresses but does not supplant earlier intuitions. Cognition, 124(2), 209–215.
Solomon, G., Johnson, S., Zaitchik, D., & Carey, S. (1996). Like father, like son: Young children’s understanding of how and why offspring resemble their
parents. Child Development, 67(1), 151.
Sousa, P., Atran, S., & Medin, D. L. (2002). Essentialism and folkbiology: Evidence from Brazil. Journal of Cognition and Culture, 2(3), 195–223.
Talanquer, V. (2006). Commonsense chemistry: A model for understanding students’ alternative conceptions. Journal of Chemical Education, 83, 811–816.
Talanquer, V. (2009). On cognitive constraints and learning progressions: The case of ‘‘structure of matter”. International Journal of Science Education, 31(15),
Talanquer, V. (2013). When atoms want. Journal of Chemical Education, 90(11), 1419–1424.
Taylor, M., Rhodes, M., & Gelman, S. (2009). Boys will be boys; cows will be cows: Children’s essentialist reasoning about gender categories and animal
species. Child Development, 80(2), 461–481.
Vosniadou, S., & Brewer, W. F. (1992). Mental models of the earth. A study of conceptual change in childhood. Cognitive Psychology, 24, 535–585.
Ware, E. A., & Gelman, S. A. (2014). You get what you need: An examination of purpose- based inheritance reasoning in undergraduates, preschoolers, and
biological experts. Cognitive Science, 38, 197–243.
Waxman, S., Medin, D., & Ross, N. (2007). Folkbiological reasoning from a cross-cultural developmental perspective: Early essentialist notions are shaped by
cultural beliefs. Developmental Psychology, 43(2), 294–308.
J.D. Coley et al. / Cognitive Psychology 92 (2017) 1–21 21