AFFECTIVE AND COGNITIVE RESPONSES TO INSECTS
AND OTHER ARTHROPODS
By
Amanda R. Lorenz-Reaves

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Entomology – Doctor of Philosophy
2017

ABSTRACT
AFFECTIVE AND COGNITIVE RESPONSES TO INSECTS AND OTHER ARTHROPODS
By
Amanda R. Lorenz-Reaves
Insects are the most abundant and diverse group of animals on Earth. Though as a group
they do far more ecological good than harm, previous studies have shown that human attitudes
toward insects are mainly negative. Attitudes have affective (emotions) and cognitive (beliefs,
mental representations) components that interact to influence behavior. Negative attitudes toward
insects are associated with negative affect such as disgust and fear, and can have negative
consequences, such as a lack of conservation funding. In addition, negative attitudes can cause
people to avoid insects and/or feel distress when insects are present, prompting a disconnection
between the public and the insect world.
To explore affective responses to insects, Chapter Two focuses on the emotion of disgust.
Disgust is associated with avoidance of objects that cause sickness (e.g. rotten meat) or
undesirable social conduct (e.g. moral disgust). Disgust is partitioned into distinct domains –
pathogen, moral, sexual, etc. To determine whether disgust stimulated by insects belongs in a
unique domain, incoming freshmen at a large public university were surveyed (Chapter Two).
Survey items pertained to moral, pathogen, and insect-specific disgust. Factor analyses indicate
that insect disgust and pathogen disgust are part of the same construct, unique from moral
disgust. This implies that insects are perceived with the same feelings of disgust felt for
pathogens.
To explore cognition associated with insects, Chapter Three and Appendix One focus on
mental models of insects and other arthropods. Mental models are internal representations of

external entities that are used to reason, make inferences, conduct thought experiments, and
anticipate future events. Drawings reflect important qualities of mental models including
knowledge categorization and organization. Drawings of insects were collected from participants
with high and low expertise in entomology. Salient insect features were indexed and principal
components analysis applied to detect underlying patterns. Two distinct components emerged –
(1) a non-winged “crawling” insect, and (2) a legless winged “flying” insect, implying that flying
and crawling insects are perceived as distinct from each other (Chapter Three). A similar
analysis of children’s drawings of insects also showed a distinction between crawling and flying
insects (Appendix One).
Finally, to explore the interaction between affective and cognitive responses to insects,
drawings of “disgusting” and “not disgusting” insects from participants sampled in Chapter
Three were compared. Participants were also surveyed to gain a quantitative measure of disgust
associated with insects. Experts exhibited significantly lower disgust responses than novices.
Additionally, the inclusion of legs on drawings of insects deemed not disgusting correlated
negatively with disgust. In contrast, the inclusion of legs on drawings of disgusting insects
correlated positively with disgust. This suggests that crawling insects may be regarded as being
more disgusting than flying insects. A multiple linear regression was conducted on the not
disgusting insect drawings to determine whether drawing a crawling insect as well as one’s
expertise level, could predict insect-associated disgust. Approximately 35% of the variation in
disgust was attributable to subject group (expertise) and the degree to which drawings aligned
with the crawling insect model. In addition, this study also demonstrates that examining
drawings of insects can be a useful tool to shed light on affect and cognition associated with
insects.

ACKNOWLEDGEMENTS
I am truly grateful to my co-advisors, Dr. Gabe Ording and Dr. Julie Libarkin. Thank you
both for being endlessly supportive and encouraging. Thank you so much for providing me with
the resources and training to complete this project. I have grown so much as a scholar and as a
person over the course of the past four years and I owe that to you both. It has been absolutely
wonderful to work with you and get to know you!
I also want to thank my committee members, Dr. Jim Smith, Dr. Suzanne Thiem, and Dr.
Bill Ravlin, for their advice and support. I have learned so much from you. In addition, I want to
thank all of the dedicated and hard-working teaching assistants with whom I have worked in ISB
201L: Knute Gunderson, Steven Nichols, Courtney Larson, Casey Shaw, Scooter Nowak,
Marissa Schuh, Chris Adams, Rachel Osborn, Nicole Quinn, Krista Buehrer, Courtney
Weatherbee, Liz Davidson-Lowe, Kristin Poley, Jeff Shoemaker, Colin O’Neil, and Paul OwenSmith. Thank you also to everyone from CISGS for your support, especially Marsha Walsh and
Dr. Stephen Thomas.
Nicole Fisher provided tremendous assistance on multiple projects; thank you so much
for the endless hours spent coding drawings! Thank you to Caitlin Callahan, Christy Steffke,
Jessica Kansman, and Katie Demeuse for your friendship and assistance collecting surveys.
Thank you also to Heather Lenartson-Kluge and the Department of Entomology at Michigan
State for allowing me to collect surveys at the conference. To Paty Jaimes, Caitlin Kirby, and
other members of the GRL, thank you so much for reading my papers and sharing wisdom on
academia, parenting, and everything in-between! Finally, I extend profound thanks to my family
for providing me with endless love and support – my husband Pat and step-son Mason, as well as
my parents, Beth and Joe, sister, Cathy, and extended family. This is for you.

iv

TABLE OF CONTENTS
LIST OF TABLES…………………………………………………………………………..…vii
LIST OF FIGURES……………………………………………………………………….……ix
CHAPTER ONE
AN INTRODUCTION TO AFFECTIVE AND COGNITIVE RESPONSES TO INSECTS AND
OTHER ARTHROPODS…………………………………………………………………………1
Research questions……………………...………………………………………...7
REFERENCES………………………………………………………………………….9
CHAPTER TWO
DISGUST IN RESPONSE TO SOME ARTHROPODS ALIGNS WITH DISGUST
PROVOKED BY PATHOGENS………………………………………...…………….………..14
ABSTRACT……………………………………………………………………………….15
INTRODUCTION………………………………………………………………………..16
METHODS……………………………………………………………………………….19
Participants………………………………………………………...…………….19
Materials…………………………………………………………………………19
Analyses………………………………………………………………………….20
RESULTS………………………………………………………………………………...21
Exploratory factor analysis………………………………………………………21
Confirmatory factor analysis…………………………………………………….23
Comparison of the means………………………………………………………..25
DISCUSSION……………………………………………………………………………26
REFERENCES………..………………………………………………………………...31
CHAPTER THREE
DRAWINGS OFFER INSIGHTS ABOUT MENTAL MODELS OF INSECTS…....................36
ABSTRACT……………………………………………………………………………….37
INTRODUCTION………………………………………………………………………..38
METHODS……………………………………………………………………………….43
Participants & data collection…………………………………………………...43
Coding & frequency of insect features…………………………………………...45
Total number of features…………………………………………………………47
Completeness of drawn insects…………………………………………………..48
Correlations between insect features…………………………………………….49
Principal components analysis…………………………………………………..49
Model scores & expertise…………………………………………………..........49
RESULTS………………………………………………………………………………..50
Participants & data collection…………………………………………………..50
Coding & frequency of insect features…………………………………………..52
Total number of features…………………………………………………………52
Completeness of drawn insects…………………………………………………..54

v

Correlations between insect features…………………………………………....55
Principal components analysis…………………………………………………..57
Model scores & expertise………………………………………………………..59
DISCUSSION…………………………………………………………………………….60
Limitations……………………………………………………………………….67
Conclusions………………………………………………………………………67
REFERENCES……...…………………………………………………………………...69
CHAPTER FOUR
IMPACTS OF AFFECT AND EXPERTISE ON DRAWINGS OF DISGUSTING AND NOTDISGUSTING INSECTS………………………………………………………………………..73
ABSTRACT………………………………………………………………………………74
INTRODUCTION………………………………………………………………………..75
Research questions………………………………………………………………76
METHODS………………………………………………………………………………77
Participants & data collection…………………………………………………..77
Insect disgust survey………………………………………………………….....78
Drawings, insect models, & insect model scores………………………………..80
Correlational analysis…………………………………………………………..81
Multiple regression……………………………………………………………..82
RESULTS……………………………………………………………………………….83
Participants & data collection………………………………………………….83
Insect disgust survey…………………………………………………………….84
Drawings, insect models, & insect model scores……………………………….86
Correlational analysis………………………………………………………….88
Multiple regression……………………………………………………………..89
DISCUSSION……………………………………………………………………………93
REFERENCES…………………………………………………………………………..97
APPENDIX…………………………………………………………………………………….101
REFERENCES……………………………..…………………………………………..115

vi

LIST OF TABLES
Table 2.1. Factor loadings for each survey item using maximum likelihood factor analysis with
promax rotation. Loadings below 0.32 are suppressed…………………………….. 23
Table 2.2. Mutivariate analysis of variance (MANOVA) to determine the effects of gender and
major on our two main Disgust Scales (Pathogen and Moral), as well as the two
subscales of Pathogen Disgust (Pathogen-General and Pathogen-Insect)………….. 26
Table 3.1. Relative differences in science and entomological expertise across our three sampled
populations…………………………………………………………………………...44
Table 3.2. Descriptive statistics for each participant group. ENT participants were expert
entomologists; GEO participants were expert geologists (non-entomologist
scientists); NSU participants were non-science undergraduate students…………….51
Table 3.3. Percentage of individuals from each participant group who drew each morphological
insect feature. Features that are part of the basic correct insect are italicized. Features
are listed in order of greatest-to-least cumulative frequency………………………...53
Table 3.4. Point-biserial correlations between insect features for all three participant groups
combined. One asterisk indicates that the relationship is significant at the level of p <
0.05; two asterisks indicate significance at p < 0.01………………………………...56
Table 3.5. Mean model scores, correctness score, and total number of features for each subject
group…………………………………………………………………………………60
Table 4.1. Descriptive statistics for each participant group (ENT, GEO, NSU). Reproduced from
Chapter Three of this dissertation……………………………………………………83
Table 4.2. Frequencies of coded insect features (all participant groups combined). In the
Difference (ND – D) column, positive values indicate that the feature was more
common in the not disgusting drawings; negative values indicate that the feature was
more common in the drawings of disgusting insects………………………………...87
Table 4.3. Mean model scores (±SD) for disgusting and not disgusting drawings across
participant groups…………………………………………………………………….87
Table 4.4. Pearson correlations between disgust score and model scores calculated from
drawings of (A) not-disgusting insects, and (B) disgusting insects. * Significant
correlation at level of 0.05. **Significant correlation at level of 0.01………………89
Table 4.5. Linear model of the predictors of Insect Disgust Score……………………………..93
Table A.1. Age and gender of participants…………..………………………………………..104

vii

Table A.2. Frequencies of the presence of each variable and also significance of differences
pre/post using paired-samples t-tests……………………………………………..108
Table A.3. Pearson correlations between major coded variables before and after the instructional
period. *p < 0.01; **p < 0.05………………………………………………….….109
Table A.4. Factor loadings for drawings pre-participation. Loadings below ±0.3 are
suppressed………………………………………………………………………...112
Table A.5. Factor loadings for post-participation drawings. Loadings below ±0.3 are
suppressed………………………………………………………………………...113

viii

LIST OF FIGURES
Figure 1.1. Typical representation of an arthropod with important features labeled……………..2
Figure 2.1. Confirmatory factor analysis with two latent variables……………………………..24
Figure 3.1. An example drawing of an insect with the coding scheme applied…………………46
Figure 3.2. Schematic representation of a complete insect (A), along with example drawings
from ENTs (B), GEOs (C), and NSUs (D)………………………………………….48
Figure 3.3. Entomologists (ENT) include more features on average than non-science
undergraduates (NSU) or non-entomologist scientists (GEO) in drawings of nondisgusting insects……………………………………………………………………54
Figure 3.4. Entomologists (ENT) draw insects that are more complete than non-science
undergraduates (NSU) or non-entomologist scientists (GEO). Complete Insect Score
was calculated by examining all non-disgusting insect drawings and giving one point
for each of six features essential to all insects: head, eyes, mouth, antennae,
segmentation, and legs. Results from a Kruskal-Wallis H Test showed that Complete
Insect Score differed across the three groups (p < 0.001). Data are shown as mean ±
SEM of Complete Insect Score, representative of drawings by 79 ENT, 92 GEO, and
101 NSU participants……………………………………………………………….55
Figure 3.5. Schematic and drawn representations of the crawling model and flying model based
on the results of our PCS. Factor loadings for each feature are also given; loadings
below ±0.32 are suppressed…………………………………………………………58
Figure 3.6. Mean model score for each of the two models from participants from each group.
Error bars denote standard error…………………………………………………….60
Figure 4.1. Crawling and flying insect models as detected through principal components analysis
(Chapter Three)……………………………………………………………………..81
Figure 4.2. Mean insect disgust score (± SE) across our three sampled populations. The y-axis
begins at 9 because the lowest possible disgust score was 9. NSUs had the highest
mean score, followed by GEOs, with ENTs having the lowest perceived disgust
associated with insects………………………………………………………………84
Figure 4.3. Mean disgust score plotted by age category for each participant group. NSU = nonscience undergraduates; GEO = non-entomologist scientists; ENT = entomologists.
The quantity of participants reporting age in each category for each group is
indicated within the bottom of each bar. Note that all NSU participants reported their
age as between 18 and 25 years. GEO and ENT participants between the ages of 18
and 25 generated mean disgust scores well under that of NSU participants………..85

ix

Figure 4.4. Two word clouds based upon participants’ recordings of the type of insect they had
intended to portray, showing dominant types of (A) not-disgusting (left) and (B)
disgusting (right) insects drawn. The relative sizes of the words indicate their
frequency in the dataset……………………………………………………………..88
Figure A.1. Examples of coded features…………………………………………………..…...107
Figure A.2. Schematic illustration of both emergent insect models from the PCA of the pre- and
post-instruction drawings. The crawling model did not change pre- and postparticipation. The flying model changed in that long body was dropped and no legs
was adopted……………………………………………………………………...111

x

CHAPTER ONE
AN INTRODUCTION TO AFFECTIVE AND COGNITIVE RESPONSES TO INSECTS AND
OTHER ARTHROPODS

1

“We are psychologically, as well as ecologically bound up with insects.”
− Adam Dodd, “Minding insects: scale, value, world”
Insects are simultaneously the most diverse and abundant animals on Earth. Their small
size allows them to capitalize on many diverse ecological niches (Wilson, 1987). They are vitally
important for numerous valuable ecological services such as pollination, dung burial,
decomposition, and soil aeration. In the U.S. alone, ecological services provided by insects are
estimated to be worth approximately $57 billion each year (Losey & Vaughan, 2006). In
addition, many human medicines are derived from insects (Costa-Neto, 2005; Cherniak, 2010).
Insects are animals belonging to the taxonomic Class Insecta. Insecta is part of a broader
taxonomic group, the Phylum Arthropoda, which also includes spiders, scorpions, ticks, mites,
centipedes, millipedes, and crustaceans. All arthropods are characterized by their hard chitinous
outer covering (exoskeleton), jointed appendages, segmented bodies, an open circulatory system,
and bilateral symmetry (Fig. 1.1).

Exoskeleton
Jointed Legs
Segmented Body
Open Circulatory
System
Bilateral Symmetry

Figure 1.1. Typical representation of an arthropod with important features labeled
There are an estimated 5-10 million total species of arthropods, the bulk of which are as
yet undescribed (Odegaard, 2000). The vast majority of arthropods are insects. Some non-insect

2

arthropods, such as spiders, are frequently confused with insects (Kellert, 1993; Shepardson,
2002). A common tendency is to lump non-insect arthropods together with insects in to one
group referred to simply as “bugs” (Shipley & Bixler, 2016). Interacting with insects is a
common experience shared by people around the globe. Insects occur everywhere that humans
do. They are present in the traditions, art, and folklore of numerous cultures (Hogue, 1987), and
provide a nutritious food source for many people (DeFoliart, 1999). It is impossible to separate
insects from the experience of human life. Human experiences with insects can be positive, for
example, watching butterflies or beekeeping. Experiences with insects can be neutral, such as
noticing an insect on the sidewalk. Experiences with insects can also be negative, such as getting
stung or finding a maggot in one’s food. Though insects as a whole do far more ecological good
than harm, and indeed, humans would likely not exist without insects (Wilson, 1987), studies
have shown that public attitudes toward insects are negatively biased (Byrne et al., 1984; Kellert,
1993; Driscoll, 1995; Bjerke et al., 1998; Schlegel & Rupf, 2010; Prokop et al., 2011; Shipley &
Bixler, 2016).
Negative public attitudes toward insects are characterized by negative emotional
responses to them such as disgust and fear (Bjerke et al., 1998; Smithsonian Institute, 2007;
Gerdes et al., 2009), and this likely encourages negative behaviors toward insects such as
avoidance or destruction. Studies addressing negative biases toward insects have given important
insights about human attitudes toward insects (as well as non-insect arthropods that are mistaken
for insects, e.g. spiders). These types of studies have shown that attitudes toward arthropods are
more negative if they are perceived as being harmful or associated with disease (e.g. parasites,
mosquitoes, ticks, etc.) (Gerdes et al., 2009; Prokop & Fančovičová, 2010), and that the disgust
elicited by insects is similar to the disgust elicited by pathogenic substances (Lorenz et al., 2014).

3

Another important insight about attitudes toward insects is that the location where an insect is
observed influences the attitudinal response – insects found inside the home are more negatively
perceived than insects found outdoors (Byrne et al., 1984; Baldwin et al., 1998). Attitudes
toward insects also vary according to the type of insect, with butterflies and lady beetles evoking
more positive attitudes than other insects (Byrne et al., 1984; Kellert, 1993; Gerdes et al., 2009;
Lorenz et al., 2014). Finally, it has also been noted that a common belief about insects is that
every insect observed in or around the home is a pest capable of causing damage (Byrne et al.,
1984; Baldwin et al., 1998).
Scientists and policymakers should be concerned about widespread negative attitudes
toward insects because they may lead to negative environmental consequences such as misuse or
overuse of pesticides by homeowners, or other widescale eliminations of insects. Additionally,
invertebrates represent the largest proportion of biodiversity in the animal kingdom, and thus
preserving biodiversity necessitates preserving invertebrates (Leather, 2013). However, a relative
paucity of conservation funding is spent on conserving insects and other invertebrates, relative to
the dollars spent conserving vertebrate animals (Cardoso et al., 2011). This may be related to
negative attitudes toward insects, since people are more willing to contribute money toward the
conservation of species for which they hold positive attitudes (Martín-López et al., 2009).
Studies suggest that participation in educational programs about insects and other
arthropods may be effective at improving attitudes toward them. For example, Wagler and
Wagler (2011) showed that pre-service elementary teachers were more likely to say that they
would include insects in their curriculum after they had participated in a program where they
were exposed to live insects. Byrne et al. (1984) reported that college-educated individuals
tended to hold more positive attitudes toward arthropods than those with a high school diploma

4

or less, and postulated this was because college-educated individuals had greater knowledge of
the environment compared to others. Participation in educational programs is often an effective
mediator of negative attitudes in other disciplines outside of entomology. For example, Randler
et al. (2012) found that students who had previously performed a dissection of a dead animal
were less disgusted by the task than students who had never performed a dissection.
In order to improve the ability of education to promote positive attitudes toward insects,
it is important to first understand the attitudes, ideas, and beliefs that people already hold about
insects. Human beings are not blank slates – they approach and interpret new information or
experiences using what they already know and feel as a basis. Thus, from an educational
standpoint, in order to facilitate the acquisition of new knowledge, it is important to be aware of
students’ pre-existing ideas and biases. Attitudes are learned “favorable or unfavorable feelings
toward objects, persons, groups, or any other identifiable aspects of our environment” (Koballa,
1988, p. 117), which can be more broadly defined as responses to stimuli (Breckler, 1984).
Attitudes are often regarded to be products of interactions between three types of response:
cognition, affect, and behavior (Breckler, 1984; Farley & Stasson, 2003; etc.). Affect refers to
emotional reactions; cognition refers to thoughts, beliefs, and mental representations; and
behavior refers to actions or intentions to act (Breckler, 1984). These three types of responses to
stimuli interact to constitute an attitude toward stimuli.
Attitudes are incorporated in mental models (Jones et al., 2011; Jung & Yim, 2015).
Mental models are internal cognitive representations of objects or systems (Vosniadou &
Brewer, 1994; Merrill, 2000; Jones et al., 2011). Mental models are representative of the current
state of one’s understanding and knowledge about a subject, and are also shaped by one’s
attitudes, beliefs, and values pertaining to the subject (Jones et al., 2011; Jung & Yim, 2015).

5

The structure of a mental model is representative of the structure of the corresponding object or
system, thus, by manipulating a mental model, an individual interacts with their knowledge
(Jones et al., 2011). Mental models are used during reasoning and problem solving tasks, as well
as to make predictions about the outcomes of events. Thus, they drive decision-making and other
behaviors. Mental models are simplified from reality in that they do not capture all minutiae and
other particulars (Jones et al., 2011). One of the most important properties of mental models is
that they are dynamic; they are created at a point of need and housed in working memory
(Vosniadou & Brewer, 1994).
Attitudes and mental models can be difficult to study because they exist only in the mind
of an individual. In previous literature, information about the composition of mental models is
often obtained by using such techniques as interviewing, diagramming, and estimating
relatedness between model components (Rowe & Cooke, 1995; Diaz, 2009). Drawing is another
useful method of eliciting information about the content and configuration of mental models
(Dove et al., 1999; Moseley et al., 2010; Libarkin et al., 2015; Quillin & Thomas, 2015). This is
because people interact with their mental models as they are drawing (Quillin & Thomas, 2015).
The action of creating a drawing requires an individual to first recall the important components
of the model, and then arrange those components in a way that is representative of the concept
being modeled. Though most studies eliciting drawings are focused on gaining insights about
cognition, drawings can also be utilized to study affect (e.g. Oster & Crone, 2004; Löfström &
Nevgi, 2012). In addition, drawings are free from limitations imposed by the collection of
spoken or written responses, such as vocabulary size and language fluency.
One method that is commonly used to gain insight from drawings is indexing (Libarkin et
al., 2015). Indexing involves detection and documentation of salient features in a drawing.

6

Insights can then be drawn from the presence or absence of salient features, as well as the
relationships between features. Typically, Thematic Content Analysis (TCA) is used to identify
meaningful patterns in relationships between features. Thematic content analysis is characterized
by the detection of common themes across participants (Anderson, unpublished data) after visual
inspection by the researcher. However, despite the best intentions it is possible for bias to occur
with human researchers, or they may miss an important theme (Anderson, unpublished data). To
avoid these potential pitfalls associated with TCA, it is beneficial to have a computer program,
rather than a researcher, search for patterns in the data. For example, Libarkin et al.’s use of
principal components analysis resulted in the emergence of underlying mental models of the
greenhouse effect (2015).

Research questions
This dissertation is aimed at exploring affective (disgust) and cognitive (mental models)
responses stimulated by interactions with insects and other arthropods. Another aim of this
dissertation is to determine how expertise influences affective and cognitive responses. Thus, the
main research questions investigated in this dissertation are as follows:
1) Is disgust elicited by insects unique from other domains of disgust? (Chapter 2)
2) Can we learn about underlying mental models of insects by examining people’s drawings
of insects? (Chapter 3)
3) How do drawings of insects differ between experts and novices, and do they suggest
major differences in their mental models? (Chapter 3)
4) What can people’s drawings of insects tell us about their level of disgust associated with
insects, and how is this influenced by expertise? (Chapter 4)

7

Results from this research will provide insights about human affect and cognition
stimulated by insects and other arthropods. These insights will assist educators in developing
new educational programs about insects that effectively challenge pre-existing misconceptions
and negative biases. The ultimate goal is for this work to facilitate improved public attitudes
toward insects, leading to increased public engagement in entomology.

8

REFERENCES

9

REFERENCES
Anderson R. Thematic Content Analysis (TCA): Descriptive presentation of qualitative data.
Unpublished data. Available at http://rosemarieanderson.com/wpcontent/uploads/2014/08/ThematicContentAnalysis.pdf Accessed July 6, 2017.
Askew, C., Çakır, K., Põldsam, L. and Reynolds, G. (2014) The effect of disgust and fear
modeling on children’s disgust and fear for animals. Journal of Abnormal Psychology, 123(3):
566–577.
Baldwin, R. W., Koehler, P. G., Pereira, R. M., and Oi, F. M. (2008) Public perceptions of pest
problems. American Entomologist, 54(2): 73-79.
Bandura, A. (1971) Social learning theory. General Learning Press, New York, New York, USA.
Breckler, S. J. (1984) Empirical validation of affect, behavior, and cognition as distinct
components of attitude. Journal of Personality and Social Psychology, 47(6): 1191–1205.
Bjerke, T., Ødegårdstuen, T. S., and Kaltenborn, B. P. (1998) Attitudes toward animals among
Norwegian adolescents. Anthrozoös, 11(2): 79–86.
Byrne, D. N., Carpenter, E. H., Thoms, E. M., and Cotty, S. T. (1984) Public attitudes toward
urban arthropods. Bulletin of the Entomological Society of America, 30(2): 40–44.
Cardoso, P., Erwin, T. L., Borges, P. A. V., and New, T. R. (2011) The seven impediments in
invertebrate conservation and how to overcome them. Biological Conservation, 144(11): 2647–
2655.
Costa-Neto, E. M. (2005) Entomotherapy, or the medicinal use of insects. Journal of
Ethnobiology, 25(1): 93–114.
Cherniak, E.P. (2010) Bugs as drugs, Part 1: insects: the new alternative medicine for the 21st
century? Alternative Medicine Review, 15: 124–135.
Davey, G. C. L., Forster, L., and Mayhew, G. (1993) Familial resemblances in disgust sensitivity
and animal phobias. Behaviour Research and Therapy, 31(1): 41–50.
DeFoliart, G. R. (2003) Insects as food: why the western attitude is important. Annual Review of
Entomology, 44: 21-50.
Diaz, A. E. (2009) Undergraduates’ mental models about insect anatomy and insect life cycles.
M.S. Thesis, California State University, Fullerton.

10

Dodd, A. (2013) Minding insects: scale, value, world. In The Management of Insects in
Recreation and Tourism, pages 23-36. Edited by R. H. Lemelin. Cambridge University Press:
Cambridge, UK.
Dove, J. E., Everett, L. A., and Preece, P. F. W. (1999) Exploring a Hydrological Concept
through Children’s Drawings. International Journal of Science Education, 21(5): 485-497.
Driscoll, J. W. (1995) Attitudes toward animals: species ratings. Society and Animals, 3(2): 139–
150.
Dunn, R. R. (2005) Modern insect extinctions, the neglected majority. Conservation Biology,
19(4): 1030–1036.
Farley, S. D., and Stasson, M. F. (2003) Relative influences of affect and cognition on behavior:
are feelings or beliefs more related to blood donation intentions? Experimental Psychology, 50:
55-62.
Gerdes, A. B. M., Uhl, G., and Alpers, G. W. (2009) Spiders are special: fear and disgust evoked
by pictures of arthropods. Evolution and Human Behavior, 30: 66–73.
Hogue, C. L. (1987) Cultural entomology. Annual Review of Entomology, 32: 181–199.
Jones, N., Ross, H., Lynam, T., Perez, P., and Leitch, A. (2011) Mental models: An
interdisciplinary synthesis of theory and methods. Ecology and Society, 16: 46.
Jung, W., and Kim, H. R. (2015) The effects of mental model on the variables in the technology
acceptance model. International Journal of Multimedia and Ubiquitous Engineering, 10(3): 171180.
Kellert, S. (1993) Values and perceptions of invertebrates. Conservation Biology, 7(4): 845–855.
Koballa, Jr., T. R. (1988) Attitude and related concepts in science education. Science Education,
72(2): 115–126.
Leather, S. R. (2013) Institutional vertebratism hampers insect conservation generally; not just
saproxylic beetle conservation. Animal Conservation, 16(4): 379–380.
Libarkin, J. C., Thomas, S. R., and Ording, G. (2015) Factor analysis of drawings: Application to
college student models of the greenhouse effect. International Journal of Science Education,
37(13): 2214-2236.
Löfström, E., and Nevgi, A. (2014) Giving shape and form to emotion: using drawings to
identify emotions in university teaching. International Journal for Academic
Development, 19(2): 99-111.

11

Lorenz, A.R., Libarkin, J.C., and Ording, G.J. (2014) Disgust in response to some arthropods
aligns with disgust provoked by pathogens. Global Ecology and Conservation, 2: 248-254.
Losey, J. E., and Vaughan, M. (2009) The economic value of ecological services provided by
insects. BioScience, 56(4): 311–323.
Martín-López, B., Montes, C., Benayas, J. (2009) The non-economic motives behind the
willingness to pay for biodiversity conservation. Biological Conservation, 139: 67–82.
Merrill, M. D. (2002). Knowledge objects and mental models. In Instructional Use of Knowledge
Objects, pages 261–280. Edited by D. Wiley. Agency for Instructional Technology and
Association for Educational Communications & Technology: Washington D.C., U.S.A.
Monge-Nájera, J. (2017) The power of short lectures to improve support for biodiversity
conservation of unpopular organisms: an experiment with worms. UNED Research Journal,
9(1).
Moseley, C., Desjean‐Perrotta, B., and Utley, J. (2010) The draw‐an‐environment test rubric
(DAET‐R): Exploring pre‐service teachers’ mental models of the environment. Environmental
Education Research, 16(2): 189-208.
Oster, G. D., and Crone, P. G. (2004) Using drawings in assessment and therapy: A guide for
mental health professionals. Psychology Press.
Prokop, P., Usak,M., Erdogan, M., Fancovicova, J., and Bahar, M. (2011) Slovakian and Turkish
students’ fear, disgust and perceived danger of invertebrates. Hacettepe University Journal of
Education, 40(40): 344–352.
Quillin, K., and Thomas, S. (2015) Drawing-to-learn: A framework for using drawings to
promote model-based reasoning in biology. CBE-Life Sciences Education, 14:es2.
Randler, C., Wüst-Ackermann, P., Vollmer, C., and Hummel, E. (2012) The relationship
between disgust, state-anxiety and motivation during a dissection task. Learning and Individual
Differences, 22(3): 419–424.
Rowe, A. L., and Cooke, N. J. (1995) Measuring mental models: Choosing the right tools for the
job. Human resource development quarterly, 6(3): 243-255.
Schlegel, J., and Rupf, R. (2010) Attitudes towards potential animal flagship species in nature
conservation: a survey among students of different educational institutions. Journal for Nature
Conservation, 18(4): 278–290.
Shepardson, D. P. (2002) Bugs, butterflies, and spiders: children's understandings about
insects. International Journal of Science Education, 24(6): 627-643.

12

Shipley, N. J., and Bixler, R. D. (2016) On the need to interpret insects: an always small but
gargantuan opportunity. Journal of Interpretation Research, 21(2): 63–72.
Smithsonian Institution Office of Policy and Analysis. (2007) The buzz on insects: a study of
attitudes and knowledge among visitors to the National Museum of Natural History. Retrieved
15, January, 2017 from
https://www.si.edu/content/opanda/docs/Rpts2007/07.09.Insects.Final.pdf.
Vosniadou, S., and Brewer, W. F. (1994) Mental models of the day/night cycle. Cognitive
Science, 18: 123-183.
Wagler, R., and Wagler, A. (2011) Arthropods: attitude and incorporation in preservice
elementary teachers. International Journal of Environmental and Science Education, 6(3): 229–
250.
Wilson, E. O. (1987) The little things that run the world (the importance and conservation of
invertebrates). Conservation Biology, 1(4): 344–346.

13

CHAPTER TWO
DISGUST IN RESPONSE TO SOME ARTHROPODS ALIGNS WITH DISGUST
PROVOKED BY PATHOGENS

Lorenz, A.R., Libarkin, J.C., and Ording, G.J. (2014) Disgust in response to some arthropods
aligns with disgust provoked by pathogens. Global Ecology and Conservation, 2: 248-254.
Global Ecology and Conservation by Elsevier. Reproduced with permission of Elsevier in the
format Thesis/Dissertation via Copyright Clearance Center. Order Detail ID: 70634046

14

ABSTRACT
Insects are disliked by the public, despite the fact that they provide valuable ecosystem
services and are vital components of ecosystems. Public support toward wildlife conservation is
influenced by attitudes toward different taxa, thus, the widespread negativity toward insects
shown by the general public almost certainly detracts from conservation efforts for them.
Negative attitudes toward insects and other invertebrates take many forms, one of which is the
feeling of disgust. Disgust has been widely researched and is typically divided into distinct
domains (e.g., moral disgust). In order to determine whether insect-specific disgust is unique
from other domains of disgust, we conducted a survey of 704 incoming freshmen at a major
Midwestern university with questions pertaining to Moral, Pathogen, and Insect-specific Disgust.
Factor analyses indicate that Insect Disgust and Pathogen Disgust are part of the same construct,
unique from Moral Disgust. Our results suggest that survey respondents perceived insects in the
same way as they would pathogens, at least in regard to disgust. This research provides insight
into how the public views insects, and will facilitate educational interventions aimed at
challenging negative attitudes toward insects. The Insect Disgust Scale will be a useful measure
of insect-related disgust in future studies.

15

INTRODUCTION
The importance of species conservation for ecosystems and human wellbeing is widely
recognized. Practices promoting the conservation of wildlife and natural resources are essential
for the preservation of biodiversity, which is crucial in all ecosystems and for all populations,
particularly in developing countries where people depend on endemic plants and animals for
medicines, food, and a source of livelihood (Adenle, 2012). Although the importance of
conservation in general is clear, a fundamental inequality exists in the types of organisms that
receive the largest conservation efforts; conservation endeavors toward vertebrate animals are
more likely to receive support than efforts toward invertebrates, fungi, or plants (Black et al.,
2001; Clark and May, 2002; Cardoso et al., 2011). For example, in 2009, the largest expenditures
of conservation dollars in the US all went toward vertebrate animals, including salmon, pallid
sturgeon, red-cockaded woodpecker, and bull trout (Buck et al., 2012).
Although invertebrates comprise 80% of all known species on Earth, they are the
recipients of only 10% of conservation funding (Cardoso et al., 2011; Collen et al., 2012). The
bias against invertebrates partly stems from the negative perception of insects by the general
public (Cardoso et al., 2011). The majority of people find insects to be scary, disgusting,
dangerous, or ugly. This is problematic for invertebrate conservation because negative attitudes
toward specific groups of organisms have been shown to adversely impact people’s willingness
to support the preservation of those organisms (Maresova & Frynta, 2007; Martín-López et al.,
2009; Knight, 2008; Prokop & Fančovičová, 2012; Prokop and Fančovičová, 2013a,b). Though
some insects are perceived positively (e.g., butterflies, dragonflies), the majority of insects as
well as other terrestrial arthropods are generally regarded in a negative light.
One prominent emotion that is often directed toward insects and their kin is disgust.

16

Disgust is considered to be, at its core, an evolutionary mechanism to avoid ingestion of harmful
substances (e.g., feces, spoiled food; Darwin, 1872/1965; Rozin & Fallon, 1987). However, the
feeling of disgust can be provoked by a diverse range of stimuli, including concrete objects (e.g.,
blood, worms, etc.) and individual behaviors (e.g., incest, stealing, etc.) that are unrelated to food
habits (Haidt et al., 1994; Oaten et al., 2009; Tybur et al., 2009, 2013; etc.). Disgust has
consequently been divided into separate ‘‘domains’’. For example, Haidt et al. (1994) created a
survey that divides disgust into seven different domains (e.g., food, sex, hygiene, animals, etc.)
and concludes that disgust is a mechanism whose primary purpose is to differentiate humans
from other animals. In contrast, some evolutionary psychologists (Tybur et al., 2009, 2013)
suggest that disgust can be divided into just three major domains: Moral (e.g., violation of
societal norms), Pathogen (e.g., infection by microorganisms), and Sexual (e.g., sexual behaviors
that may be damaging to one’s reproductive fitness). Thus, according to this interpretation,
disgust is not only a mechanism to avoid disease, but also functions as a regulator of mate choice
and social relations. We chose to model our Insect Disgust scale on the survey developed by
Tybur et al. (2009).
Logically, feelings of disgust inspired by insects can be anticipated to align most closely
with Pathogen Disgust, rather than Moral or Sexual Disgust. Insects and other arthropods share
commonalities with Pathogens in that they can occur in ‘‘outbreak’’ numbers, are of small size,
and often exhibit large populations and rapid reproduction rates. In addition, there are many
arthropod species that are ‘‘disease-relevant’’ by being either actively involved in the
transmission of disease (e.g., mosquitoes, fleas, and ticks), or associated with unhygienic
conditions (e.g., some flies). In one study, ratings of disgusting pictures of insects correlated
strongly with Pathogen Disgust (Prokop & Jančovičová, 2013). In contrast, there are no or few

17

conceptual links between insects and moral issues (Prokop & Jančovičová, 2013), or insects and
human sexual habits. In our survey, we included both the Pathogen Disgust scale from Tybur et
al. (2009) as well as the Moral Disgust scale, in order to compare disgust in response to insects
with these two previously validated domains of disgust. We did not include the Sexual Disgust
scale because it is not relevant to insect-related disgust, and because the inclusion of the Moral
Disgust scale already provided an effective comparison with the Pathogen Disgust Scale and our
Insect Disgust Scale.
The current study investigated the disgust responses of incoming freshmen at a large
Midwestern university. We chose to focus our invertebrate-specific survey items on a
combination of neutral insects (e.g., ants, crickets, bugs) as well as stereotypically unpopular or
disease-relevant insects and arachnids (e.g., cockroaches, scorpions, spiders). We avoided the
inclusion of charismatic insects that were not anticipated to evoke disgust, with the exception of
one item that was specific to butterflies. However, this item was removed from analysis once it
became clear that subjects responded to the butterfly question differently than to the other insectrelated questions (see Section 3).
We postulated that disgust in response to insects would emerge as a unique construct
when compared to disgust in response to non-insect stimuli, with this expectation based on prior
research suggesting that Insect and Pathogen Disgust, although different, would be correlated.
We also hypothesized that demographic variables would correlate with Insect Disgust. In many
studies on disgust, women display higher disgust sensitivity than men (Davey, 1994; Tucker &
Bond, 1997; Gerdes et al., 2009; Oaten et al., 2009; Prokop & Jančovičová, 2013; etc.). This may
be resultant of the traditionally higher parental investment exerted by women, although this
theory has not yielded any strong support (Prokop & Jančovičová, 2013). Other variables are

18

known to affect disgust sensitivity, such as cultural affiliation (Prokop & Fančovičová, 2010)
and political affiliation (Inbar et al., 2011), as well as participation in educational programs that
feature the disgusting object (Randler et al., 2012). However, gender is considered to be the most
dominantly influential demographic (Berger & Anaki, 2014). In our study, we chose to analyze
gender and college major, since our respondents were of similar ages and were all occupied as
full-time students at the same university. Sherman and Sherman (1998) reported lower disgust
sensitivity in nursing majors compared to other majors, though this was specific to items related
to their profession (e.g., bodily fluids). We specifically postulated that women would exhibit
higher disgust responses than men. We also postulated that non-science majors, who likely had
less biology background than science majors, would exhibit higher levels of disgust.

METHODS
Participants
College freshmen attending a university orientation program completed several surveys,
including a survey measuring disgust. We report on an analysis of 704 completed disgust
surveys. The study population was 49% male, had a median age of 18 years old, and an average
age of 17.9 ± 02 years. Sixty-six percent of participants had declared majors in a STEM
(Science, Technology, Engineering, or Mathematics) field.

Materials
Participants completed a survey containing 23 items related to disgust (Table 1). We
utilized disgust scales created by Tybur et al. (2009) to measure Moral and Pathogen-specific
Disgust. A set of items specifically related to insects was also added. One insect-related item

19

came from Tybur et al. (2009) (‘‘seeing a cockroach run across the floor’’); remaining items
were created for this study but were inspired by Tybur et al. (2009) as well as Bixler and Floyd
(1999). Morality concepts incorporated into items included lying, cheating, or stealing;
pathogens included blood, vomit, and excrement; and insects included mosquitoes, ants, and
cockroaches (Table 1). Survey participants responded to each item by rating their disgust level
on a scale from A to D; with A being ‘‘not at all disgusting’’, B ‘‘somewhat disgusting’’, C
‘‘very disgusting’’, and D ‘‘extremely disgusting’’. The scale also included demographic
questions relating to gender and major. This research followed all human subject protocols as
required by institutional IRB. Survey questions are provided in Table 1.

Analyses
All statistical analyses were conducted in SPSS 21.0, except for a confirmatory factor
analysis run in AMOS 21.0. We ran exploratory factor analysis and confirmatory analysis in
order to investigate the unidimensionality of Insect Disgust items and their relationship to
Pathogen and Moral Disgust. A multivariate analysis of variance (MANOVA) was also
performed to investigate the variance in scores on identified disgust scales that could be
explained by the common demographic variables of gender and area of study. Disgust scores
were calculated for each survey respondent by calculating the mean of their scores for each
identified disgust scale. Disgust scores in this analysis were the dependent variables, and gender
and major served as two independent variables.

20

RESULTS
Exploratory factor analysis
An initial exploratory factor analysis was used to determine the relationships between
disgust related to insects, morals, and pathogens. Forty-four surveys were discarded due to
missing or compromised data. 704 surveys remained and were included in the analysis.
Exploratory factor analysis was then used to identify covariance among the 23 survey items. To
test for normality, we tested each item for skewness and kurtosis, and discovered that all items
had skewness or kurtosis values less than |2| (Tybur et al., 2009), with the exception of a
butterfly item (‘‘Feeling a butterfly land on your arm’’). The butterfly item also proved
problematic upon examination of correlation matrices, which revealed low (below 0.3) and
irregular Pearson Correlations. These results were interpreted to reflect the fact that butterflies
are perceived quite differently (i.e., non-disgusting) from the rest of the insects included in the
survey, and the item was consequently removed from our analyses. The following analysis
considers the remaining 22 items.
We used factor analysis to examine underlying patterns in the data. Our data followed the
assumption of multicollinearity (VIF values <2). The Kaiser–Meyer–Olkin measure of sampling
adequacy was 0.904, which is above the recommended threshold of 0.6. Additionally, Bartlett’s
2

test of sphericity was significant (χ (231) = 4723.3, p < 0.001). Exploratory Factor Analysis
(EFA) on the remaining 22 items revealed a total of four eigenvalues greater than one: 5.956,
3.054, 1.227, and 1.037. Based on eigenvalues, the first factor explained 27% of the variance, the
second factor explained 13.9% of the variance, and the third and fourth factors each explained
<6% of the variance. Each survey item also had diagonals greater than 0.5 in the anti-image
correlation matrix, indicating that all questions should be included in the analysis (Neill, 2008).

21

We chose a two-factor model that explained 41% of the variance due to the previous theoretical
validation of the Moral and Pathogen scales in Tybur et al. (2009), as well as the ‘‘leveling off’’
of the eigenvalues in the scree plot after two factors (Cattell, 1966). Upon extraction, promaxrotated factor loadings suggested that the ‘‘Insect’’ and ‘‘Pathogen’’ items load onto one factor,
with ‘‘Moral’’ items loading on a second factor. Thus, we combined our ‘‘Insect’’ and
‘‘Pathogen’’ questions into one latent variable. For purposes of clarity in further discussion and
analysis, we henceforth will refer to our insect-specific items as the ‘‘Pathogen-Insect Disgust
Scale’’, and the pathogen questions from Tybur et al. (2009) as the ‘‘Pathogen-General Disgust
Scale’’. Discussion of the ‘‘Pathogen Disgust Scale’’ refers to insect and general pathogen items
collectively. We conducted reliability analyses on our two main Disgust Scales, which yielded a
high Cronbach’s alpha for both the Pathogen Disgust Scale (0.873) and the Moral Disgust Scale
(0.828). In addition, we also conducted reliability analysis on the Pathogen-Insect Disgust Scale,
in order to determine if this subset of items alone could be utilized effectively in future studies.
Cronbach’s alpha for insect-related items was high (0.838), and for general pathogen items was
acceptable (0.725), indicating that both components of the ‘‘Pathogen Disgust Scale’’ were
reliable metrics even when separated. Rotated factor loading values from our analysis with two
factors extracted were utilized in the subsequent confirmatory analysis (Table 2.1).

22

Table 2.1. Factor loadings for each survey item using maximum likelihood factor analysis with
promax rotation. Loadings below 0.32 are suppressed.
Item

Insect/Pathogen

Standing close to a person who has body odor.
Seeing some mold on old leftovers in your refrigerator.
Stepping on dog poop.
Sitting next to someone who has red sores on their arm.
Accidentally touching a person’s bloody cut.
Shaking hands with a stranger who has sweaty palms.
Finding a scorpion in your shoe.
Seeing an ant crawl across the floor.
Finding a bug in your shirt.
Watching a spider make its web.
Eating a chocolate-covered cricket.
Feeling a mosquito bite you.
Accidentally touching a spiderweb.
Watching a centipede crawl across your leg.
Seeing a cockroach run across a countertop.
Stealing from a neighbor.
A student cheating to get good grades.
Deceiving a friend.
Intentionally lying during a business transaction.
Forging someone’s signature on a legal document.
Shoplifting a candy bar from a convenience store.
Cutting to the front of a line to purchase the last few tickets to a show.

Moral

0.463
0.508
0.551
0.493
0.485
0.486
0.54
0.521
0.715
0.606
0.525
0.53
0.721
0.598
0.658

-

0.584
0.734
0.548
0.718
0.668
0.677
0.559

Confirmatory factor analysis
Based on the results of our exploratory factor analysis, we conducted a confirmatory
factor analysis (CFA). This technique allows significance testing of the structure of a
2

hypothetical model. In addition to χ , other measures of goodness-of-fit were used that account
for sample size and parsimony. These include RMSEA (Root Mean Square Error of
Approximation; values of <0.06 indicate good fit), CFI (Comparative Fit Index; values of >0.95
indicate good fit) and SRMR (Standardized Root Mean Squared Residual; values of <0.08
indicate good fit) (Hu & Bentler, 1999). Rotated factor loadings as well as the scree plot from

23

our exploratory analysis suggested that two factors should be extracted for our model, thus we
used two latent variables: Moral Disgust and Pathogen Disgust (including both insect and
2

general pathogen items; Fig. 2.1). Goodness-of-fit tests indicated good fit, χ (208, N = 704) =
654.614, p < 0.001, CFI = 0.902, RMSEA = 0.055, SRMR = 0.0476. Thus, confirmatory factor
analysis of the model suggested by our exploratory factor analysis yielded good fit, indicating
that the model is a match to our data.

Figure 2.1. Confirmatory factor analysis with two latent variables.

24

Comparison of the means
We assessed normality of the two primary Disgust scales: Moral Disgust and Pathogen
Disgust, as well as the two subsets of Pathogen Disgust: Pathogen-Insect Disgust and PathogenGeneral Disgust. All four scales showed normal Q–Q plots and frequency distributions, and also
exhibited homogeneity of variance (Levene’s test, p > 0.05 for all four scales), although the
Shapiro–Wilk test results from all four scales exhibited significant deviation from normality (p <
0.001). After considering both the visual examinations and normality test results, and taking the
large sample sizes into account, we decided to proceed with parametric tests in our analyses
(Ghasemi & Zahediasl, 2012).
In order to compare the effects of gender and area of study on scores from all four
Disgust scales (differentiating between Pathogen-General and Pathogen-Insect), we conducted a
multivariate analysis of variance (MANOVA; Table 2.2). Disgust scores for women (M = 2.909,
SD = 0.451) were found to be significantly higher than Disgust scores for men (M = 2.652, SD =
0.450) across all scales (p < 0.05). In contrast, both science majors (M = 2.742, SD = 0.456) and
non-science majors (M = 2.857, SD = 0.485) exhibited similar Disgust scores across all scales (p
> 0.1). There was no interaction between gender and area of study on Disgust scores (p > 0.5).

25

Table 2.2. Mutivariate analysis of variance (MANOVA) to determine the effects of gender and
major on our two main Disgust Scales (Pathogen and Moral), as well as the two subscales of
Pathogen Disgust (Pathogen-General and Pathogen-Insect).
Independent
Dependent Variable
df
F
Sig.
Variable
Gender
Pathogen Disgust
2
16.208
< 0.001
Moral Disgust
2
17.29
< 0.001
Pathogen-General Disgust
2
3.438
0.033
Pathogen-Insect Disgust
2
22.845
< 0.001
Major
Pathogen Disgust
1
1.162
0.281
Moral Disgust
1
0.01
0.92
Pathogen-General Disgust
1
0.169
0.681
Pathogen-Insect Disgust
1
1.663
0.198
Gender*Major
Pathogen Disgust
2
0.311
0.733
Moral Disgust
2
0.096
0.909
Pathogen-General Disgust
2
0.118
0.889
Pathogen-Insect Disgust
2
0.437
0.646

DISCUSSION
The ecological and economical importance of invertebrates cannot be overstated. Insects
and other terrestrial arthropods are valuable commodities because they provide many ecosystem
services that benefit both human and environmental interests (Kellert, 1993). These ecosystem
services include pollination, organic matter decomposition, and pest control. In the United States
alone, insect-mediated ecosystem services have been valued at $57 billion annually (Losey and
Vaughan, 2006).
Despite their importance, insects provoke largely negative emotions in people (Kellert,
1993; Bjerke and Østdahl, 2004; Schlegel and Rupf, 2009). For example, insects ranked number
two in a list of topics which urban students reported as frightening, second only to snakes (Bixler
et al., 1994). In addition, Shepardson (2002) noted that many children’s perceptions of insects

26

are largely negative and emphasize harmful aspects of human–insect interactions such as bites
and stings. People’s perceptions of organisms are influential in determining their willingness to
conserve those organisms (Knight, 2008; Martín-López et al., 2009; Prokop & Fančovičová,
2012; Prokop and Fančovičová, 2013a,b). Thus, it is important to gain a better understanding of
the fundamental ways in which insects and other arthropods are perceived by the general public.
We postulated that disgust experienced in response to insects and other arthropods is a
construct unique from other previously described domains of disgust, albeit with expectations of
correlation with pathogen-related disgust. To test this hypothesis, we conducted a survey of over
700 incoming freshmen at a large Midwestern university with questions pertaining to three
different topics known to raise disgust responses: morality, pathogens, and insects. We tested our
data for underlying patterns using exploratory and confirmatory factor analysis. Contrary to our
hypothesis, our results suggested that disgust in response to pathogens and disgust in response to
insects are part of the same construct. Our data shows that the insects included in our survey
were perceived in the same way as pathogens. Prokop and Jančovičová. (2013) documented
similar results when they measured disgust in young adolescents in response to pictures of
insects. In their study, Prokop and Jančovičová (2013) utilized the same Pathogen and Moral
Disgust Scales from Tybur et al. (2009) and documented a significant positive correlation
between the subjects’ ratings of Pathogen Disgust and the ratings of disgusting insect pictures,
and no significant correlation between the insect pictures and Moral Disgust scores. These results
are interesting because although many insects do pose health risks, the majority of insect species
are fairly innocuous and many are beneficial to human interests. Many of the insects included in
our survey were not associated with disease risk, including crickets and ants. Why then, did the
respondents still feel disgusted by insects that pose no threats or risk of disease? Early in our

27

analysis, we noted a differential response to a butterfly survey item that led to that item being
dropped from the analyses. Thus, it is probable that any other marked differences in responses to
a particular insect item would have been similarly distinct. Davey et al. (1998) also documented
a disgust response to harmless insects among students from varying countries. In contrast, a
study comparing affective responses of students after viewing disease-relevant and diseaseirrelevant pictures of arthropods revealed differential disgust responses, indicating that subjects
responded to disease-irrelevant insects with a lesser degree of disgust than to their diseasecausing counterparts (Prokop & Fančovičová, 2010). Additionally, a study by Gerdes et al.
(2009) demonstrated that people direct greater disgust responses toward disease-relevant or
dangerous insects than toward harmless insects. The current study supports the finding that
insects and pathogens are viewed similarly, regardless of the disease-relevance of specific
insects.
The insects-as-pathogens model fits with the established theory of disease avoidance as a
driver of disgust (Matchett & Davey, 1991; Oaten et al., 2009). The perception of insects as
pathogens is one possible explanation for the widely held negative attitudes toward insects by the
general public. Additionally, demographic variables are known to correlate with feelings about
insects. Previous studies of disgust have shown a gender bias in terms of disgust and fear toward
specific animals, with females showing higher sensitivity than males (Davey, 1994; Tucker &
Bond, 1997; Gerdes et al., 2009; Oaten et al., 2009; Prokop & Fančovičová, 2010; Prokop et al.,
2010; Prokop & Jančovičová, 2013; etc.). The results of this study are consistent with this
pattern. In terms of college major, we expected to see lower Pathogen-Insect Disgust from
students majoring in a science (STEM—Science, Technology, Engineering or Math) field, since
those students are likely to have a stronger background in the biological sciences and previous

28

studies suggest that habituation to distasteful objects such as insects decreases the disgust
response toward those objects (Bixler & Floyd, 1999; Randler et al., 2012). However, we
observed no difference in terms of Pathogen-Insect Disgust between STEM and non-STEM
students. This may be due to the fact that the students we surveyed were incoming freshmen and
had not yet experienced college-level courses in STEM disciplines.
Insects in general share certain commonalities with pathogens. Like pathogens, we often
do not notice insects due to their small size, and some do indeed cause harm to humans.
However, being disgusted by the vast majority of insects promotes unnecessary anxiety and the
avoidance of many invertebrates that are potentially beneficial. Additionally, an exaggerated
sense of disgust toward insects in general poses problems for society at large. For example, the
practice of entomophagy (eating insects) has been proposed as a global solution to world hunger,
and one of the major barriers to its progress is the disgust response toward insects, particularly in
western cultures (see Defoliart, 1999).
Our study has shown that insects appear to be perceived in the same way as pathogens by
incoming freshmen at a large Midwestern university. Whether this perception of insects is
generalizable toward the general public remains to be examined. Future research should
investigate which characteristics of insects specifically influence disgust, as well as investigate
how targeting these characteristics in educational programming can impact public support for
invertebrate conservation efforts. For example, Wagler and Wagler (2012) demonstrated that
elements of external insect morphology affected preservice teachers’ willingness to teach about
insects. Current estimations suggest that only 0.5% of total invertebrate diversity has been
assessed by the International Union for the Conservation of Nature (IUCN) for its Red List,
which is responsible for determining the endangerment status of individual taxa (Leather, 2013).

29

The sheer overwhelming diversity of insects, especially in comparison with vertebrate groups,
has prevented accurate large-scale estimates of relative numbers of endangered insect species
(Wilcove & Master, 2005). The potential loss of these insect species may have untold effects on
both the health of ecosystems and economic stability.
In terms of conservation efforts, it is extremely important to understand the emotions and
attitudes that people direct toward specific groups of animals, to recognize where those emotions
derive from, and to generate interventions to challenge those negative perceptions. The
Pathogen-Insect Disgust survey described in this study can be utilized at a large scale to
determine insect-related disgust sensitivity, and may prove useful in settings such as educational
outreach. Hopefully, as people become better educated about harmless neutral insects, they will
come to be recognized as different from pathogens. This would translate to a separation of insect
items from pathogen items during a factor analysis, rather than the clustering observed here—
such a divergence of scales would indicate that insect education has been effective. We welcome
future studies that evaluate the extent to which students and the general public differentiate
insects from pathogens.

30

REFERENCES

31

REFERENCES
Adenle, A.A. (2012) Failure to achieve 2010 biodiversity’s target in developing countries: How
can conservation help? Biodiversity and Conservation, 21(10): 2435–2442.
Berger, U., and Anaki, D. (2014) Demographic influences on disgust: Evidence from a
heterogeneous sample. Personality and Individual Differences, 64: 67–71.
Bixler, R.D., Carlisle, C.L., Hammltt, W.E., and Floyd, M.F. (1994) Observed fears and
discomforts among urban students on field trips to wildland areas. Journal of Environmental
Education, 26: 24–33.
Bixler, R.D., and Floyd, M.F. (1999) Hands on or hands off? Disgust sensitivity and preference
for environmental education activities. Journal of Environmental Education, 30: 4–11.
Bjerke, T., and Østdahl, T. (2004) Animal-related attitudes and activities in an urban population.
Anthrozoos, 17: 109–129.
Black, S.H., Shepard, M., and Allen, M.M. (2001) Endangered invertebrates: the case for greater
attention to invertebrate conservation. Endanger. Species UPDATE, 18: 42–50.
Buck, E.H., Corn, M.L., Alexander, K., Sheikh, P.A., and Melz, R., 2012. U.S. Congressional
Research Service. The endangered species act (ESA) in the 112th congress:
Conflicting values and difficult choices (R41608), in: Federation of American Scientists CRS
Reports Online; Accessed: February 26, 2014.
Cardoso, P., Erwin, T. L., Borges, P. A. V., and New, T. R. (2011) The seven impediments in
invertebrate conservation and how to overcome them. Biological Conservation, 144:
2647–2655.
Cattell, R. B. (1966) The scree test for the number of factors. Multivariate Behavavior Research,
1: 245–276.
Clark, J. A., and May, R. M. (2002) Taxonomic bias in conservation research. Science,
297(5579): 191–192.
Collen, B., Böhm, M., Kemp, R., and Baillie, J. E. M. (2012) Spineless: Status and Trends of the
World’s Invertebrates. Zoological Society of London, United Kingdom.
Darwin, C. (1965) The Expression of the Emotions in Man and Animals. University of Chicago
Press, Chicago, Originally published in 1872.
Davey, G. C. L., (1994) Self-reported fears to common indigenous animals in an adult UK
population: the role of disgust sensitivity. British Journal of Psychology, 85: 541–554.

32

Davey, G. C. L., McDonald, A. S., Hirisave, U., Prabhu, G. G., Iwawaki, S., Jim, C. I.,
Merckelbach, H., de Jong, P. G., Leung, P. W. L., and Reimann, B. C. (1998) A crosscultural
study of animal fears. Behaviour Research and Therapy, 36: 735–750.
Defoliart, G. R., (1999) Insects as food: Why the western attitude is important. Annual Review of
Entomology, 44: 21–50.
Gerdes, A., Uhl, G., and Alpers, G. W. (2009) Spiders are special: fear and disgust evoked by
pictures of arthropods. Evolution and Human Behavior, 30: 66–73.
Ghasemi, A., and Zahediasl, S. (2012) Normality tests for statistical analysis: a guide for nonstatisticians. International Journal of Endocrinology and Metabolism, 10: 486–489.
Haidt, J., McCauley, C., and Rozin, P. (1994) Individual differences in sensitivity to disgust: A
scale sampling seven domains of disgust elicitors. Personality and Individual Differences, 16:
701–713.
Hu, L., and Bentler, P. M. (1999) Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternatives. Structural Equation Modeling, 6: 1–55.
Inbar, Y., Pizarro, D., Iyer, R., and Haidt, J. (2011) Disgust sensitivity, political conservatism,
and voting. Social Psychological and Personality Science, 3: 537–544.
Kellert, S. R. (1993) Values and perceptions of invertebrates. Biological Conservation, 7: 845–
855.
Knight, A. J. (2008) ‘Bats, snakes and spiders, Oh my!’ How aesthetic and negativistic attitudes,
and other concepts predict support for species protection. Journal of Environmental Psychology,
28: 94–103.
Leather, S. R. (2013) Institutional vertebratism hampers insect conservation generally; not just
saproxylic beetle conservation. Animal Conservation, 16: 379–380.
Losey, J. E., and Vaughan, M. (2006) The economic value of ecological services provided by
insects. Bioscience, 56: 311–323.
Maresova, J., and Frynta, D. (2007) Noah’s ark is full of common species attractive to humans:
The case of boid snakes in zoos. Ecological Economics, 64: 554–558.
Martín-López, B., Montes, C., Benayas, J. (2009) The non-economic motives behind the
willingness to pay for biodiversity conservation. Biological Conservation, 139: 67–82.
Matchett, G., Davey, G. C. L. (1991) A test of a disease-avoidance model of animal phobias.
Behavior Research and Therapy, 29: 91–94.
Neill, J. (2008) Writing up a Factor Analysis. Retrieved from

33

http://www.bwgriffin.com/gsu/courses/edur9131/content/Neill2008_WritingUpAFactor
Analysis.pdf.
Oaten, M., Stevenson, R. J., and Case, T. I. (2009) Disgust as a disease-avoidance mechanism.
Psychological Bulletin, 135: 303–321.
Prokop, P., and Fančovičová, J. (2010) The association between disgust, danger and fear of
macroparasites and human behaviour. Acta Ethologica, 13: 57–62.
Prokop, P., and Fančovičová, J. (2013a) Self-protection versus disease avoidance: The perceived
physical condition is associated with fear of predators in humans. Journal of Individual
Differences, 34: 15–23.
Prokop, P., and Fančovičová, J. (2013b). Does colour matter? The influence of animal warning
coloration on human emotions and willingness to protect them. Animal Conservation, 16: 458–
466.
Prokop, P., and Jančovičová, M. (2013) Disgust sensitivity and gender differences: An initial test
of the parental investment hypothesis. In: Problems of Psychology in the 21st Century, vol. 7. pp.
40–48.
Prokop, P., Usak, M., and Fančovičová, J. (2010) Risk of parasite transmission influences
perceived vulnerability to disease and perceived danger of disease-relevant animals. Behavioural
Processes, 85: 52–57.
Randler, C., Hummel, E., and Prokop, P. (2012) Practical work at school reduces disgust and
fear of unpopular animals. Society and Animals, 20: 61–74.
Rozin, P., and Fallon, A. E. (1987) A perspective on disgust. Psychological Review, 94: 23–41.
Schlegel, J., and Rupf, R. (2009) Attitudes toward potential animal flagship species in nature
conservation: A survey among students of different educational institutions. Journal for Nature
Conservation, 18: 278–290.
Shepardson, D. P. (2002) Bugs, butterflies, and spiders: children’s understandings about insects.
International Journal of Science Education, 24: 627–643.
Sherman, N. C., and Sherman, M. F. (1998) Sensitivity to disgust among college women by
major. Psychological Reports, 82: 1192–1194.
Tucker, M., and Bond, N. W. (1997) The roles of gender, sex role, and disgust in fear of animals.
Personality and Individual Differences, 22: 135–138.
Tybur, J. M., Lieberman, D., and Griskevicius, V. (2009) Microbes, mating, and morality:
Individual differences in three functional domains of disgust. Journal of Personality and Social
Psychology, 97: 103–122.

34

Tybur, J. M., Lieberman, D., Kurzban, R., and DiScioli, P. (2013) Disgust: Evolved function and
structure. Psychological Review, 120: 65–84.
Wagler, R., and Wagler, A. (2012) External insect morphology: A negative factor in attitudes
toward insects and likelihood of incorporation in future science education settings. International
Journal of Environmental and Science Education, 7: 313–325.
Wilcove, D. S., and Master, L. L. (2005) How many endangered species are there in the United
States? Frontiers in Ecology and the Environment, 3: 414–420.

35

CHAPTER THREE
DRAWINGS OFFER INSIGHTS ABOUT MENTAL MODELS OF INSECTS

36

ABSTRACT
Insects are widely disliked despite the many ecosystem services they provide. Negative
attitudes toward insects promote harmful environmental practices and hinder conservation
efforts. To challenge this negativity, it is necessary to understand existing attitudes and the
knowledge that underlays them. One way to do this is to study the content of mental models.
Mental models are internal representations of external phenomena that are used to reason, make
inferences, conduct thought experiments, and anticipate future events. Drawings reflect qualities
of mental models such as content and organization. We collected drawings of insects from
individuals across different levels of entomological and general scientific expertise, ranging from
non-science undergraduate students to professional entomologists. Insect features included in the
drawings were indexed and a principal components analysis was applied to find underlying
variation in patterns of features. This analysis resulted in the emergence of two components that
each resembled a distinct insect – a legged but wingless “crawling” insect, and a winged but
legless “flying” insect. These two models of insect were depicted at all levels of entomological
expertise, suggesting that the way that insects move about impacts how they are conceptualized
in mental models. From a teaching and learning perspective, this suggests that comparing flying
versus crawling may be a more approachable way to characterize insects to the public, rather
than traditional taxonomic classifications. The major difference between novice and expert
(entomologists) drawings was that experts’ drawings contained more features and were more
accurate on average. The most commonly depicted features across all participants were distinct
head, antennae, long-shaped body, legs, and wings. This study provides insights about the
content and organization of people’s mental models of insects and shows how mental models
become more complex as knowledge develops.

37

INTRODUCTION
Insects are extremely important to the ecology of the natural world. They provide
valuable ecosystem services such as the pollination of many flowering plants, the aeration of
soil, and the burial of dung. One study estimated that wild insects provide ecosystem services
worth approximately $57 billion per year in the United States alone (Losey & Vaughan, 2006).
While some insects do cause distress and represent legitimate danger by biting and stinging (e.g.
disease transmission by mosquitoes), the vast majority of the beetles, ants, flies, dragonflies,
caterpillars, and other insects that people typically encounter are beneficial. Despite their many
benefits, studies have indicated that negative attitudes toward insects and other arthropods are
extremely widespread (Byrne et al., 1984; Kellert, 1993; Bixler & Carlisle, 1994; Driscoll, 1995;
Schlegel & Rupf, 2010). People frequently respond to interactions with insects with negative
emotions such as disgust, fear, alarm, and dislike (Bjerke et al., 1998; Smithsonian Institute,
2007; Gerdes et al., 2009). Extreme examples of this include phobias and delusory parasitosis
(Hinkle, 2000). We argue that it is important for educators to recognize and challenge negative
perceptions of insects, as they result in a variety of negative consequences.
Negative attitudes toward insects may hamper conservation efforts for invertebrates.
People are less willing to contribute money toward conservation of species that they dislike
(Martín-López et al., 2007), and in general, people tend to have a higher affinity for mammals
and other vertebrates than for insects (Bjerke et al., 1998; Schlegel & Rupf, 2010). A potential
consequence of this is that little of the available conservation funding goes toward preserving
insects (Leather, 2009; Leather, 2013). Butterflies are the notable exception to this trend (Shlegel
& Rupf, 2010), however, butterflies are perceived differently than other insects, and many
people do not identify butterflies as even being insects (Kellert, 1993). Additionally, negative

38

interactions with insects can detract from people’s overall experience of nature (Bixler et al.,
1994). Thus, it follows that they may be less likely to spend their time and money on outdoor
recreation (Bixler & Floyd, 1997). Another example of a negative consequence that occurs as a
result of negative bias toward insects is the misuse of pesticides (Shipley & Bixler, 2016).
Baldwin et al. (1998) surveyed Florida residents, and reported that a majority of respondents
stated that merely seeing a live insect in their home was sufficient reason to apply pesticide. This
means that there does not even have to be an infestation occurring for people to apply pesticides
to their homes. A similar survey, conducted by Potter and Bessin (1998) in Lexington, KY, also
found that the sight of just a couple of insects inside the home would prompt most participants to
spray, regardless of whether the insect was harmful or not. This is important because the misuse
of pesticides has negative consequences both for the environment and for public health (Shipley
& Bixler, 2016). Most relevant to our study, negative conceptions toward insects likely form an
impediment to teaching and learning about them. This has been acknowledged in previous works
on entomological education (Matthews et al., 1997; Smithsonian Institute, 2007; Wagler &
Wagler, 2011).
Learning is a constructive process whereby knowledge is sorted and categorized,
connections are made between new and previously existing knowledge, and knowledge is
organized into representational structures termed mental models (Bransford, 2000). Mental
models are dynamic internal representations of external phenomena (Jones et al., 2011).
Individuals use their mental models to reason, make inferences, conduct thought experiments,
and anticipate future events. Mental models do not represent exact replicas of external objects or
processes; rather, they are simplified and may not be entirely accurate (Jones et al., 2011). They
are modified by an individual’s experiences, knowledge, and attitudes. Mental models are

39

dynamic in that they change over time as new information is learned (Jones et al., 2011). They
also affect how and whether new information is incorporated. Though a collection of facts may
be presented to a learner, there is no guarantee that the information will be incorporated into the
learner’s mental model. This is particularly so when new information contradicts the previously
existing mental model (Jones et al., 2011). When people are confident in their mental model, in
that they have a strong conviction that their model is correct, they tend to become more resistant
to alternative viewpoints that contradict their beliefs (Lord et al., 1979; Jones et al., 2011). This
implies that people who have established negative attitudes toward insects are less likely to learn
new things about them that might change their views. Additionally, previous research shows that
when students feel negatively toward a topic, their interest in that topic, as well as motivation to
learn about it, is lowered (Holstermann et al., 2009). Positive interest and motivation are
extremely important for effective learning (Schiefele, 1991). Thus, widespread negative
conceptions of insects present a problem for successful entomological education.
The goal of this study was to investigate the content and organization of people’s mental
models of insects. The examination of mental models is challenging, since they cannot be
directly measured or observed – they exist only in the mind of an individual. Previous studies
have utilized techniques such as written surveys, interviews, diagramming, and estimating
relatedness between model components in order to evaluate mental models (Rowe & Cooke,
1995; Diaz, 2009). Another methodology that has been used to assess mental models is by asking
subjects to draw them (Dove et al., 1999; Shepardson, 2002; Diaz, 2009; Bartoszeck et al., 2011;
Bartoszeck & Bartoszeck, 2012; Libarkin et al., 2015). Drawings are visual representations of
conceptual knowledge, and as such, they can help to inform teaching practices by illuminating
student understanding (Dove et al., 1999; Van Meter & Garner, 2005; Ainsworth et al., 2011). In

40

addition, capturing data in the form of drawings is valuable because they are quick and efficient
to collect, and have the additional advantage of being robust to barriers of language or literacy.
Drawings reflect characteristics of mental models including knowledge categorization
and organization. The action of making a drawing requires the participant to interact with their
mental model of an object or process (Van Meter & Garner, 2005). Even the action of making a
drawing can inform the configuration of a mental model (Van Meter & Garner, 2005; Quillin &
Thomas, 2014). In accordance with the generative theory of drawing construction (Van Meter &
Garner, 2005), Quillin and Thomas (2014) provide a diagram of the relationship between internal
mental processes and the creation of a drawing (Fig. 3.1). Though drawings have been used as a
means of unpacking student understanding and examining mental models, most studies that
analyze drawings rely on qualitative methods such as thematic content analysis (Dove et al.,
1999). While qualitative methods are extremely valuable for deriving important insights,
quantitative methods are also needed because they offer a more objective approach and are often
quicker to implement. Studies that have utilized quantitative methods to analyze drawings
provide real insights into student thinking (e.g. Bowker, 2007; Diaz, 2009; Libarkin et al., 2015).
Previous studies have used drawings to probe mental models of insects in children (Shepardson,
2002; Bartoszeck et al., 2011), college students (Diaz, 2009), and adults (Bartoszeck &
Bartoszeck, 2012). For example, Bartoszeck et al. (2011) scored children’s drawings of insects
as belonging to hierarchical levels depending on the features included and their accuracy. In this
way, Bartoszeck et al. (2011) were able to discern that children build up their mental models of
insects starting with the most prominent features (body, wings, etc.). Shepardson (2002) … Diaz
(2009) collected insect drawings from college students in an effort to understand their mental
models of insects. She scored drawings based on the presence and configuration of salient insect

41

features, with accurate and complete insects receiving the highest scores. This approach was
combined with interviews and proved insightful for characterizing aspects of college students’
mental models of insects. For example, results showed that students are often unaware that
insects such as butterflies possess eyes and legs. We sought to apply a more rigorous quantitative
approach, applying the method used by Libarkin et al., (2015), that includes a factor analysis to
evaluate student drawings of the greenhouse effect. Similar to Diaz (2009), the authors first
deconstructed each drawing into its salient features, and scored them based on presence (1) or
absence (0). They then applied exploratory factor analysis, which resulted in the detection of four
distinct models that students hold about the greenhouse effect. In our study, we applied this
technique to drawings of insects to detect underlying mental models.
Though previous studies have given us insights about mental models of insects in people
who do not have much entomology training, what seems to be missing from the literature is an
examination of mental models in people who have developed expertise in science and
entomology. Comparisons of experts and novices within a discipline are useful for exploring
mental models, since experts represent the culmination of the learning process. Thus, experts can
be seen as representing maximum knowledge and novices representing minimum. By comparing
expert and novice mental models, it is possible to gain insights into the learning process. Experts
not only have greater knowledge of their discipline than novices, but they also organize their
knowledge differently (Bransford, 2000). For example, experts engage in more efficient
“chunking” of information that allows them to hold more information in their working memory
(Bransford, 2000). Thus, we would expect insect experts to have mental models of insects that
differ from the mental models of novices. Additionally, experts tend to be more abstract in their
thinking than novices, and are more effective at solving problems within their discipline. Expert-

42

novice comparisons are frequently collected in studies of conceptual change, since experts, who
began as novices at some point, clearly think differently and approach problems in their area of
expertise differently than novices (Chi et al., 1981).
This study will shed light on people’s ideas of what constitutes or defines an insect. In
this study, we attempted to answer the following two research questions:
1) Can factor analysis applied to salient features included in drawings of insects discern
underlying model(s) of insects that contain qualities of mental models?
2) How do drawings of insects differ between experts and novices, and do they suggest
major differences in their mental models?
We predict that drawings of insects do represent characteristics of underlying mental
model(s) of insects, and that drawings created by experts will differ from those created by
novices.
METHODS
Participants & data collection
We followed all human subjects protocols as approved by the institutional review board
of our university. In order to capture a spectrum of entomological and general science expertise,
we collected drawings of insects from three broad populations: 1) Attendees at the conference of
a professional entomological society, hereafter referred to as ENTs, 2) Attendees at the
conference of a professional geological society, hereafter referred to as GEOs, and 3) NonScience undergraduate students enrolled in an environmental science course at a large
Midwestern university, hereafter NSUs. ENTs and GEOs were assumed to have an equivalent
level of general science expertise (Table 3.1). ENTs were assumed to have a higher degree of

43

entomological expertise than either GEOs or NSUs (Table 3.1). NSUs were assumed to have
lower degrees of both science and entomological expertise (Table 3.1).

Table 3.1. Relative differences in science and entomological expertise across our three sampled
populations.
Entomological
Population
Science Expertise
Expertise
ENT
High
High
GEO
High
Low
NSU
Low
Low

The ENT and GEO groups were sampled during the same year, about one month apart.
The NSUs who comprised our novice group were sampled the following year. At the time of the
NSU data collection, no examination of the ENT or GEO drawings had yet occurred.
Participation in this research was voluntary and did not impact the students’ grades in any way.
The drawing task consisted of two pages, each with an empty rectangle covering most of
an 8.5”x11” page. In order to facilitate a related study (see Chapter 4), participants were
originally instructed to make two drawings: One of an insect that they considered disgusting on
the first page, and one of an insect that they did not consider to be disgusting on the second page.
Below the rectangle, on the same page, participants were prompted to identify the type of insect
they had drawn. Additionally, participants were verbally instructed that they could draw any
insect of their choice, and those who inquired as to whether they could draw spiders or other
non-insect arthropods were instructed that it was acceptable to do so. This was because these
animals are commonly confused with insects, and we assume that other participants who were
inclined to draw a non-insect may not have asked. For this current Chapter Three, only the “not
disgusting” insect drawings were evaluated, since these drawings included a larger variety of

44

insects than the “disgusting” insect drawings. The disgusting insect drawings will be discussed in
an upcoming chapter (see Chapter 4).
For the ENT and GEO groups, the drawing task was administered at one location within
the main exhibit hall at each conference. Thus, all ENT and GEO participants were convenience
sampled as they passed by the researchers’ stations. Scholars at both conferences were offered a
snack as an incentive for participation. Time was not limited for these participant groups, but the
drawing task generally took about 2-5 minutes. For the NSUs, the drawing task was administered
on the first day of class before any instruction had taken place. Students were given five minutes
to complete the drawing task. Since NSU participation occurred during regular class time,
students were not offered an incentive, but completed the survey voluntarily.

Coding & frequency of insect features
We followed the method of Libarkin et al. (2015) to code each drawing. This coding
scheme was established iteratively by three of the co-authors, and involved breaking the drawing
down into its salient features (Fig. 3.1). These salient features were morphological insect features
such as legs, antennae, wings, etc. All three raters began by examining a subset of ten randomly
selected drawings together and listing prominent features from each drawing. Each feature was
given a definition that was mutually agreed upon by all three authors. It should be noted that all
insects share a suite of six basic morphological features: legs, antennae, eyes, mouth, head, and
body segmentation. Most, though not all, insect species also have wings during the adult stage,
but there are some common exceptions, such as earwigs and worker ants. Because of this, we did
not include wings as one of the essential features, though we did code for them. Also, though all
insects possess a head, a thorax, and an abdomen, with many insects it is difficult to distinguish

45

the division between the thorax and abdomen. For this reason, we coded for the presence of a
head, which is usually very easy to distinguish in most insects, separately from thorax and
abdomen. The thorax and abdomen we lumped together as one variable that we called “body
shape,” which had two levels: Oval and elongate.
Another aspect of the insect body that we coded for was segmentation. As mentioned
above, it can be difficult to interpret the relative positions of two of the major body segments, the
thorax and abdomen. Therefore, we coded a body as being segmented if there was any sort of
clear division between thorax and abdomen (Fig. 3.1). In addition, since on insects the individual
body parts themselves are divided into segments, we also coded a body as segmented if there
were clear divisions between these intra-segment segments.

Wings
Antennae

Eyes

Stinger

Mouth

Segmentation
Distinct
Head

Legs

Figure 3.1. An example drawing of an insect with the coding scheme applied.

46

Through an iterative process, we settled on 11 major morphological features of insects to
include in our coding (Table 3.3). These 11 features naturally contain the six basic features that
are characteristic of all insects. Once the coding scheme was finalized, each author
independently coded 20 drawings, after which an intraclass correlation was conducted. The
average measures intraclass correlation across the three raters was 0.92 (min. = 0.91 and max. =
0.93). An intraclass correlation close to 1.0 suggests that each coder’s analysis of the drawings is
consistent with the others’. Given this high level of agreement, the remaining drawings were
coded by one rater. See Figure 3.1 for an example of coding; see Figure 3.2 for a representation
of a complete insect as well as example drawings from each participant group. All statistical
analyses were performed using SPSS (Version 23.0, SPSS Inc., Chicago, IL).
We calculated the frequency at which each coded feature was drawn within each
participant group separately. Since our sample sizes for each participant group were unequal,
frequency was calculated as a percentage of the total population for each group.

Total number of features
We compared the mean total number of features included in the drawings (out of a
possible 11) for each subject group. Since ENTs are expected to have greater knowledge of
insect morphology, we anticipated that on average, their drawings would contain more details
(features) than those of novices. For this variable, the highest possible score is eleven, and the
lowest possible score is one, since the body-shape variable can be considered here to count as a
constant – everyone drew a body on their insect. After satisfying the assumption of normality,
we compared the three participant groups using a one-way ANOVA, followed by post-hoc
Tukey’s HSD.

47

(A)

(B)

Complete Insect:

Contains 6 basic features common to all insects

• 
• 
• 
• 
• 
• 

Head
Legs
Eyes
Mouth
Antennae
Segmentation

(D)

(C)

Figure 3.2. Schematic representation of a complete insect (A), along with example drawings
from ENTs (B), GEOs (C), and NSUs (D).
Completeness of drawn insects
As part of our analysis, we wanted to explore how often the drawings included an
accurate scientific model of an insect. We consider that a drawing of a “complete” insect should
include the six major features described above: Legs, eyes, mouth, antennae, head, and
segmentation (Fig. 3.2). If we assign one point for including each of these features, we can
calculate a “completeness” score for each participant, with the highest possible score being six
and the lowest score zero. These data failed the assumption of normality; consequently, we

48

utilized a non-parametric test, the Kruskal-Wallis H test, to determine whether Insect
Completeness Score varied across our participant groups.

Correlations between insect features
In order to evaluate whether the data were suitable for PCA, we ran correlational analyses
on drawing data from all three expertise levels combined. We conducted Spearman’s rank-order
correlations between the different insect features, which were coded as dichotomous (1 = feature
present, 0 = absent). The intent of the correlational analysis was to determine whether potentially
interesting relationships might exist between insect features in the drawings, indicating the
possible existence of meaningful patterns that represent underlying mental models. Once relevant
criteria had been met, we ran a principal components analysis on all of the drawings.

Principal components analysis
We utilized the technique of principal components analysis (PCA) to reduce
dimensionality among the drawings. This analysis was conducted on drawings from all three of
the participant groups, which increased the total sample size to a level acceptable for principal
components analysis. We did not include a rotation, in order to allow variables to cluster on more
than one factor if necessary (Libarkin et al., 2015).

Model scores & expertise
Based on the groupings suggested by our PCA, we computed scores indicating the degree
of alignment of a drawing with the crawling and flying models. Model scores were calculated for
each participant for each of the two models by giving one point for the presence of each of the

49

model components. Thus, two scores were computed for each drawing, one for the crawling
model and one for the flying model. For the crawling model scoring, a particular drawing could
attain a maximum of six points, one for each of the six features which loaded on to the first
factor in our PCA (Head, Legs, Eyes, Mouth, Segmented Body, No Wings). Thus, if a drawing
included legs, eyes, mouth, and a segmented body, the drawing would receive a score of 4 for the
crawling model. For the flying model, a drawing could attain a maximum of 5 points
(Segmentation, Eyes, Mouth, Wings, Long Body).
Since participant group is a nominal variable, we utilized dummy coding in order to make
the variable more suitable for MANOVA. We chose the ENT group as the baseline for
comparison, and thus created two dummy variables: one that we called ENT vs. NSU, which
represents the difference between ENT and NSU participants; and one that we called ENT vs.
GEO, which represents the difference between the ENT and the GEO participants.
In order to determine whether model scores differed statistically across our three
participant groups, we conducted a MANOVA with two dependent variables (Crawling Model
Score; Flying Model Score) and one independent variables (participant group).

RESULTS
Participants & data collection
The entomologists we surveyed were mainly undergraduates, graduates, faculty, and
professionals in entomology. A total of 97 ENTs participated in the survey, though eight cases
were excluded due to missing/incomplete data. In an effort to restrict the sample only to those
who had had training in entomology, and thus a high level of entomological expertise, we also
excluded those participants who acknowledged having taken fewer than three entomology

50

courses in their lifetime, amounting to 18 total individuals excluded. Thus, we report analyses of
drawings from 79 ENTs (Table 3.2).
A total of 104 GEOs participated in the survey. In an effort to restrict the sample to only
those individuals who had a science background but no formal instruction in entomology, we
excluded six participants who answered “yes” when asked if they had previously taken courses
in entomology. In addition, we also excluded two participants who were not either working
toward or in possession of a science degree. Finally, four additional participants were excluded
from the analysis because they did not complete a drawing. Thus, we report on data from 92
GEOs. These individuals were mainly undergraduates, graduate students, faculty, and
professionals in geoscience disciplines (Table 3.2). We surveyed a total of 112 NSUs. Eleven
surveys were excluded from the analysis due to missing/incomplete data. Consequently, we
report on a total of 101 NSU drawings (Table 3.2).
Table 3.2. Descriptive statistics for each participant group. ENT participants were expert
entomologists; GEO participants were expert geologists (non-entomologist scientists); NSU
participants were non-science undergraduate students.
Expertise
Level

N

ENT

79

GEO

92

NSU

101

Age
(M±SD)

Gender

Highest Degree Held

In Entomology:
16% Bachelors’ in progress
37.2% Female
38.9 ± 14.1
26% holding Bachelors’
62.8% Male
18% holding Masters’
40% holding PhD
In Geoscience:
29% Bachelors’ in progress
48.4% Female
34.02 ± 14.7
23% holding Bachelors’
51.6% Male
23% holding Masters’
25% holding PhD
All participants working
53.5% Female
19.81 ± 2.75
toward Bachelor’s degrees
46.5% Male
in non-science fields

51

Coding scheme and frequency of insect features
The frequencies at which each of our 11 coded features occurred across each participant
group are listed in Table 3.3. Eight out of the 11 coded insect features were drawn by greater
than 50% of all participants combined – these features are, in order of cumulative percentage
from greatest to least: Distinct head, antennae, legs, long-shaped body, wings, segmentation,
eyes, and mouth. Oval-shaped body, coded as one variable together with long-shaped body, was
also drawn by many participants and is included in Table 3.3.
Seven insect features were drawn by over 60% of ENTs, indicating that these features are
most prominent in the mental models of experts: Distinct head, antennae, legs, long-shaped
body, wings, eyes, and segmentation. The most commonly drawn feature within the ENT group,
as well as when compared to the other two groups, was the distinct head. Within the GEOs, the
most commonly drawn variable was legs. GEOs and ENTs both drew legs on their insects more
commonly than NSUs. The most commonly drawn variable both within the NSUs and compared
to the other two groups was antenna. We also noted that ENTs were twice as likely to include
eyes, mouths, and segmentation on their insects than either of the other two groups.

Total number of features
Skewness and kurtosis values divided by their respective standard errors were both less
than ±1.96, and the Q-Q plot was normal, leading us to conclude that the data are suitable for
parametric analysis (Field, 2013). A one-way ANOVA comparing the total number of features
drawn for each group was significant: F(2,269) = 17.257, p < 0.001. A post-hoc Tukey’s HSD
was applied to the data and showed that on average, ENTs differed significantly from both GEOs
and NSUs (p < 0.001). NSU and GEO participants did not differ significantly from each other (p

52

= 0.42). The mean (± SD) total number of features drawn by ENTs was 5.29 (±1.78), compared
with GEOs at 3.87 (± 1.65) and NSUs at 4.17 (± 1.54). Thus, experts drew one to two more
features on average than either of the non-expert groups (Figure 3.3).
Table 3.3. Percentage of individuals from each participant group who drew each morphological
insect feature. Features that are part of the basic correct insect are italicized. Features are listed
in order of greatest-to-least cumulative frequency.
Coded Feature

NSU

GEO

ENT

1. Distinct head

64.4

66.3

87.3

2. Antennae

78.2

57.6

67.1

3. Body Shape: Long

72.3

59.8

65.8

Body Shape: Oval

27.7

40.2

34.2

4. Legs

53.5

68.8

68.4

5. Wings

61.4

45.7

64.6

6. Eyes

36.6

34.8

64.6

7. Segmentation

30.7

31.5

64.6

8. Mouth

11.9

16.3

34.2

9. Stinger

5.9

2.2

7.6

10. Hair

2.0

3.3

1.3

0

0

3.8

11. Pincers / Cerci

53

Mean
Number
Features
MeanTotal
Total Number
of Features

10
10

8
6
44
22
0

NSU
NSU

GEO
GEO

ENT
ENT

Figure 3.3. Entomologists (ENT) include more features on average than non-science
undergraduates (NSU) or non-entomologist scientists (GEO) in drawings of non-disgusting
insects.
Completeness of drawn insects
We consider that a drawing of a correct, “complete,” insect should include the six major
features found on all insects: Legs, eyes, mouth, antennae, head, and segmentation. We
calculated an Insect Completeness Score for each participant; scores ranged from zero to six,
with one point given for the inclusion of each of the correct features. The mean (± SD) Insect
Completeness Score was highest in the ENT participants, at 3.86 (± 1.49), followed by the GEO
group (2.79 ± 1.49) and the NSU group (2.75 ± 1.42), which were very similar to each other. The
Insect Completeness Score data did not meet the assumption of normality across the three
participant groups, thus, we proceeded with non-parametric testing for these data. A KruskalWallis H test was significant, Χ2(2) = 26.81, p < 0.001, indicating that Insect Completeness
Score differed across the three groups. The mean rank Insect Completeness Score was highest
for ENTs (174.47), followed by GEOs (121.16), followed by NSUs (120.77) (Fig. 3.4).

54

Though ENT participants drew a greater number of correct insect features than GEO or
NSU participants, on average, ENTs created drawings that only contained 64% of a complete
insect. Principal components analysis allows us a mechanism for exploring this variability.

Mean
Insect
Score
Mean
InsectCompleteness
Completeness Score

6	

*	
4	

2	

0	

NSU
NSU

GEO
GEO

ENT
ENT

Figure 3.4. Entomologists (ENT) draw insects that are more complete than non-science
undergraduates (NSU) or non-entomologist scientists (GEO). Complete Insect Score was
calculated by examining all non-disgusting insect drawings and giving one point for each of six
features essential to all insects: head, eyes, mouth, antennae, segmentation, and legs. Results
from a Kruskal-Wallis H Test showed that Complete Insect Score differed across the three
groups (p < 0.001). Data are shown as mean ± SEM of Complete Insect Score, representative of
drawings by 79 ENT, 92 GEO, and 101 NSU participants.

Correlations between insect features
In order to further investigate relationships between insect features, we conducted pointbiserial correlational analyses between the 11 coded features (Table 3.4). Results showed a
strong negative relationship between legs and wings (r = −0.43, p < 0.001), indicating that an
individual who drew legs was less likely to also include wings, or vice versa. Drawing a distinct
head on an insect correlated with the inclusion of other features such as legs (r = 0.5, p < 0.001),

55

a segmented body (r = 0.39, p < 0.001), and eyes (r = 0.26, p < 0.001). Additionally, eyes
correlated strongly with mouth (r = 0.45, p < 0.001), as well as segmentation (r = 0.33, p <
0.001). In general, correlational analyses showed that many features were significantly correlated
(p < 0.05) between features, with many having Pearson coefficients above 0.3 (Table 3.4). We
then utilized a PCA in order to explore the variability in the data

Head

Legs

Antennae

1

0.50*
*

−0.06

1

Wings

Body
Shape

Segmentation

Eyes

Mouth

Sting
-er

Hair

−0.18*
*

−0.14*

0.39**

0.26*
*

0.13*

0.07

0.04

−0.02

−0.15*

−0.43*
*

−0.24*
*

0.28**

0.28*
*

.25**

0.08

0.11

0.06

1

0.22**

0.14*

0.04

−0.15
*

−0.11

−0.003

−0.02

−0.11

1

0.26**

−0.13*

0.04

0.02

−0.12*

0.07

−0.07

1

0.18**

0.07

−0.01

0.001

−0.08

0.002

1

0.32*
*

0.21**

0.06

0.13*

−0.02

1

0.45**

0.12*

0.20*
*

0.02

1

0.04

0.18*
*

0.05

1

0.14*

−0.02

1

0.19*
*

Hair

Stinger

Cerci

Mo
-uth

Eyes

Segm
-ent

Body
Shape

Wings

Antennae

Legs

Head

Table 3.4. Point-biserial correlations between insect features for all three participant groups
combined. One asterisk indicates that the relationship is significant at the level of p < 0.05; two
asterisks indicate significance at p < 0.01.
Cerci

1

56

Principal components analysis
We chose to include the following eight coded insect features in the PCA: Body shape,
Segmentation, Head, Eyes, Mouth, Legs, Antennae, and Wings. These variables were chosen
because they were included in the drawings of at least 10% of at least one group (i.e., ENT,
GEO, NSU) of participants (Table 3.3). The resulting Kaiser-Meyer-Olkin (KMO) measure of
sampling adequacy was 0.64, which meets the recommended threshold of 0.6 (Costello &
Osborne, 2005). Additionally, Bartlett’s test of sphericity was significant (Χ2= 376.53, df = 28,
p < 0.001). The diagonals in the anti-image correlation matrix were all above the recommended
threshold of 0.5, and communalities after the initial extraction were also all above 0.5, which is
above the recommended threshold of 0.3 (Costello & Osborne, 2005). Components were
extracted at eigenvalues greater than one, and in addition, we examined the scree plot as an
indication of how many components to extract (Catell, 1966). This analysis resulted in three
components, accounting for 62.9% of the variation in the data set. However, the eigenvalue for
the third component is quite low (1.1), and the scree plot is indicative of a two-factor solution.
When we repeated the analysis with two as the set number of components to extract, the
antennae variable failed to load on either component, indicating that it does not belong in the
solution. For this reason, we decided to drop antennae from the analysis.
We repeated the PCA, using the following seven features: Segmentation, long-shaped
body, head, eyes, mouth, wings, and legs. The KMO measure for this analysis was 0.641.
Bartlett’s test was again significant (Χ2= 349.4, df = 21, p < 0.001), all anti-image correlation
coefficients were above 0.5, and all communalities were above 0.3, with most above 0.5. Two
components were extracted with eigenvalues of 2.33 and 1.50. Together, these two components
accounted for 54.7% of the variation in the data. The scree plot is shown in Figure 3.5. The

57

component matrix is reported in Figure 3.5; variables with loadings less than ±0.32 are
suppressed.
We can consider these two components as two distinct models of insect (Libarkin et al.,
2015) (Fig. 3.5). The first model (component) contains segmentation, a distinct head, legs, no
wings, eyes, and a mouth. We will refer to this as the crawling model. The second model
(component) contains a long-shaped body, segmentation, wings, eyes, and a mouth. We will
refer to this as the flying model. We tested the internal consistency of these two models by
calculating Cronbach’s alpha. For the crawling model (6 features) Cronbach’s alpha was 0.66,
which implies acceptable internal consistency. For the flying model (5 features) Cronbach’s
alpha was 0.44, which is less robust.

Model A

WINGS	

Model B

(Schematic Representation)

(Schematic Representation)

Component Matrix
Segmentation 0.61
Distinct head 0.72
Eyes 0.60
Mouth 0.50
Legs 0.79
Wings −0.42
Example Aligned Drawings

Component Matrix
Long body 0.67
Segmentation 0.32
Eyes 0.51
Mouth 0.42
Wings 0.63

Example Aligned Drawings

Figure 3.5. Schematic and drawn representations of the crawling model and flying model based
on the results of our PCS. Factor loadings for each feature are also given; loadings below ±0.32
are suppressed.

58

Model scores & expertise
In order to address our second research question and determine whether a relationship
existed between alignment with the models and expertise, we conducted a MANOVA. Our
independent variable was participant group, which is coded as nominal. Our two dependent
variables were Crawling and Flying Model Scores, both of which were ratio/interval. Skewness
and kurtosis values over respective standard errors were mostly below ±1.96 for all combinations
of dependent and independent variables, indicating a normal or near-normal distribution.
Levene’s test was not significant for scores for either model, thus satisfying the assumption of
homogeneity of variance. In addition, Box’s test was non-significant (Box’s M = 2.496,
F(6,1407046.4) = 0.411, p = 0.872), satisfying the assumption of homogeneity of covariance
matrices. The MANOVA examined the relationship between the scores for each model and
participant group. Overall, the MANOVA was significant (F (2,268) = 11.01, p < 0.001, Roy’s
Largest Root = 19.175, partial η2 = 0.125). Participant group had a significant main effect on
both the Crawling Model Score (F(2,269) = 11.538, p < 0.001, partial η2 = 0.079) and the Flying
Model Score (F(2,269) = 16.445, p < 0.001, partial η2 = 0.109). Given the significant results of
this test, we proceeded with Tukey’s HSD post hoc test (Field, 2013).
Mean scores for the crawling model were significantly higher in ENT participants than in
either NSUs (p < 0.001) or GEOs (p = 0.004). NSU participants had slightly lower scores but did
not differ statistically from GEU participants for the crawling model (p = 0.269) (Table 4, Figure
5). For the flying model, post-hoc testing showed that ENT participants again had significantly
higher scores than either NSU (p < 0.001) or GEO participants (p < 0.001). GEO and NSU
participants did not differ statistically in their mean alignment with the flying model (p = 0.352),
though NSUs had slightly higher scores (Table 3.5, Figure 3.6).

59

Table 3.5. Mean model scores, correctness score, and total number of features for each subject
group.
Crawling Insect
Flying Insect
Correctness
Total Features
Score (M±SE)

Score (M±SE)

Score (M±SE)

(M±SE)

ENT

3.54 ± 0.188

2.94 ± 0.153

3.86 ± 0.175

5.29 ± 0.201

GEO

2.73 ± 0.174

1.88 ± 0.126

2.76 ± 0.156

3.87 ± 0.172

NSU

2.36 ± 0.165

2.13 ± 0.117

2.75 ± 0.142

4.17 ± 0.153

Mean Number of Features

6
5
4
3
2
1
0
NSU

GEO

ENT

NSU

Crawling Model

GEO

ENT

Flying Model

Figure 3.6. Mean model score for each of the two models from participants from each group.
Error bars denote standard error.
DISCUSSION
In the present study, we analyzed drawings in order to shed light on the content and

60

organization of mental models of insects. We also investigated how mental models differ
between people with advanced training in entomology (ENTs), people with advanced training in
science but not entomology (GEOs), and people with minimal or no training in entomology or
other science (NSUs). Mental models are employed when a person thinks and reasons about a
topic (Jones et al., 2011). When constructing a mental model, one must select which components
to include, incorporate those components together, and organize them into a coherent framework
(Van Meter & Garner, 2005; Quillin & Thomas, 2014). Drawings capture aspects of mental
models, since the act of making a representational drawing requires soliciting a mental image.
Thus, drawings can be thought of as imperfect external representations of dynamic internal
representations. Drawings are commonly used in science education research to study how people
think about explanatory processes, such as the greenhouse effect. The results of the present study
support previous works that have indicated that drawings can also be used to study how people
think about objects or entities, such as insects (e.g. Shepardson, 2002; Diaz, 2009). In addition,
this study demonstrates the utility of quantitative analyses for detecting patterns in drawings
from many individuals.
Two underlying models emerged from our PCA: (1) A crawling model with legs and
specifically without wings, and (2) a flying model with wings. Individuals from all three
participant groups created drawings that aligned with the two models. This suggests that these
basic means of characterizing insects – as crawling or flying – may represent two distinct mental
models of insect. These two models are interesting because they represent the two spaces that
people primarily observe and interact with insects – on surfaces and in the air. Experts (ENTs)
had higher scores for each model than either of the novice groups (GEO or NSU). This is likely
due to the fact that entomologists drew more features on their insects than the other groups.

61

Beyond the importance to insect education, this result also shows that breaking down drawings
into salient features and conducting PCA is a useful method for determining underlying
relationships between variables.
The discovery of the two insect models lends support to previous findings indicating that
flying and crawling are useful distinctions for characterizing insects to the public. For example,
Baldwin et al. (2008) surveyed Florida residents about their use of pesticides, and utilized the
categories of “flying” and “crawling” to specify different categories of pests. Crawling pests
were cockroaches and ants; flying pests were mosquitoes, wasps, and flies. Results showed that
participants viewed flying pests as being more harmful to human health than crawling pests, but
cited crawling pests as the larger issue in terms of prevalence and the cost of control.
Interestingly, the authors suggest in their discussion that the reason crawling insects were
perceived as an actionable threat by the majority of participants was “probably due to a low
tolerance for crawling insects, rather than actual damage caused by the pests” (p. 77, Baldwin et
al., 2008). Furthermore, many of the flying insects drawn by NSUs were butterflies; likely not
perceived as threatening or damaging.
A similar computer-based analysis of children’s drawings of insects (see Appendix I) also
yielded two models, one of which is a crawling model, the other of which is a flying model. This
is extremely interesting because it indicates that crawling and flying insects are seemingly
distinct from each other not only across levels of knowledge, but also across the life span.
Additionally, the flying versus crawling dichotomy also emerged in a recent analysis of
children’s comments regarding various different types of insects (Breuer et al., 2015). Cluster
analysis of the comments revealed a divide between insects that traveled mainly by crawling or
were perceived to crawl, and insects that were mainly perceived in flight, with the flying insects

62

receiving more positive comments and inspiring less disgust than the crawling insects (Breuer et
al., 2015). This lends support to the idea that insects’ different modes of travel (flying vs.
walking) may be a useful way to characterize them in educational settings rather than traditional
taxonomy-based classification. Though it seems clear from both Baldwin et al. (2008) and
Breuer et al. (2015) that crawling insects in general are more negatively regarded than flying
insects, the role of affect is outside the scope of the current study. Chapter Four of this
dissertation addresses this relationship.
Though the PCA allows us to reduce dimensionality in our data, we also gleaned
important insights about how expert and novice mental models of insect morphology differ by
examining correlations between features and frequency of features. Both correlational and PCA
showed an overall negative relationship between the inclusion of legs and wings in the drawings.
Novice (GEO or NSU) participants who drew one were more likely than experts to not include
the other. Interestingly, uniquely among NSU participants, wings were more frequently included
than legs. We noted that many of the drawings with prominent wings were butterflies, which are
an entrée point into the world of insects for many people. Diaz (2009) collected drawings of
insects from a similar population of college students and interviewed a subset of those students
about the contents of their drawings. Qualitatively, Diaz (2009) makes note of the fact that
students who drew insects with wings but not legs, usually butterflies and lady beetles, did so
because they believed that these insects do not possess legs. She also notes that some students
did not draw eyes or a mouth on their insects because they believed that insects also do not
possess those features. This result was also apparent in our data set – frequency of eyes and
mouths was much lower for the two novice groups than for experts. In addition, the occurrence
of eyes and mouth was positively correlated with expertise. One of the most defining

63

characteristics of insects is their segmentation. Diaz (2009) reported that most students drew
their insect with distinct segments (including the head). Segmentation was positively correlated
with expertise in our data, and experts drew segmentation more frequently than novices. We
suggest that eyes, mouths, and segmentation are more advanced concepts that become more
commonly integrated into mental models of insects upon the achievement of expertise. Experts
included a greater number of features, on average, than novices. One simple explanation for this
is that experts possess greater knowledge of insect morphology than novices. For example, the
majority of extant insects possess both eyes and mouths, and all insects have segmented bodies –
we would expect entomologists to know that.
Drawings of insects created by people who had expertise in entomology differed in
several important ways from drawings created by novices. Drawings of insects by trained
entomologists (ENTs) were more detailed and more complete than drawings made by people
who did not have this training. Interestingly, participants with general training in science (GEOs)
but who lacked expertise in entomology created drawings that were similar in both completeness
and the amount of detail to drawings made by college students who had neither science nor
entomology training (NSUs). This implies that experience specifically in entomology, rather than
science in general, is responsible for changes to mental models of insects. Previous studies also
acknowledge that expertise development changes the way one connects with content – experts
organize and process information within their discipline differently than non-experts do, and
their knowledge is much greater (Chi et al., 1981). One possible reason that entomologists
included more features in their drawings than novices is concerned with working memory.
Working memory is the site of interaction between attention, perception, and memory. It
“temporarily stores information as part of the performance of complex cognitive tasks”

64

(Baddeley, 1992). The contents of working memory, including mental models, are used in
completing cognitive tasks. Thus, working memory is engaged when creating a drawing because
the individual must rely on what they already know. In children, working memory capacity was
found to be a significant predictor of the number of dog-like features included in drawings of
dogs (Panesi & Morra, 2016). Previous work has shown that experts have greater and more
complex working knowledge within their discipline than novices (Chi et al., 1981; Peters, 2000).
Though entomologists drew insects that were more accurate than novices, their drawings
largely fell short of the full six features that we designated as defining a correct and complete
insect: Head, segmentation, eyes, mouth, antennae, and legs. Does this mean that the
entomologists we surveyed were not aware that insects possess all of these features? If the
drawings are reflective of mental models, and mental models encompass one’s knowledge, then
does it follow that features missing from the drawing are also missing from the individual’s
knowledge? Though previous research has shown that even seasoned experts may hold
misconceptions about certain aspects of their field (Lewis & Linn, 1994), we feel that this is
unlikely in this case. The six essential insect features are extremely important to the functioning
of an insect and are covered even in the most basic of introductory entomology courses. Why
else might it be that drawings by ENTs are not all perfectly correct and complete representations
of insects? We suggest that experts do know that insects have all of these features but for some
reason do not include all of them at once. This could be either the result of a conscious decision
not to include certain features in the drawing, or an unconscious choice. Because we did not
interview our subjects, we cannot go back and ask them if they made a conscious decision not to
draw all the features, or whether they were not aware of it at the time.

65

An additional interesting finding is that the majority of the drawings we analyzed
depicted insects in their adult (mature) life stage; this was true across all three participant groups.
Life stage was not referred to in the survey prompt. Anecdotally, we did notice the presence of
larvae such as maggots when participants were prompted to draw an insect that they thought was
disgusting, though again depictions of adults far outnumbered depictions of larvae. This suggests
that mental models of insects are primarily composed of adult insects, rather than larval forms. In
addition, larval forms seem to be associated with negative affect, in this case, disgust. In support
of this is Wagler and Wagler’s (2012) study, which documented more negative attitudes toward
larval forms than adult forms. Another explanation for the paucity of larval insects depicted in
our study is that people are simply less aware of larval insects. This makes sense because
immature insects are more likely to be found hidden within or underneath objects, since their
exoskeletons are less hardened than in adulthood, and thus they need more protection. It is
unclear whether the relative lack of immatures in our sample is due to lack of awareness or
dislike. Perhaps ontologically, adult insects are viewed as the most important because they are
the most visible in our world. We suggest that educational programs should place more emphasis
on teaching the immature forms of insects. Insects are most vulnerable in the larval stage,
particularly larval stages that occur in water, which are often utilized as bioindicators (Merritt et
al., 2008). Larval forms of insects are also important food sources for human populations across
the globe. If people are not aware of the existence of larval insects, they may be less inclined to
protect and conserve them.

66

Limitations
One important thing to note is that though we collected drawings, which are
representative of knowledge, we did not quantify nor experimentally manipulate the amount,
relevance, or complexity of our participants’ knowledge of insects and other arthropods. Rather,
we assumed that the experts’ greater formal training, as indicated by their considerable amount
of undergraduate- and/or graduate-level entomological coursework, would be a satisfactory
indicator of expertise. Precisely defining expertise levels poses a challenge to researchers since
there is much variation in how expertise levels are defined across the literature (see Hoffman,
1998). In general, scholars tend to group those who have had significant formal education in
their fields toward the expert end of the continuum, and those who have had little or no formal
training toward the novice end (Chi et al., 1981; Lewis & Linn, 1994). We acknowledge that this
is a grossly oversimplified classification, since there are additional factors that contribute to
expertise other than coursework (Hoffman, 1998). Our three participant groups were sampled
from three diverse populations (non-science college students, non-entomologist scientists, and
entomologist scientists) and are different enough from each other that a valid comparison across
expertise levels is warranted.

Conclusions
Individuals enter a learning environment with their own unique pre-existing mental
models that they use to interpret their experiences and plan their behavior (Jones et al., 2011).
Teaching and learning in the sciences must be thoughtfully planned with attention paid to
facilitating positive attitudes, since positive interest and motivation are important for effective
learning (Schiefele, 1991). Evidence suggests that fact-based science learning is not adequate in

67

regards to improving attitudes (Koballa & Crawley, 1985). Thus, it is important to assess
students’ existing attitudes and knowledge prior to instruction. For instance, in the field of
entomology, learners may be constrained by negative conceptions of insects. Our study suggests
that people organize their mental models of insects based on whether they are observed to
primarily crawl or fly. We believe that this may be a more approachable way to characterize
insects to the public, rather than emphasizing traditional taxonomic classifications. In addition,
we found that experts’ mental models are more likely to include fine-grained features such as
segmentation, eyes, and mouths; with novices’ mental models generally sticking to large-grained
features such as wings. One final major insight from this study is that, across the spectrum of
expertise we sampled, mental models of insects tend to include the mature/adult life stage rather
than the immature stage.

68

REFERENCES

69

REFERENCES
Baddeley, A. (1992) Working Memory. Science, 255(5044): 556.
Baldwin, R. W., Koehler, P. G., Pereira, R. M., and Oi, F. M. (2008) Public perceptions of pest
problems. American Entomologist, 54(2): 73–79.
Bartoszeck, A., and Bartoszeck, F. K. (2012) Educação de jovens e adultos: Estudo exploratório
do conceito de insetos. Estação Científica (UNIFAP), 2(1): 33–41.
Bartoszeck, A., da Silva, B. R., and Tunnicliffe, S. D. (2011) Children’s concept of insect by
means of drawings in Brazil. The Journal of Emergent Science, 2: 17-24.
Bixler, R. D., Carlisle, C. L., Hammitt, W. E., and Floyd, M. F. (1994) Observed fears and
discomforts among urban students on field trips to wildland areas. The Journal of Environmental
Education, 26(1): 24–33.
Bixler, R. D., and Floyd, M. F. (1997) Nature is scary, disgusting, and uncomfortable.
Environment and Behavior, 29(4): 443–467.
Breuer, G. B., Schlegel, J., Kauf, P., and Rupf, R. (2015) The importance of being colorful and
able to fly: Interpretation and implications of children’s statements on selected insects and other
invertebrates. International Journal of Science Education, 37(16): 2664–2687.
Chi, M. T. H. (2008) Three types of conceptual change: Belief revision, mental model
transformation, and categorical shift. In S. Vosniadou (Ed.), Handbook of research on
conceptual change (pp. 61-82). Hillsdale, NJ: Erlbaum.
Chi, M. T. H., Feltovich, P. J., & Glaser, R. (1981) Categorization and representation of physics
problems by experts and novices. Cognitive Science, 5(2): 121–152.
Diaz, A. E. (2009) Undergraduates’ mental models about insect anatomy and insect life cycles.
M.S. Thesis, California State University, Fullerton. Retrieved 27, January, 2017 from
http://search.proquest.com.proxy1.cl.msu.edu/docview/305177877/abstract/51ACC4349EF94C3
APQ/1.
Gerdes, A. B. M., Uhl, G., and Alpers, G. W. (2009) Spiders are special: Fear and disgust evoked
by pictures of arthropods. Evolution and Human Behavior, 30(1): 66–73.
Hinkle, N. C. (2000) Delusory parasitosis. American Entomologist, 46(1): 17–25.
Hoffman, R. R. (1998) How can expertise be defined? Implications of research from cognitive
psychology. In Exploring expertise (pp. 81-100). Palgrave Macmillan UK.

70

Holstermann, N., Grube, D., and Bögeholz, S. (2009) The influence of emotion on students’
performance in dissection exercises. Journal of Biological Education, 43(4): 164–168.
Jones, N., Ross, H., Lynam, T., Perez, P., & Leitch, A. (2011) Mental models: An
interdisciplinary synthesis of theory and methods. Ecology and Society, 16(1): 46.
Kellert, S. R. (1993) Values and perceptions of invertebrates. Conservation Biology, 7(4): 845–
855.
Koballa, T. R., and Crawley, F. E. (1985) The influence of attitude on science teaching and
learning. School Science and Mathematics, 85(3): 222-232.
Leather, S. R. (2013) Institutional vertebratism hampers insect conservation generally; not just
saproxylic beetle conservation. Animal Conservation, 16(4): 379–380.
Lewis, E. L., and Linn, M. C. (1994) Heat energy and temperature concepts of adolescents,
adults, and experts: Implications for curricular improvements. Journal of Research in Science
Teaching, 31(6): 657–677.
Libarkin, J. C., Thomas, S. R., and Ording, G. (2015) Factor analysis of drawings: Application to
college student models of the greenhouse effect. International Journal of Science Education,
37(13): 2214-2236.
Lord, C. G., Ross, L., and Lepper, M. R. (1979) Biased assimilation and attitude polarization:
The effects of prior theories on subsequently considered evidence. Journal of Personality and
Social Psychology, 37(11): 2098–2109.
Losey, J. E., and Vaughan, M. (2006) The economic value of ecological services provided by
insects. BioScience, 56(4): 311–323.
Martín-López, B., Montes, C., and Benayas, J. (2007) The non-economic motives behind the
willingness to pay for biodiversity conservation. Biological Conservation, 139(1-2): 67–82.
Merritt, R. W., Cummins, K. W., & Berg, M. B. (2008) Introduction to the Aquatic Insects of
North America: Text. Dubuque, IA: Kendall Hunt Publishing.
Osborne, J. W., & Costello, A. B. (2005) Best practices in exploratory factor analysis. Practical
Assessment, Research & Evaluation, 10(7): 1-9.
Panesi, S., & Morra, S. (2016) Drawing a dog: The role of working memory and executive
function. Journal of Experimental Child Psychology, 152: 1–11.
Peters, H. P. (2003) From information to attitudes? Thoughts on the relationship between
knowledge about science and technology and attitudes toward technologies. In M. Dierkes & C.
von Grote (Eds.) Between Understanding and Trust. The Public, Science and Technology (pp
265–286). The Netherlands: Harwood Academic Publishers.

71

Potter, M. F., & Bessin, R. T. (1998) Pest control, pesticides, and the public: Attitudes and
implications. American Entomologist, 44(3): 142–147.
Quillin, K., & Thomas, S. (2015) Drawing-to-learn: A framework for using drawings to promote
model-based reasoning in biology. CBE-Life Sciences Education, 14(1): es2.
Rowe, A. L., & Cooke, N. J. (1995) Measuring mental models: Choosing the right tools for the
job. Human Resource Development Quarterly, 6(3): 243—255.
Schiefele, U. (1991) Interest, learning, and motivation. Educational Psychologist, 26(3-4): 299–
323.
Shepardson, D. P. (2002) Bugs, butterflies, and spiders: Children’s understandings about insects.
International Journal of Science Education, 24(6): 627–643.
Smithsonian Institution Office of Policy and Analysis. (2007) The buzz on insects: a study of
attitudes and knowledge among visitors to the National Museum of Natural History. Retrieved
15, January, 2017 from
https://www.si.edu/content/opanda/docs/Rpts2007/07.09.Insects.Final.pdf.
Thompson, F., & Logue, S. (2006) An exploration of common student misconceptions in
science. International Education Journal, 7(4): 553-559.
Wagler, R., & Wagler, A. (2011) Arthropods: attitude and incorporation in preservice elementary
teachers. International Journal of Environmental and Science Education, 6(3): 229–250.
Van Meter, P., & Garner, J. (2005) The promise and practice of learner-generated drawing:
Literature review and synthesis. Educational Psychology Review, 17(4): 285–325.

72

CHAPTER FOUR
IMPACTS OF AFFECT AND EXPERTISE ON DRAWINGS OF DISGUSTING AND NOTDISGUSTING INSECTS

73

ABSTRACT
Negative attitudes toward insects are very common among the public. Attitudes have
behavioral, cognitive, and affective components. The goal of this study was to explore the
affective response to insects, disgust, in the context of mental models of insects in drawings. We
collected drawings of “disgusting” and “not disgusting” insects from individuals across different
levels of entomological and general scientific expertise, ranging from non-science undergraduate
students to professional entomologists. Insect features included in the drawings were indexed
and a principal components analysis was applied to find underlying variation in patterns of
features. This analysis resulted in the emergence of two components that each resembled a
distinct insect – a legged but wingless “crawling” insect, and a winged but legless “flying”
insect. In addition, all participants completed a disgust survey indicating their relative level of
disgust stimulated by insects and non-insect arthropods. We utilized a multiple linear regression
in order to determine if insect-associated disgust could be predicted by the combination of insect
features included and subject group of the individual participants. The analysis indicated that
approximately 35% of the variation in disgust was attributable to subject group (expertise) and
the degree to which drawings aligned with the crawling insect model. We conclude that crawling
insects may be regarded as being more disgusting than flying insects, that educational and
professional backgrounds play a role in mediating disgust toward insects, and that examining
drawings of insects can be a useful tool to shed light on affect associated with insects.

74

INTRODUCTION
Emotions, or affective responses, play an important role in cognition, in that they provide
information in addition to semantic knowledge (Clore & Huntsinger, 2007). Affective reactions
are important to the formation of attitudes (Breckler, 1984). For example, people tend to assign
higher values to objects that inspire positive emotions and lower values to objects that inspire
negative emotions (Clore & Huntsinger, 2007). Insects are invertebrate animals that stimulate a
variety of emotions in humans, ranging from awe and happiness to disgust and fear. Though
insects are intrinsically valuable to human beings due to their important positions in the world’s
food chains and the ecosystem services (Losey & Vaughan, 2006), multiple studies show that
negative attitudes toward insects are very common among the public (Kellert, 1993; Bixler et al.,
1994; Bjerke et al., 1998; Schlegel & Rupf, 2010). In order to encourage public valuation of
insects, it is important to challenge negative attitudes toward insects.
Previous studies indicate that people with greater knowledge of something are less likely
to perceive it as harmful (Boete & Moran, 2016). For example, professionals who work with bats
are less likely to perceive them as dangerous (Boete & Moran, 2016). Along this line,
comparisons of experts and novices within the same discipline have also shown that experts
perceive their object of study differently than novices (Chi et al., 1981). Both visual processing
(Tanaka & Curren, 2001) and knowledge organization (Chi et al., 1981) differ between experts
and novices. Evidence from brain-imaging studies suggests that experts become emotionally
habituated to their object of expertise (Cheng et al., 2007). This means that experts exhibit more
muted emotional responses to their object of study than non-experts.
In this study, we explored the representation of emotion in drawings of insects from
individuals with different levels of insect and scientific expertise. Examining drawings has been

75

shown to be a useful way to probe understanding about a topic of interest (Dove et al., 1999;
Ainsworth et al., 2011; Libarkin et al., 2015, etc.). In addition, drawings can also be used to
detect or evaluate emotion (e.g. Oster & Crone, 2004; Löfström & Nevgi, 2012), though this is
much less common in the literature than using drawings to evaluate understanding. Drawings of
insects have previously been utilized in studies of children’s understandings about insects
(Shepardson, 2002; Snaddon & Turner, 2007), though to our knowledge, no prior studies exist
that utilize drawings of insects created by adults.
In particular, we focused on the emotion of disgust. Disgust is a negative reaction that is
both emotional and physiological, and is associated with a characteristic facial expression that
involves scrunching one’s nose and raising the upper lip (Haidt et al., 1994). Disgust activates
regions of the brain associated with hunger and eating, and the word disgust itself means “bad
taste” (Phillips et al., 1997). Disgust is thought to be an evolutionary mechanism to promote
avoidance of potential disease-causing agents such as feces, blood, and rotten meat (Oaten et al.,
2009), though disgust is also elicited in response to moral transgressions and violation of sexual
norms (Tybur et al., 2009). Disgust associated with the perception of insects is linked to
pathogen avoidance (Lorenz et al., 2014). Sensitivity to disgust is usually measured using a scale
(Likert-type survey) (Haidt et al., 1994; Tybur et al., 2009; Lorenz et al., 2014).
Research questions
This study builds off the previous chapter of my dissertation, Chapter Three, in that it
utilizes data from the same participants. Chapter Three describes an investigation in to mental
models of insects using principal components analysis (PCA) of salient features included in
drawings of regular (not disgusting) insects. The PCA resulted in the characterization of two
major models of insects that occurred in the drawings: a crawling model, and a flying model.

76

This present work investigates the same set of drawings as Chapter Three, but considers
an additional dimension of cognition: affect. In addition to the drawings of regular (not
disgusting) insects that were explored in Chapter Three, this current work also reports on an
additional set of drawings in which the same participants drew an insect they considered
disgusting.
The main goal of this chapter is to examine how disgust influences people’s internal
representations of insects in the context of expertise. Drawings created by insect experts,
professional entomologists, were compared to drawings created by insect novices, who were
non-science major undergraduates. In addition, both were compared to a third group of
individuals who were professional scientists outside of the field of entomology. In Chapter
Three, it was found that drawings of experts aligned more closely with the crawling and flying
models than the drawings of non-experts. Building off this finding, this current Chapter is framed
around the following research questions:
1) Do insect experts differ in their reported level of insect-associated disgust from
insect novices?
2)

How do drawings of disgusting insects differ from drawings of not disgusting
insects?

3) Can disgust toward insects be predicted by examining drawings of insects?

METHODS
Participants & data collection
We collected drawings from three populations of individuals at three levels of insect
expertise: non-science undergraduate students (NSUs), scientists with no specialized training in

77

entomology (GEOs), and scientists with extensive training in entomology (ENTs). ENTs were
individuals participating in the annual meeting of a major American professional entomological
society, and thus were mainly students and professionals in entomology. We assume that our
ENT population has greater expertise about insects than either GEOs or NSUs, and
approximately equivalent general science expertise as our GEO participants. GEOs were
individuals participating in the annual meeting of a major American professional geological
society, and thus were mainly students and professionals in the geosciences. GEOs were
assumed to have greater expertise in general science than NSUs, but similar expertise about
insects. Both ENTs and GEOs were presumed to have greater expertise in general science than
NSUs. Descriptive statistics for our participants can be found in Table 1 (reproduced from
Chapter Three of this dissertation). Participants were asked to draw one insect they did not find
disgusting, and on a separate page, one insect that they did find disgusting. Participants were also
asked to record the type of insect they had drawn. In addition, we collected demographic data
such as gender and highest degree held.

Insect disgust survey
In addition to the two drawings, participants also completed a brief Likert-scaled survey
of insect-associated disgust (Lorenz et al., 2014; Chapter Two of this dissertation). Participants
ranked their disgust level in response to each survey item on a 4-point Likert scale with the
following rankings: “Not at all disgusting,” “Somewhat disgusting,” “Very disgusting,” and
“Extremely disgusting.” The disgust survey consisted of the following nine survey items:
(1) Finding a scorpion in your shoe.
(2) Seeing an ant crawl across the floor.

78

(3) Finding a bug in your shirt.
(4) Watching a spider make its web.
(5) Eating a chocolate-covered cricket.
(6) Feeling a mosquito bite you.
(7) Accidentally touching a spiderweb.
(8) Watching a centipede crawl across your leg.
(9) Seeing a cockroach run across a countertop.
Insect disgust scores were calculated for each participant by weighting the different
response options from one (not at all disgusting) to four (extremely disgusting), and taking the
sum of all responses from each participant. Thus, the lowest possible disgust score was nine, and
the highest possible score was 36. We calculated Cronbach’s alpha on all nine items for all
participants together in order to confirm the reliability of the survey. The disgust survey showed
good reliability with Cronbach’s α = 0.866. We conducted a Kruskal-Wallis H test in order to
determine whether disgust differed significantly across our participant groups.
Previous studies on disgust report higher mean disgust levels for females relative to
males (e.g. Davey, 1994). Since the gender data did not ascribe to a normal distribution, we used
a non-parametric test to determine whether disgust differed between the genders. Another
suggestion from the literature (e.g. Curtis et al., 2004) is that disgust sensitivity declines with
age. We collected data from GEO and ENT participants across the lifespan, from age 19 to age
78. In contrast, NSU participants all reported their ages as between 18 and 25 years, as was
expected because NSU participants were undergraduate students. We compared mean insect
disgust score for each participant group in four age categories: 18-25, 26-40, 41-55, and 56-78
years. Due to severely unequal sample sizes as well as the absence of NSU data for ages over 25,

79

we were unable to perform a statistical analysis to determine whether age affects insect disgust
outside of expertise. Thus, we created a graph of the data in order to make a qualitative
comparison.

Drawings, insect models, & insect model scores
Both sets of drawings were coded by breaking each drawing into salient insect features
and recording their presence or absence. A detailed description of our sampling and coding
methodology can be found in Chapter Three. Eleven distinct insect features were coded: head,
antennae, legs, body shape (oval vs. elongate), wings, eyes, segmentation, mouth, stinger, hair,
and pincers/cerci. The average frequency at which each feature occurred for both the disgusting
and not disgusting drawings was calculated and compared between disgusting and not disgusting
drawings across all participants, in order to gain an idea of overall trends.
In the previous analysis of the non-disgust drawings (Chapter Three), PCA was utilized
without rotation in order to reduce dimensionality in the data. From this analysis, two
components emerged. These two components each reflected a distinct model or type of insect.
The first component represents a crawling insect, and the second component represents a flying
insect (Fig. 4.1). These two components can be considered to represent common themes in
people’s representations of insects. Scores were calculated for each model for each participant,
indicating the degree to which the features included in the model were present in each drawing.
For example, the crawling model contains six features: head, segmentation, eyes, mouth, legs,
and wings lacking. Thus, for their crawling model score, each participant was given one point for
the inclusion of all six features. Note that for this model, the inclusion of the “wings absent”
feature means not drawing wings. Thus, it is possible for a drawing to score a maximum of six

80

points if they drew a head, segmentation, legs, eyes, and a mouth, and no wings. The minimum
possible score for the crawling model is negative one. For the flying model, the maximum score
is five since the model contains five features: Long-shaped body, segmentation, wings, eyes, and
mouth. Thus, the maximum possible score is five, and the minimum is zero. In addition, a
“complete insect score” was calculated for all drawings. The complete insect score estimates the
degree to which a drawing represents an accurate scientific model of an insect (or other
arthropod). A complete insect should possess a distinct head, legs, eyes, mouth, antennae, and
segmentation.

Crawling Model

Flying Model

(Schematic Representation)

(Schematic Representation)

WINGS	

Figure 4.1. Crawling and flying insect models as detected through principal components
analysis (Chapter Three).
Correlational analysis
In order to explore potential relationships between insect disgust and the drawings, pointbiserial Pearson correlations were conducted between disgust score, gender, crawling model
score, flying model score, and participant group for both the disgusting and not disgusting
drawings. Participant group, as a nominal variable representing relative expertise, was dummycoded in order to be appropriate for the analysis. The ENT group, the experts, was chosen as the
baseline with which to compare the other participants, the non-experts. Thus, two dummy

81

variables were created, ENT vs. NSU (compares ENTs to NSUs), and ENT vs. GEO (compares
ENTs to GEOs).

Multiple regression
A multiple linear regression analysis was conducted in order to determine whether
expertise and model scores could predict insect-associated disgust. This test was run for the notdisgusting insect drawings only. This is because a greater variety of insects was represented in
the not-disgusting drawings than in the disgusting drawings. Thus, the not-disgusting drawings
were more representative of a typical insect. In addition, sample size for the not-disgusting
drawings was greater than for the disgusting drawings. Before conducting the regression, tests
were performed to verify normality, linearity, homogeneity of variances, homoscedasticity,
outliers, and influential points (Field, 2013). We conducted a multiple linear regression to predict
Insect Disgust Score from Subject Group (ENT, GEO, NSU) and Crawling Model Score. The
Crawling Model Score was included because it correlated significantly with Insect Disgust Score
in the not-disgusting drawings. Flying Model Score also correlated significantly with Insect
Disgust Score in the not-disgusting drawings, but the correlation coefficient was relatively low;
thus, it was not included in the regression model. Crawling Model Score is considered to be
continuous, since it is a quantitative measure of the number of features that align with the model.
Age was not included as a predictor in the analysis since our participant groups were different in
terms of age distribution. Subject Group, as a categorical variable, was dummy-coded into the
same two dummy variables utilized in the correlational analyses. Again, the ENT group was
selected to use as a baseline.

82

RESULTS
Participants & data collection
Full descriptive statistics for our participants are reproduced in Table 4.1. A total of 294
individuals were surveyed. Two hundred and forty-one drawings of disgusting insects, and 272
drawings of not-disgusting insects were analyzed. For the drawings of disgusting insects, out of a
total of 112 NSUs, 100 completed an applicable drawing. Twelve individuals either did not
complete a drawing, or their drawing was not applicable to the study (e.g. drawing does not
depict an insect or other arthropod). Out of a total of 96 GEOs, 80 completed a drawing; and out
of a total of 87 ENTs, 65 completed an applicable drawing. ENTs numbers are lower because
many did not complete a drawing, instead writing in that they did not find any insects to be
disgusting. For the drawings of not-disgusting insects, out of the total 112 NSUs, 101 completed
an applicable drawing. For GEOs, 92 completed an applicable drawing; for ENTs, 80 individuals
completed an applicable drawing.
Table 4.1. Descriptive statistics for each participant group (ENT, GEO, NSU). Reproduced
from Chapter Three of this dissertation.
Expertise
Level

N

Age (M±SD)

ENT

79

38.9 ± 14.1

GEO

92

34.02 ± 14.7

NSU

101

19.81 ± 2.75

Gender

Highest Degree Held

In Entomology:
16% Bachelors’ in progress
37.2% Female
26% holding Bachelors’
62.8% Male
18% holding Masters’
40% holding PhD
In Geoscience:
29% Bachelors’ in progress
48.4% Female
23% holding Bachelors’
51.6% Male
23% holding Masters’
25% holding PhD
All participants working toward
53.5% Female
Bachelor’s degrees in non46.5% Male
science fields

83

Insect disgust survey
Out of a total of 294 participants, 285 completed the disgust survey. Disgust scores
declined with entomological expertise, with the mean disgust score (± SD) for ENTs at 14.25 (±
4.15), for GEOs at 18.37 (± 5.24), and for NSUs at 22.76 (± 5.73) (Fig. 4.2).

25	

Mean	Disgust	Score	

23	
21	
19	
17	
15	
13	
11	
9	

NSU	

GEO	

ENT	

Figure 4.2. Mean insect disgust score (± SE) across our three sampled populations. The y-axis
begins at 9 because the lowest possible disgust score was 9. NSUs had the highest mean score,
followed by GEOs, with ENTs having the lowest perceived disgust associated with insects.
Since the disgust data for each gender did not conform to a normal distribution, in order
to determine whether reported disgust differed by gender, a Mann-Whitney U test was utilized,
combining data from all three participant groups. This test was significant (U = 5346.5, p <
0.001), indicting that males and females differ in their reported levels of disgust across all three
participant groups. Mean rank score was higher for females (167.19) than males (109.15).

84

All NSU participants reported their age as between 18 and 25 years. In comparison, GEO
and ENT participants between the ages of 18 and 25 generated mean disgust scores well under
that of NSU participants. Thus, the higher mean insect disgust score exhibited by the NSU group
cannot be attributed solely to their collective lower age. However, greater age may still be
confounding because older ENT and GEO participants have more career experience. Expertise
level was not explicitly measured, though it is likely safe to assume that expertise increases with
age/experience. However, note from the graph (Fig. 4.3) that for the ENT group, mean disgust is
greatest in the 18-25 years group, with the next three age groups (26-78 years) scoring similarly
to each other. So, it would appear from this graph that insect disgust is present in experts at every
age, but is relatively constant after age 25.

Mean Disgust Score

25
20
15
10
5

89

41

13

22

34

16

16

11

10

NSU

GEO

ENT

GEO

ENT

GEO

ENT

GEO

ENT

0
18-25 years

26-40 years

41-55 years

56-78 years

Age

Figure 4.3. Mean disgust score plotted by age category for each participant group. NSU = nonscience undergraduates; GEO = non-entomologist scientists; ENT = entomologists. The quantity
of participants reporting age in each category for each group is indicated within the bottom of
85

Figure 4.3 (cont’d)
each bar. Note that all NSU participants reported their age as between 18 and 25 years. GEO and
ENT participants between the ages of 18 and 25 generated mean disgust scores well under that of
NSU participants.
Drawings, insect models, & insect model scores
There were both differences and similarities in the frequency of inclusion of the different
insect features between the disgusting and not disgusting insect drawings (Table 4.2). Nondisgusting drawn insects were nearly 30% more likely to have wings, 18% more likely to have
antennae, 13% more likely to have a long-shaped body, and 8% more likely to have a distinct
head than the disgusting drawings. Disgusting drawings were 20% more likely to have legs, 13%
more likely to have an oval-shaped body, nearly 10% more likely to have a mouth, and 6% more
likely to have pincers/cerci. By contrast, features that were included at similar rates in both types
of drawings (within 5% of each other) were eyes, segmentation, stinger, and hair. In order to
qualitatively explore the types of insects drawn, we created word clouds (www.tagul.com)
composed of all recorded insects drawn for both the disgusting and not disgusting insects (Figure
4.4). Common among the not disgusting insects were butterflies, lady beetles, dragonflies, ants,
bees, and praying mantids. Common among the disgusting insects (though some are non-insect
arthropods) were cockroaches, ticks, maggots, centipedes, mosquitoes, flies, and spiders. Note
that bees and ants were common among both disgusting and not disgusting entries.
Calculated scores for both the crawling model and flying model, along with complete
insect score, were higher among experts (Table 4.3). This was found to be significant for the not
disgusting drawings in the previous chapter. Since higher scores are indicative of having more
features in common and thus greater alignment with the models, this suggests that ENTs as a
group drew insects that were more closely aligned to the models than either GEOs or NSUs.

86

Model scores were tested between disgusting and not disgusting drawings, with the crawling
model scores generally larger for disgusting than not-disgusting drawings across all three
participant groups (Table 4.3).
Table 4.2. Frequencies of coded insect features (all participant groups combined). In the
Difference (ND – D) column, positive values indicate that the feature was more common in the
not disgusting drawings; negative values indicate that the feature was more common in the
drawings of disgusting insects.
Not Disgusting
Disgusting
Difference
Insect Feature
Drawings (ND)
Drawings (D)
(ND – D)
1.
2.
3.
4.
5.

Distinct Head
Antennae
Legs
Wings
Body Shape: Long
Body Shape: Oval
6. Eyes
7. Segmentation
8. Mouth
9. Stinger
10. Hair
11. Pincers/Cerci

71.7%
68.0%
63.0%
57.0%
66.2%
33.8%
44.1%
40.8%
19.9%
5.1%
2.2%
1.1%

63.8%
49.2%
83.5%
27.8%
52.4%
47.6%
47.6%
41.7%
29.5%
5.5%
4.7%
7.1%

7.9
18.8
-20.5
29.3
13.8
-13.8
-3.5
-0.9
-9.6
-0.4
-2.5
-6.0

Table 4.3. Mean model scores (±SD) for disgusting and not disgusting drawings across
participant groups.
Not Disgusting
Disgusting
Crawling Model Flying Model
Crawling Model Flying
Model
NSU
2.36 (±1.66)
2.13 (±1.18)
3.37 (±1.09)
1.89 (±1.11)
GEO
2.73 (±1.66)
1.88 (±1.21)
3.14 (±1.23)
1.86 (±1.14)
ENT
3.54 (±1.67)
2.94 (±1.36)
3.71 (±1.40)
2.34 (±1.43)

87

Figure 4.4. Two word clouds based upon participants’ recordings of the type of insect they had
intended to portray, showing dominant types of (A) not-disgusting (left) and (B) disgusting
(right) insects drawn. The relative sizes of the words indicate their frequency in the dataset.

Correlational analysis
Point-biserial Pearson correlations were conducted between Insect Disgust Score,
Crawling Model Score, Flying Model Score, as well as the dummy-coded participant group
variables, ENT vs. NSU and ENT vs GEO. For the not-disgusting insect drawings, Insect
Disgust Score was negatively correlated with Subject Group and the Crawling Model Score
(Table 4.4). In addition, the analyses showed that Insect Disgust Score correlated significantly
with the Crawling Model Score in the not-disgusting insect drawings only. Insect Disgust Score
did not correlate with the Crawling Model Score in the disgusting insect drawings. Disgust score
correlated weakly with Flying Model Score for both the disgusting and not disgusting drawings.

88

Table 4.4. Pearson correlations between disgust score and model scores calculated from
drawings of (A) not-disgusting insects, and (B) disgusting insects. * Significant correlation at
level of 0.05. **Significant correlation at level of 0.01.
Crawling
Flying
(A) NOT
Disgust
ENT vs
ENT vs
Gender
Model
Model
DISGUSTING
Score
NSU
GEO
Score
Score
Disgust Score

1

ENT vs. NSU

0.50**

-0.01

0.37**

-0.32**

-0.16**

1

-0.55**

0.10

-0.21**

-0.09

1

0.02

-0.04

-0.22**

1

-0.18**

-0.18**

1

0.41**

ENT vs. GEO
Gender
Crawling
Model Score
Flying Model
Score

1

Gender

Crawling
Model
Score

Flying
Model
Score

-0.04

0.37**

-0.10

-0.17**

-0.55**

0.09

-0.01

-0.08

1

0.03

-0.14*

-0.07

1

-0.04

-0.02

1

0.57**

(B)
DISGUSTING

Disgust
Score

ENT vs.
NSU

ENT vs.
GEO

Disgust Score

1

0.50**
1

ENT vs. NSU
ENT vs. GEO
Gender
Crawling
Model Score
Flying Model
Score

1

Multiple regression
A multiple linear regression was conducted using the forced entry method to predict
Insect Disgust Score from Subject Group and Crawling Model Score. In this analysis, only the
not-disgusting insect drawings were utilized. Prior to the analysis, participant group, a nominal

89

variable, was dummy-coded in to two dichotomous variables, ENT vs. GEO and ENT vs. NSU.
In this way, the regression will compare these two groups to the ENT group, who represent a
high level of expertise. In order to avoid issues with missing data, missing cases were deleted
listwise.
Despite the fact that females on average scored higher than males on the Insect Disgust
Survey, gender was not utilized as a predictor of insect-associated disgust in the regression
analysis. This is because, though females tend to report stronger emotions than males, brainimaging studies have revealed similar activity for males and females during emotional
processing (Schienle et al., 2005; Bluhm, 2013). The tendency of females to report greater
emotion thus may be a product of culture rather than a biological difference between the genders
(Bluhm, 2013). In addition, the purpose of this study, as stated in the research questions, is to
determine whether disgust can be predicted from relative expertise about insects and drawings of
insects, not about effects of gender. For these reasons, gender was not included in the regression
model.
Before conducting the regression, in order to make sure the data were suitable for the
analysis, test were conducted to verify the assumptions of linearity, outliers and influential
points, normality, multicollinearity, independent errors, homogeneity of variances, and
homoscedasticity (Field, 2013).
Linearity between the Insect Disgust Score and Crawling Model Score was examined
utilizing a scatterplot. The scatterplot suggested that a linear relationship exists between Insect
Disgust Score and Crawling Model Score, meeting the assumption of linearity. The dummy
variables were not tested for linearity because as dichotomous variables the relationship is
automatically linear. In testing for outliers and influential points, two individual cases had

90

standardized residuals that were >2.5. Field (2013) states that as long as the total number of cases
with standardized residuals > 2.5 is less than one percent of the total data set, then these cases
should not unduly influence the model. In support of this, Cook’s distance was 0.004, which is
less than 1, and the mean Mahalanobis distance was 3.0, which is less than the critical values of
alpha for chi-square analysis for two degrees of freedom (Field, 2013). The correlation matrix
was examined for multicollinearity issues. Crawling Insect Score correlated significantly with
the NSU dummy variable (r = −0.21, p < 0.001). Though these two predictor variables correlated
significantly with each other, the correlation coefficient is relatively small, and thus, issues with
collinearity are unlikely. Crawling Score did not correlate with the GEO dummy variable (r =
−0.041, p = 0.501). In addition, VIF values were all less than 10, and Tolerance values were
greater than 0.2 (Field, 2013), again indicating no problems with multicollinearity. The DurbinWatson statistic was utilized to test for independence of errors. The Durbin-Watson value was
1.951; since this number was close to two, independence of errors was assumed (Field, 2013).
Continuing to explore the data to determine if all assumptions of multiple linear
regression are met, the data were tested for normality. When the skewness and kurtosis values
were divided by their standard error, the result was less than ±1.96 for most combinations of
variables, indicating a near-normal distribution. In addition, the histogram and P-P plot were also
indicative of a normal distribution. Homogeneity of variances in Insect Disgust Scores was tested
using Levene’s test. For the ENT vs. NSU (dummy variable) disgust scores, variances based on
means were equal for both ENT and NSU participants, F(1,264) = 2.003, p = 0.158. For the ENT
vs. GEO (dummy variable) disgust scores, variances based on means were significantly different
for ENT and GEO participants, F(1, 264) = 7.876, p = 0.005. However, the variance ratio
between ENT and GEO participants was 1.538, which is fairly low, indicating that the

91

significance is likely an artifact of large and unequal sample sizes (Field, 2013). For the
Crawling Model Score, variances based on means were equal, F(6, 269) = 1.626, p = 0.140. We
conclude from this analysis that the assumption of homogeneity of variances is acceptably met
by our data. Finally, we tested for homoscedasticity by examining the scatter plot of zpred vs.
zresid. The resulting plot resembled an overall random array, indicating no issues with
homoscedasticity.
The linear regression was conducted to determine whether Crawling Model Score and
Subject Group could predict Insect Disgust Score. The result of this analysis yielded a significant
model, (F(3, 262) = 47.873, p < 0.001, R2 = 0.354, R2Adjusted = 0.347). This indicates that
approximately 35% of the variation in Insect Disgust Scores is predicted by differences in
Crawling Model Score and Subject Group. Each of the three predictor variables were significant
additions to the model, as indicated by their standardized beta values. For Crawling Model Score
(standardized β = −0.181, p < 0.001), this value indicates that as the Crawling Model Score
increases by one standard deviation (1.73), the Insect Disgust Score decreases by −0.181
standard deviations (−0.181 ✕ 6.12), or 1.11.
The dummy variable of ENT vs. NSU had a standardized β of 0.616 (p < 0.001). This
indicates that the difference in disgust score for a person in the NSU group compared to a person
in the ENT group was significant. The GEO vs. ENT dummy variable had a standardized β of
0.283 (p < 0.001), thus, the difference in disgust score for a person in the GEO group compared
to a person in the ENT group was also significant.

92

Table 4.5. Linear model of the predictors of Insect Disgust Score.
b
SE b
β

p

Constant

16.327

0.853

-

p < 0.001

Crawling Model
Score

−0.639

0.182

0.616

p < 0.001

ENT vs NSU

7.836

0.781

0.283

p < 0.001

ENT vs GEO

3.640

0.774

−0.181

p = 0.001

DISCUSSION
The results of this study clearly indicate that insect experts differ in their reported level of
insect-associated disgust from insect novices. This finding is in line with a trend, documented in
previous literature, that experts perceive their object of expertise differently from novices (Chi et
al., 1981; Boete & Moran, 2016). However, we are not able to discern from our results whether
this lower disgust is truly the result of the process of developing expertise, or whether the
individuals who choose to become entomologists tend to have lower disgust toward insects at the
outset. Nevertheless, the possibility that having knowledge and experience with insects acts to
reduce one’s negative emotions associated with them is still a plausible one. This is because
experimental studies have found that habituation can indeed influence affect toward a stimulus
(Rozin, 2008; Viar-Paxton & Olatunji, 2012), and that experts are known to become emotionally
habituated to disgusting stimuli (Cheng et al., 2007). In addition, exposure to a disgusting
stimulus across multiple contexts has been found to be more effective at reducing disgust than
exposure in one context (Viar-Paxton & Olatunji, 2012). Therefore, it seems very likely that the
ENTs greater experience and exposure to insects caused them to feel less disgust associated with
insects. Training in science, regardless of exposure to insects, appears to also mediate insect-

93

associated disgust, as our GEO participants scored approximately mid-way between NSUs and
ENTs on the Insect Disgust Survey.
Another goal of this study was to determine whether drawings of disgusting insects differ
from drawings of not disgusting insects. That is, are there qualities of insects that particularly
evoke disgust? What does a disgusting insect look like, in comparison to a non-disgusting insect?
From analysis of the drawings, we noted a clear trend that had to do with the inclusion of legs
versus wings; legs were more commonly included on disgusting insects, wings were more
commonly included on not disgusting insects, and the two features are strongly negatively
correlated with each other (Chapter Three). Though all insects have legs, save for those in their
larval or pupal life stages, participants were more likely to include them when evoking
disgusting insects. This may indicate that people are more likely to label an insect as disgusting
when they perceive it to be crawling on its legs. These results could be related to the fact that
parasites like fleas and lice are crawling pests – they do not possess wings. Disgust as an emotion
is associated with aversion to potentially damaging stimuli, thus, it is possible that people in
general have a greater aversion to crawling insects in order to facilitate avoidance of these types
of macroparasites. Disgust is very relevant to disease; in fact, disgust has been postulated to be
evolutionarily driven by the desire to avoid pathogens and other undesirable stimuli (Curtis
et al., 2004; Oaten et al., 2009; Tybur et al., 2009). We noted that the insects (and other
arthropods) that were drawn most often in the disgusting category were insects that were
associated with disease and pathogens. This aligns with a trend in the literature of diseaserelevant insects being perceived as more disgusting than disease-irrelevant insects (Curtis et al.,
2004; Prokop & Fančovičová, 2010). Curtis et al. (2004) asked people to rate the disgustingness
of contrasting pairs of disease-relevant and irrelevant stimuli. They found that disease-irrelevant

94

insects (tent caterpillars, wasp) were rated similarly to disease-relevant animals (parasitic worms,
louse). Curtis et al. (2004) attribute this similarity to the fact that the insects are also
“biologically relevant” (p. S132), in that they represent danger. The authors state, “Further work
should reveal whether the response pattern to disgust-relevant insects varies with the exposure
that people have had to different species in different ecological zones” (p. S132). In this current
study, we did not test or ask about participants’ actual behaviors associated with disgusting
versus not disgusting insects, however, one study found that people were more likely to take
action such as using pesticides upon finding a crawling pest than a flying one (Baldwin et al.,
2008).
There was no relationship between crawling model score and disgust score for the
disgusting drawings, despite the fact that some features of the crawling model (legs, mouth,
segmentation) were more common among disgusting drawings than not disgusting drawings, and
crawling insect model scores were higher for disgusting than not disgusting drawings. This may
indicate that people included these features regardless of their reported level of disgust toward
insects. A logical conclusion here is that despite individual differences in reported disgust, there
seems to be a consensus (lack of variation) in that everyone thinks the same insect features (or
same types of insects) are disgusting. This suggests that educators and researchers can better
assess levels of insect-associated disgust by asking people to draw insects that they do not
consider disgusting.
The third research question addressed in this study is, can the intensity of an individual’s
disgust toward insects be predicted by examining their drawings of insects? We tested this for
the not-disgusting drawings only. We found that participants’ crawling model scores were a
significant predictor of insect-associated disgust, in that higher crawling model scores predicted

95

lower disgust scores, though this effect was relatively small. Subject group, and thus, expertise,
was a stronger predictor of insect-associated disgust than crawling model score. Being either a
NSU or GEO were significant predictors of disgust score; each were associated with greater
disgust than the ENT participants. It is worth noting that being an NSU was the strongest
predictor of insect disgust.
We conclude that educational and professional backgrounds play a role in mediating
disgust toward insects, and that examining drawings of insects can be a useful tool to shed light
on affect associated with insects.

96

REFERENCES

97

REFERENCES
Ainsworth, S., Prain, V., and Tytler, R. (2011) Drawing to learn in science. Science, 333(6046):
1096–1097.
Baldwin, R. W., Koehler, P. G., Pereira, R. M., and Oi, F. M. (2008) Public perceptions of pest
problems. American Entomologist, 54(2): 73–79.
Bixler, R. D., Carlisle, C. L., Hammltt, W. E., and Floyd, M. F. (1994) Observed fears and
discomforts among urban students on field trips to wildland areas. The Journal of Environmental
Education, 26: 24–33.
Bluhm, R. (2013) Self-fulfilling prophecies: the influence of gender stereotypes on functional
neuroimaging research on emotion. Hypatia, 28(4): 870–886.
Boëte, C., and Morand, S. (2016) Bats and academics: how do scientists perceive their object of
study? PLOS ONE, 11(11): e0165969.
Breckler, S. J. (1984) Empirical validation of affect, behavior, and cognition as distinct
components of attitude. Journal of Personality and Social Psychology, 47(6): 1191–1205.
Cheng, Y., Lin, C. P., Liu, H. L., Hsu, Y. Y., Lim, K. E., Hung, D., and Decety, J. (2007)
Expertise modulates the perception of pain in others. Current Biology, 17(19): 1708–1713.
Chi, M. T. H., Feltovich, P. J., and Glaser, R. (1981) Categorization and representation of
physics problems by experts and novices. Cognitive Science, 5(2): 121–52.
Clore, G. L., and Huntsinger, J. R. (2007). How emotions inform judgment and regulate
thought. Trends in Cognitive Sciences, 11(9): 393-399.
Curtis, V., Aunger, R., and Rabie, T. (2004) Evidence that disgust evolved to protect from risk of
disease. Proceedings of the Royal Society of London B: Biological Sciences, 271(Suppl 4):
S131–33.
Davey, G. C. L. (1994) Self-reported fears to common indigenous animals in an adult UK
population: the role of disgust sensitivity. British Journal of Psychology, 85(4): 541–554.
Dove, J. E., Everett, L. A., and Preece, P. F. W. (1999) Exploring a hydrological concept through
children’s drawings. International Journal of Science Education, 21(5): 485–497.
Field, A. (2013) Discovering statistics using IBM SPSS statistics. Fourth ed. Sage.
Haidt, J., McCauley, C., and Rozin, P. (1994) Individual differences in sensitivity to disgust: a
scale sampling seven domains of disgust elicitors. Personality and Individual Differences, 16(5):
701–713.

98

Kellert, S. R. (1993) Values and perceptions of invertebrates. Conservation Biology, 7(4): 845–
855.
Lorenz, A. R., Libarkin, J. C., and Ording, G. J. (2014) Disgust in response to some arthropods
aligns with disgust provoked by pathogens. Global Ecology and Conservation, 2: 248–254.
Losey, J. E., and Vaughan, M. (2006) The economic value of ecological services provided by
insects. BioScience, 56(4): 311–323.
Löfström, E., and Nevgi, A. (2014) Giving shape and form to emotion: using drawings to
identify emotions in university teaching. International Journal for Academic Development,
19(2): 99–111.
Oaten, M., Stevenson, R. J., and Case, T. I. (2009) Disgust as a disease-avoidance mechanism.
Psychological Bulletin, 135(2): 303–21.
Oster, G. D., and Crone, P. G. (2004) Using drawings in assessment and therapy: A guide for
mental health professionals. Psychology Press.
Phillips, M. L., Young, A. W., Senior, C., Brammer, M., Andrew, C., Calder, A. J., Bullmore, E.
T., Perrett, D. I., Rowland, D., Williams, S. C. R., Gray, J. A., and David, A. S. (1997) A specific
neural substrate for perceiving facial expressions of disgust. Nature, 389(6650): 495–498.
Prokop, P., and Fančovičová, J. (2010) The association between disgust, danger and fear of
macroparasites and human behaviour. Acta Ethologica, 13: 57–62.
Rozin, P. (2008) Hedonic ‘adaptation’: specific habituation to disgust/death elicitors as a result
of dissecting a cadaver. Judgment and Decision Making; Tallahassee, 3(2): 191.
Schienle, A., Schäfer, A., Stark, R., Walter, B., and Vaitl, D. (2005) Gender differences in the
processing of disgust- and fear-inducing pictures: an fMRI study. Neuroreport, 16(3): 277–280.
Shepardson, D. P. (2002) Bugs, butterflies, and spiders: children’s understandings about insects.
International Journal of Science Education, 24(6): 627–643.
Snaddon, J. L., and Turner, E. C. (2007) A child’s eye view of the insect world: perceptions of
insect diversity. Environmental Conservation, 34: 33–35.
Tanaka, J. W., and Curran, T. (2001) A neural basis for expert object recognition. Psychological
Science, 12: 43–47.
Tybur, J. M., Lieberman, D., and Griskevicius, V. (2009) Microbes, mating, and morality:
individual differences in three functional domains of disgust. Journal of Personality and Social
Psychology, 97: 103–22.

99

Viar-Paxton, M. A., and Olatunji, B. O. (2012) Context effects on habituation to disgust-relevant
stimuli. Behavior Modification, 36(5): 705–722.
Wilcove, D. S., and Master, L. L. (2005) How many endangered species are there in the United
States? Frontiers in Ecology and the Environment, 3: 414–420.

100

APPENDIX

101

INTRODUCTION
Public attitudes toward insects and other arthropods are overwhelmingly negative, despite
the fact that both ecologically and economically arthropods are among the most important
creatures on the planet. The exact origin of negative attitudes toward insects remains somewhat
mysterious. In particular, childhood seems to be an important time for the formation of lasting
beliefs and values pertaining to wildlife (Deruiter & Donnelly, 2002). For example, one study
suggested that the most effective time for building empathy and appreciation toward wildlife is
between the first and fifth grade (ages 6-10 years) (Kellert, 1985). In addition, Shepardson
(2002) surveyed elementary school students from kindergarten through the fifth grade about their
perceptions of insects, and noted that the emergence of negative attitudes toward insects started
around first grade (ages 6-7 years).
The experience of keeping an animal as a pet can enhance an individual’s sense of
empathy (Daly & Morton, 2006; Daly & Suggs, 2010). Children who keep pets are better
informed about their pets’ biology than children who do not keep pets (Prokop et al., 2008).
Keeping pet animals in an educational context has been reported by teachers to be beneficial for
children’s socio-emotional development (Daly & Suggs, 2010). In addition, exposure to animals
which formerly were negatively perceived has been shown to enhance children’s’ appreciation
for those animals (Lindemann-Matthies, 2005).
A previous study by Korte et al. (2005) introduced live Madagascar Hissing Cockroaches
(Gromphadorhina portentosa) as pets in a first grade classroom, and observed the children as
they “got to know” the cockroaches and made observations about their behavior. The children
were interviewed at two time points – just before the cockroaches were first introduced to the
class, and again after several weeks had passed and the children had studied the cockroaches and

102

handled them several times. The children reacted positively to the roaches by the end of the
study.
In order to challenge negative perceptions toward insects, it is important to habituate
individuals as children to the diverse forms, functions, modes of living, and beneficial services
provided by the insect world. Habituation has been shown to be an effective means of developing
positive interest in items that were previously regarded as distasteful (Holstermann et al., 2012).
In the following study, we introduced live pet insects to elementary school classrooms and
collected drawings from students in order to detect changes in the children’s mental models of
insects.
METHODS
Study Design and Data Collection
As part of a larger series of surveys designed to assess elementary school students’
attitudes and knowledge towards insects, students in three 1st/2nd (combined) grade classrooms
and three 3rd/4th (combined) grade classrooms at a local public school completed a drawing task.
The task was administered to students pre- and post-participation in a series of targeted lessons
structured around interactions with live pet insects (hissing cockroaches, blue death-feigning
beetles, and superworms). Thus, each student made two drawings, with the exception of some
who were absent on the days that the surveys were given. Drawings were completed in the
students’ regular classrooms during normal class time. Students were instructed not to talk to
each other as they completed the task, so as not to influence one another’s drawings. The
students were instructed by the first author that the insect they were to draw could be of any kind
that they wished, but should be a “real” type of insect (i.e., not imaginary). In a few
circumstances students asked, “Can I draw a spider?” and were instructed that that would be

103

acceptable despite the fact that spiders are not insects. The reason for this is that children tend to
classify spiders as insects, and children commonly draw spiders when they are asked to draw
insects (Shepardson, 2002). The pre-participation surveys were given between September 30 and
October 6, 2015, before the start of the insect lessons. The post-participation surveys were given
between December 9-11, 2015, after the lessons had finished. Topics covered in the insect
lessons included external insect morphology, sociality, defense mechanisms, and life cycles.

Participants
One hundred and sixty three students from six classrooms (three 1st/2nd grade combined
and three 3rd/4th grade combined) at public elementary school in Michigan, U.S., participated in
this study during the autumn and early winter of 2014. Participants ranged from 6 to 10 years of
age. We report on the analysis of 140 pre- and 161 post-lesson drawings. Sample size differed
between the two treatments mainly due to differential student attendance during sampling days.
In addition, an illustration that decorated the front of the pre-participation survey may have
influenced some students in their drawings. In an attempt to eliminate this potential confound,
we removed from our analysis all pre-participation drawings (n=21) that bore strong similarity to
the cover illustration and removed the cover illustration for the post-survey.

Table A.1. Age and gender of participants.

Pre
Post

n
140
161

Mean Age (SD)
7.69 (1.18)
7.71 (1.2)

104

% Male
43.6
47.2

Coding of drawings
Several methods for interpreting drawings quantitatively have been utilized in previous
analyses of children’s drawings, such as indexing of salient features (Libarkin et al., 2015),
computer-based pattern analyses (Forbus et al., 2011) and grouping drawings into pre-defined
categories (Dove et al. 1999; Shepardson, 2002). In our study, each child’s drawing was
analyzed by coding the individual features of each insect. An iterative process was used to
develop the coding scheme. All three raters began by examining a subset of ten randomly
selected drawings together and listing prominent features from each drawing. Each feature was
given a definition that was mutually agreed upon by all three authors. Once the coding scheme
was finalized, each author independently coded the same 20 drawings, after which an intraclass
correlation was conducted.

Data analysis
Data were analyzed using SPSS 21.0 (IBM, Inc., Chicago, IL, U.S.A.). We calculated the
frequencies of all coded variables from both of our two sampling periods. In order to determine
whether the frequency of coded features changed pre to post, we compared the pre and post
frequencies of each feature in paired-samples t-tests. In addition, we conducted a correlation
analysis between all coded variables to determine whether the data were likely to contain
underlying models. After the method of Libarkin et al. (2015), who utilized principal
components analysis (PCA) to identify underlying models in undergraduates’ drawings of the
greenhouse effect, we wanted to determine whether children’s representations of insects
contained discernible patterns that could be elucidated via PCA. We then utilized point-biserial
correlations in order to determine the relationship between alignment of drawings with models

105

(the factor score) and the use of descriptive codes (e.g. two eyes, six legs, etc.), as well as the
relationship between factor score and the demographic variables gender and grade level.

RESULTS
Coding of drawings
The average measures intraclass correlation across the three raters was 0.92 (min. = 0.91
and max. = 0.93). An intraclass correlation close to 1.0 suggests that each coder’s analysis of the
drawings is consistent with the others’. Given this high level of agreement, the remaining
drawings were coded by one rater. Coding of the children’s drawings resulted in the
identification of 22 distinct insect features that existed in the drawings collectively (Table A.2).
Several of these features, or codes, were automatically related to each other. For example, we
coded for the presence or absence of some of the most common and obvious insect features such
as eyes, mouth, etc., and additionally, we coded for descriptors of those features (should they be
present) such as number of eyes, shape of the mouth, decoration on the wings, etc. (see Table
A.2). Thus, the presence of some of the major codes (eyes, mouth, etc.) automatically
guaranteed the use of an additional descriptive code. Because these major features are related to
their descriptors, but coded separately, we utilized only the major codes such as “mouth” and
“eyes” in our factor analysis.
The drawings in our sample were created by children representative of a range of age
levels (from 6 to 10 years of age). Not surprisingly, we noted quite a bit of variation within the
drawings across the sample. However, there were some consistently obvious patterns that are
reflected in our coding scheme. For example, all of the children drew some sort of “body” for
their insect. The precise shapes of the bodies differed to an extent, but all could be classified as

106

either more rounded or elongate in shape. To reflect this pattern, we coded body shape as one
dichotomous variable with the two mutually-exclusive conditions round and long. In the results
section, a positive factor loading for the variable “body shape” indicates a round body as part of
the model, and a negative loading indicates a long body. If the variable does not load, that
indicates that body shape was variable and is not essential to the model.

Figure A.1. Examples of coded features.

Frequency of coded features
The most commonly drawn insect features, both pre- and post-lessons, were legs, eyes,
head, and antennae, respectively. Table A.2 gives the frequencies, as well as the results of
significance testing between the pre- and post-participation occurrence of each coded feature.

107

Only a few features significantly differed in their frequency pre v. post. Of the major
codes, only head (p=0.005) and mouth (p=0.037) showed significant increases. Of the other
coded features, the frequency-of-use of antennae (p=0.005), stinger (p=0.004), and
detail/decoration on the body (p=0.013) increased significantly. The frequency of drawings with
eight legs decreased significantly (p=0.023). There was also a marginally-significant decrease in
insects with long, pointed mouthparts (p=0.059).

Table A.2. Frequencies of the presence of each variable and also significance of differences
pre/post using paired-samples t-tests.
Characteristics
Before
After
Sig.
Instruction
Instruction
(n=140)
(n=160)
% Present
% Present
Body Shape
Oval
Long
Head
Eyes
Mouth
Legs
Wings
Antennae
Cerci
Stinger
Hair

49.3
50.7
66.4
69.3
42.9
71.4
37.1
52.1
2.9
5
5

41
59
75.2
75.2
45.3
66.5
46.6
58.4
1.2
13
6.2

0.79
0.005
0.09
0.037
0.502
0.205
0.005
0.158
0.004
0.809

Correlational analysis
We investigated correlations between all coded characteristics prior to factor analysis.
Sample sizes consisted of 140 pre-participation drawings and 160 post-participation drawings.
Pre- and post-participation data were tested in two separate correlations. We noted that many of
the most commonly drawn coded features correlated with other features with coefficients above

108

0.3 (Tables A.3), indicating the potential presence of underlying latent factors in the data (Field,
2013). In light of this result, we continued with our investigation of whether the pre- and postparticipation data met the qualifications for factorability.
Also of note is the strong negative correlation, both pre- and post-, between legs and
wings. We can infer from this that children who drew legs on their insect were unlikely to also
include conspicuous wings, and vice versa. This relationship is also reflected in the following
analyses, and is expanded upon in the discussion section.

Table A.3. Pearson correlations between major coded variables before and after the
instructional period. *p < 0.01; **p < 0.05.
PRE
Head
Shape
Legs
Wings
Eyes
Mouth
Head
1
.187*
.281**
Shape
1
.370**
-.344**
.192*
Legs
.187*
.370**
1
-.528**
.333**
Wings
-.344**
-.528**
1
Eyes
.281**
.192*
.333**
1
.389**
Mouth
.
.389**
1
POST
Head
Shape
Legs
Wings
Eyes
Mouth

Head
1
.187*
.261**
-.183*
.301**
.264**

Shape
.187*
1
.271**
-.272**

Legs
.261**
.271**
1
-.655**

Wings
-.183*
.272**
-.655**
1

Eyes
.301**

Mouth
.264**

1
.408**

.408**
1

PCA of pre-participation drawings
To analyze the pre-data, we ran a PCA with no rotation and latent factors extracted at
eigenvalues greater than one, which is Kaiser’s criterion (Field, 2013). In addition, examination
of the point of inflection of the scree plot was also utilized as a criterion for determining the
number of latent factors (Catell, 1960; Costello & Osborne, 2005). We employed the major

109

codes of head, mouth, eyes, legs, wings, and body shape as (oval vs. long) in our statistical
model, as these characteristics were the most frequently drawn major features of insects across
the sample. We report a Kaiser-Meyer-Olkin measure of sampling adequacy of 0.645, which is
above the recommended threshold of 0.5, and in addition Bartlett’s test of sphericity was
significant (χ2 (15) = 124.829, p < 0.001) (Field, 2013). Furthermore, the anti-image correlation
matrix revealed that correlation coefficients on the diagonal were all above 0.5, and
communalities for the five variables utilized in our model were all above 0.3, further indicating
the validity of factor analysis as an analytical approach to these data (Field, 2013).
Two factors emerged from the analysis with eigenvalues greater than one, with the first
factor explaining 35.8% of the variance, and the second factor explaining 21.9%, for a total of
57.7% of variation in the data explained by two latent factors. In addition, examination of the
point of inflection of the scree plot also indicated the presence of two factors, hereafter referred
to as models. Un-rotated loadings for coded features on each model are given in Table A1.4. A
hypothetical illustration of each model is included in Figure A1.2, along with examples of
drawings that appear to fit each model.

110

Pre-Participation

Post-Participation

Crawling Model
WINGS

Flying Model

Crawling Model
Head
Eyes
Mouth
Legs
Oval Body
No Wings
Head
Eyes
Mouth
Wings
Long Body

UNCHANGED

Flying Model
LEGS

Head
Eyes
Mouth
Wings
No Legs

Figure A.2. Schematic illustration of both emergent insect models from the PCA of the pre- and
post-instruction drawings. The crawling model did not change pre- and post-participation. The
flying model changed in that long body was dropped and no legs was adopted.
The analysis suggests the presence of two separate constructs underlying the drawings.
To determine the internal consistency of the two models, we calculated Cronbach’s α for each
model. Negatively-loading items were reverse-coded before conducting the analysis (Field,
2013). We report that Cronbach’s α = 0.621 for the Crawling Model (Oval Body, Legs, Eyes,
Mouth, Head, No Wings), and α = 0.238 for the Flying Model (Long Body, Eyes, Head, Wings,
Mouth). Note that these two models are similar to the models of insects from Chapters three and
four. In alignment with previous literature (Field, 2013), we interpret these alpha values to
suggest that the Crawling Model is a reliable construct. However, the low alpha achieved by the
Flying Model suggests it is not truly representative of underlying constructs.

111

Table A.4. Factor loadings for drawings pre-participation. Loadings below ±0.3 are suppressed.
Coded Feature
Body Shape (+ Oval, − Long)
Head
Eyes
Mouth
Legs
Wings

Crawling Model Flying Model
0.624
-0.312
0.460
0.357
0.624
0.522
0.371
0.682
0.783
-0.636
0.613

PCA of post-participation drawings
For the post drawings, we also conducted a separate PCA with no rotation and factors
extracted at eigenvalues greater than one. We utilized the same coded features as in the pre-data
for our analysis, in order to determine whether the models changed or remained the same pre- to
post-participation in the insect lessons. The Kaiser-Meyer-Olkin measure of sampling adequacy
for the post-data was 0.62, which again is above the recommended threshold of 0.5, and
Bartlett’s test of sphericity was significant (χ2(15) = 166.0, p < 0.001) (Field, 2013).
Communalities were above 0.3 for all variables, and diagonals in the anti-image correlation
matrix were all above 0.5.
Similar to the pre-participation results, our analysis of post-participation drawings
revealed two latent factors with eigenvalues greater than one. The first factor explains 34.64% of
the variation, and the second factor explains 24.63%. Together these two latent factors account
for 59.27% of the variance in the data. In addition, the scree plot was also indicative of the
presence of two factors. Loadings for the post-data are given in Table A.5. The Crawling Model
remains the same. Post-participation, the Flying Model changes in that body shape falls out of
the model and not-legs is added. Please see Figure A.2 for visual depictions of these changes.

112

Analysis of the internal consistency of each model indicates that, post-participation, Cronbach’s
α = 0.605 (acceptable) for the Crawling Model, and α = 0.361 (low) for the Flying Model.

Table A.5. Factor loadings for post-participation drawings. Loadings below ±0.3 are
suppressed.
Coded Feature
Crawling Model
Flying Model
Body Shape (+ Oval, − Long)
0.518
Head
0.611
0.331
Eyes
0.432
0.675
Mouth
0.373
0.696
Legs
0.777
-0.370
Wings
-0.711
0.492
DISCUSSION
The two insect models that emerged from the analysis are particularly interesting because
they represent the two locomotory modalities by which children experience insects – crawling on
the ground or other surface, and flying through the air. Another interesting finding is the
mutually exclusive nature of legs and wings in the drawings, as indicated by our correlational
and factor analyses. In short, factor loadings indicated that drawings of flying insects generally
showed the wings but not the legs, and drawings insects with legs generally did not have wings.
In nature, the vast majority of insects have both legs and wings, and thus this tendency to draw
one but not the other is incongruent with reality and may be an especially important area to target
during instruction about insects.
Breuer et al. (2015) documented a similar result in their investigation of children’s
conceptions of insects. Breuer et al. (2015) analyzed children’s comments in response to
exposure to images of different types of insects. Using a cluster analysis, Breuer et al. (2015)
determined that the children’s comments about the different insects that had or appeared to have

113

wings grouped separately from comments about the insects (and other invertebrates) that did not
have wings. It appears, then, that locomotory modality may be particularly important to
children’s conceptions of insects.
Though we did not explicitly investigate the relationships between children’s attitudes
toward insects and their drawings of them, it is very likely that a relationship exists (Jolley,
2010). The exact origins of negative attitudes toward insects are murky and have not been fully
explicated. It has been postulated that humans are evolutionarily adapted to avoid insects
because some species pose dangerous threats to humans (Ohman et al., 1985). Cultural
environment, parental attitudes toward insects, and personal experiences of insects also play
important roles that influence the strength and valence of attitudes toward insects (Breuer et al.,
2015; Prokop et al., 2010).
It is quite striking how similar these models and relationships between features are to
models that emerged from an analysis of adults’ drawings of insects (Chapter 3). This suggests
that flying and crawling – legs and wings – are perceived as distinct from each other from
childhood through the lifespan.

114

REFERENCES

115

REFERENCES
Barraza, L. (1999) Children’s drawings about the environment. Environmental Education
Research 5: 49-66.
Bjerke, T., and Østdahl, T. (2004) Animal-related attitudes and activities in an urban population.
Anthrozoos, 17: 109–129.
Breuer, G. B., Schlegel, J., Kauf, P., and Rupf, R. (2015) The importance of being colorful and
able to fly: interpretation and implications of children’s statements on selected insects and other
invertebrates. International Journal of Science Education, 37(16): 2664–2687.
Cardoso, P., Erwin, T., Borges, P. and New, T. (2011) The seven impediments in invertebrate
conservation and how to overcome them. Biological Conservation, 144(11): 2647-2655.
Cattell, R. B. (1966) The scree test for the number of factors. Multivariate Behavioral Research
1: 629-637.
Daly, B., and Morton, L. L. (2006) An investigation of human-animal interactions and empathy
as related to pet preference, ownership, attachment, and attitudes in children. Anthrozoös, 19(2):
113-127.
Daly, B., and Suggs, S. (2010) Teachers’ experiences with humane education and animals in the
elementary classroom: implications for empathy development. Journal of Moral Education, 39:
101-112.
De Winter, J. C. F., and Dodou, D. (2016) Common factor analysis versus principal component
analysis: A comparison of loadings by means of simulations. Communications in Statistics –
Simulation and Computation, 45: 299-321.
Dove, J. E., Everett, A., and Preece, P. F. W. (1999) Exploring a hydrological concept through
children’s drawings. International Journal of Science Education, 21(5): 485-497.
Field, A. (2013) Discovering statistics using IBM SPSS statistics. Sage.
Forbus, K., Usher, J., Lovett, A., Lockwood, K., & Wetzel, J. (2011). CogSketch: Sketch
understanding for cognitive science research and for education. Topics in Cognitive
Science, 3(4), 648-666.
Jolley, R. P. (2010) Children & Pictures: Drawing and understanding. Wiley-Blackwell.
Kidd, A.H., and Kidd, R. M. (1990) Social and environmental influences on children’s attitudes
toward pets. Psychological Reports, 67: 807-818.

116

Kellert, S.R. (1993) Values and perceptions of invertebrates. Biological Conservation, 7(4): 845855.
Losey, J.E., and Vaughan, M. (2006) The economic value of ecological services provided by
insects. BioScience, 56(4): 311-323.
Martín-López, B., Montes, C., and Benayas, J. (2007) The non-economic motives behind the
willingness to pay for biodiversity conservation. Biological Conservation, 139: 67-82.
McLean, M., Henson, Q., and Hiles, L. (2003) The possible contribution of student drawings to
evaluation in a new problem-based learning medical programme: a pilot study. Medical
Education, 37(10): 895-906.
Ohman, A., Dimberg, U., and Ost, L.-G. (1985) Animal and social phobias: Biological
constraints on learned fear responses. In S. Reiss & R. R. Bootzin (Eds.), Theoretical issues in
behavior therapy (pp. 123-178). New York: Academic Press.
Prokop, P., A. Tolarovičová, A., Camerik, A. M., and Peterková, V. (2010) High school
students’ attitudes towards spiders: A cross‐cultural comparison. International Journal of
Science Education, 32(12): 1665-1688.
Sallam, M. Eds. Mejia, D., and Lewis, B. 1998. Insect damage: damage on post-harvest.
International Centre of Insect Physiology and Ecology. Food and Agriculture Organization of the
United Nations.
Shepardson, D. (2002) Bugs, butterflies, and spiders: children's understandings about insects.
International Journal of Science Education, 24(6): 627-643.
Snaddon, J. and Turner, E. (2007) A child's eye view of the insect world: perceptions of insect
diversity. Environmental Conservation, 34: 33-35.

117