Behavior problems are a leading reason for euthanasia or surrender of cats, dogs, and horses.1–10 In a 2009 survey10 of > 2,700 horse owners, 60% reported that they had euthanized an unwanted horse, and behavior was listed among the top 5 reasons that a horse was unwanted. In the United States it is estimated that 670,000 dogs and 860,000 cats are euthanized annually in shelters.11 A 1999 survey estimated that an additional 224,000 dogs and cats were euthanized annually in American veterinary practices because of behavior problems.4 Although in-practice euthanasia numbers have not recently been reassessed, behavior problems may be as prevalent as 85% among dogs.12 Given the gravity of these numbers, all avenues of intervention should be evaluated.
Most cats and dogs have visited a veterinarian ≤ 1 year prior to relinquishment to a shelter for a behavior problem, placing veterinarians on the front lines of prevention7 and making knowledge in animal behavior a necessary component of day 1 competency. Day 1 competency is defined by the AVMA as adequate clinical experience to diagnose, treat, and prevent mental or physical disease.13 Despite this, surveyed 2016 graduates of AVMA-accredited veterinary schools did not uniformly report being prepared for practice on day 1 with respect to behavior.14 In a 2014 survey-based study,15 only 22 of 30 (73%) participating AVMA Council on Education–accredited schools in the United States offered formalized behavior education of any duration (range, < 4 days to 1 full semester), and the offerings in 19 AVMA Council on Education–accredited schools outside of the United States were unclear. Only 97 of 362 (26.8%) graduating seniors surveyed in 2016 reported that their behavior curriculum prepared them for diagnosis and treatment of behavior problems after graduation,14 indicating that some curricula may lack sufficient depth or rigor. These results suggest that veterinary students graduate with a wide range of educational experience and knowledge regarding evidenced-based behavior management and commonly held behavior misconceptions; some graduates may even be completing veterinary school with the same level of behavior knowledge that was present at the start of their curriculum. To the authors' knowledge, information regarding behavior knowledge among veterinary students prior to receiving postgraduate training in animal behavior has yet to be published.
To diagnose, treat, and prevent behavior problems, graduates need a strong knowledge base regarding animal body language and motivation as well as basic concepts of learning theory. Accurately interpreting an animal's body language can help a clinician understand an animal's motivation and create an appropriate treatment plan. Familiarity with and interpretation of the body language of companion animals also necessarily informs day-to-day safety and owner perception of care in the veterinary workplace.16,17 Treatment plans for behavior problems center on combinations of behavior modification, avoidance of triggers for unwanted behavior, and, when warranted, medical treatment.18–20 Selection of appropriate medications for treatment of behavior problems is a highly involved process and is often beyond the scope of introductory courses because it requires integration of many separately taught clinical topics. However, implementation of appropriate management plans is within the scope of an introductory course. It requires understanding how behavior changes in accordance with the way animals learn. Of critical importance is the recognition of the differences and interactions between associative-emotional learning and operant-rational learning, especially outside of highly controlled laboratory settings, such as in clinical practice and daily life.20–23
As veterinary professionals, graduates should have a scientifically sound knowledge base that is consistent with current peer-reviewed articles and textbooks on animal behavior. That said, students likely acquire much of their information and perceptions about animal behavior from nonscientific sources, such as television, lay books, and magazines and other popular cultural publications including websites. These sources likely include misleading information and information about animal behavior that is not substantiated by the scientific literature (ie, behavior myths). The perpetuation of inaccurate or unsubstantiated information undermines graduates' ability to provide care in this critically important arena and can contribute to insufficient or possibly inappropriate treatment recommendations. Such recommendations can risk harm to patient or client, damage to the human-animal bond, and violations of the veterinarian's oath.24–26
A common myth is that animals have the same motivations for their behaviors as people and that these are displayed with the same physical signals or body language. The ascription of anthropomorphic motivations to animals is a long-standing and well-debated topic for which there is a lack of consensus on terminology.27,28 Although there is often good agreement in terminology used by lay people,29 there are inherent problems with the use of descriptions that apply to human emotional states such as spite, jealousy, deception, and guilt when referring to companion animals. Because different species use postures and gestures differently, there is a large potential for inaccuracy when interpreting these signs.30–33 For example, experimental investigation of the so-called guilty look in dogs revealed that the display of postures that signal apologetic or guilty feelings in human body language (eg, lowering the head, looking up, widening the eyes, and hunching at the shoulders) was instead a fearful response to owner cues; the use of these signals was dependent on the owner's actions and was independent of the dog's behavior.31
Dominance as a motivational state or diagnosis for behavior problems, particularly among dogs and horses, is also a topic of contention.30,34–42 When initially described for dogs in the 1960s, dominance referred to resource-holding ability during bone-in-pen tests of domestic dogs.43 These tests were devised on the basis of behavior observations of unrelated, wild-caught wolves placed together in captivity.44 The use of this term has been inappropriately expanded in popular culture to include a variety of intraspecies and interspecies aggressions, relationships, unruly behaviors, and failures to comply with or meet human expectations.34 Treatment plans based on a dominance framework may lead to aversive correction or punishment-based training that may seem valid at face value but can worsen aggression toward the original target or toward the technique implementer, most often the owner.45 Associative learning, first described by Watson and Reyner in the 1920s, likely underlies the potential for increased aggression when punishment-based methods are used.21 Fear and anxiety are intervening variables that change learning and behavior output, particularly in response to positive punishment (ie, adding something aversive to decrease a behavior).45,46 Failure of veterinarians to understand how animals learn and the unintended effects of punishment on animals can lead to potentially detrimental recommendations for addressing behavior problems or the inability to intervene regarding the use of such techniques. Aversive training techniques and dominance frameworks can fracture the human-animal bond by furthering misconceptions about the relationship47 and can increase the risk of relinquishment, consideration of rehoming, or euthanasia instead of seeking support for behavior-related problems.48 Veterinarians should also be aware of misconceptions about animal learning, most notably the practice of intensive early handling of neonatal foals in an attempt to alter their behavior later in life, despite evidence that it has limited long-term effectiveness.46,49–55
The purpose of the study reported here was to assess incoming first-year veterinary students' knowledge of basic learning theory, body language, motivational states, and dominance in horses, cats, and dogs, particularly with respect to common behavior myths; to investigate factors potentially associated with this knowledge (including sources of information and companion animal ownership); and to determine the effects of a semester-long, 2-credit animal behavior course on student knowledge of these subjects. We hypothesized that students who reported previous animal behavior courses and peer-reviewed journal articles as the sources of their knowledge about animal behavior would score higher on a basic companion animal behavior knowledge survey than those who reported other sources such as dog trainers, veterinarians, and popular-culture (ie, pop-culture) sources such as television and magazine articles including those found online; that students reporting pop-culture information sources would more frequently choose answers consistent with common behavior myths over evidenced-based answers on the survey; and that students would have higher scores on the same animal behavior knowledge survey after participating in the introductory animal behavior course. We also hypothesized that higher survey scores would predict higher scores on the final examination for the behavior course.
Materials and Methods
First-year veterinary students at The Ohio State University College of Veterinary Medicine were invited to participate in the study before and at the completion of a 1-semester (2-credit) Introduction to Animal Behavior course. The course consisted of 24 lecture hours and 8 laboratory contact hours. Eighty-two of the 162 (51%) eligible students were from Ohio; the remainder were from 25 other states and 1 country outside of the United States.a The study was completed in accordance with the facility's institutional review board educational exemption (No. 2017E0533).
A survey was created that included demographic and knowledge-base sections. The survey was adapted from a psychometrically tested survey on animal behavior education and myths among shelter volunteers (deemed a semi-knowledgeable population).b Content was shifted away from animal shelter vignettes; questions related to equine behavior were added, and the entire survey was reevaluated by teachers of the animal behavior course. A paper version of the adapted survey was pilot tested among a dichotomous sample of individuals at The Ohio State University College of Veterinary Medicine (n = 8 respondents) who were identified by a DACVB as having either a high or low to moderate degree of behavior knowledge. The questions were refined for clarity according to comments from individuals who participated in the pilot study, and an online versionc was created (Supplementary Appendix S1, available at: avmajournals.avma.org/doi/suppl/10.2460/javma.256.10.1153).
Students were emailed links to the survey before the first lecture of their animal behavior course, and the survey opened after a brief introduction to the course. Students were allotted 15 minutes of in-class time to complete the survey, and extra time was offered if needed. On the last day of the course, the same survey was provided to students with the same amount of in-class time allotted for its completion. Participation in the preclass and postclass surveys was voluntary and offered to all students. Responses were collected and filed anonymously and confidentially online.
In the demographic section, students were asked whether they owned a dog, cat, or horse and whether they had taken an undergraduate animal behavior course. They were also asked to select sources of their companion animal behavior knowledge: previous animal behavior course, peer-reviewed journal articles, books, magazines or online pop-culture articles (treated as 1 combined selection), television shows, dog trainers, a family veterinarian, or personal experience. Multiple responses were allowed. The demographic portion of the survey further included a self-assessment section in which students rated their own skill or comfort level on a Likert-type scale from 1 (indicating no knowledge or not being comfortable) to 7 (indicating expert knowledge or being extremely comfortable). Three subjects were listed for self-assessment: methods to reduce undesired behavior (normal or abnormal) in companion animals, understanding learning theory for training or how to train, and understanding motivation of behavior in companion animals. The students ranked their skill or comfort level for understanding motivation for each species of interest (dog, cat, and horse) separately.
The knowledge-base section of the survey focused on dog, cat, and horse body language; motivations for behavior; and basic learning theory definitions and applications. It consisted of 23 items with true-or-false, multiple-choice, and SATP response options. These prompts or questions and the correct responses were developed from peer-reviewed and textbook publications by DACVB authors. Correct responses for items related to body language and motivation were supported by information in publications related to canine and feline19,20,30,56–58 or equine57,59–61 body language. Common behavior myths and misconceptions comprised distractors (incorrect response options) included in the survey. These included anthropomorphisms (eg, attributing behavior to spite, boredom, jealousy, or meanness), use of punishment or exposure to the fear-inducing stimulus to ameliorate fearful behavior, and generalizability of emotional states and cognitive function between vastly different experiences. Consequently, each wrong answer represented a myth. To assess misconceptions about dominance, the term was used only as a distractor; dominance was specifically not presented in terms of current, evidence-based definitions such as a fluid attribute of repeated agonistic encounters between members of the same species for resource holding, characterized by persistent yielding of 1 animal.19,20,43,60,62–64 It was included only as a motivation or inherent trait. This also prevented knowledge or misconceptions about dominance from influencing questions when dominance was in neither the prompt nor the distractors. Learning theory questions assessed basic understanding of operant conditioning and initial reward rates but did not delve into more complex discussions such as duration of behavior extinguishment or persistence reward schedules.21,46
Before analysis, data were summarized as total scores (a sum of all correct responses in the knowledge base section), species-specific subscores (correct responses related to dog, cat, or horse behavior), and topic-specific subscores (correct responses related to companion animal body language and motivation, learning or training, and dominance). Items with binary (true or false) and multiple-choice response options each had 1 correct answer worth 1 point. For items with SATP instructions, 1 point was assigned for each correct response selected, and 1 point was assigned for each distractor not selected (as if each was an independent question with a true-or-false response option). Because of overlap in the subject matter, some questions were included in > 1 subscore. The highest possible subscores for knowledge of dog, cat, and horse behavior were 26, 9, and 9 points, respectively. The highest possible subscore for body language and motivation was 16 points, and that for learning or training was 6 points. To evaluate knowledge about dominance, items in which the prompt or question included dominance were scored and a count of incorrect responses was generated. A count of instances when dominance distractors were selected as responses was also generated. These counts were summed to create a dominance index (maximum of 18 points). Unlike the other subscores, a higher dominance index suggested less knowledge about dominance.
The animal behavior course included lecture material focused on information from peer-reviewed literature, hands-on animal-handling laboratories (with cats, dogs, cattle, and horses), and in-class cases focused on clinically relevant behavior problems in domesticated animals. This included case assignments in laboratory format as well as case presentations in didactic form. The course was not specifically designed to debunk common myths about animal behavior but was intended to supply students with a solid knowledge base about behavior that could be applied in later courses and in clinics, including the ability to identify behaviors that indicate an animal's intent or motivation.a Assessments included a midterm examination, a final examination, and 5 laboratory assignments. The laboratory assignments for the course were specifically designed for low discriminatory value (ie, most students were expected to score 100%). Thus, to determine whether the survey scores reflected knowledge gained by students in the course, the total postclass survey score was evaluated as a predictor of the final examination score, rather than students' course grade. The final examination was written by 2 investigators who were course instructors (MEH and KLP) prior to survey development and was not altered for purposes of the study. Course instructors were masked to student identity and student survey scores until final grades were posted. At that time, 1 instructor was unmasked to survey identification numbers, but not scores or responses. This allowed linking the final examination score to the survey score through an identification number key without breaching anonymity.
Owing to the use of SATP questions and a technical glitch regarding the selection of the responses for these questions in the preclass survey, percentage scores (percentage of the number of possible points earned) were used rather than raw scores for comparisons of survey data. In the preclass survey, 1 SATP question about motivation for inappropriate urination of cats was inadvertently not set to accept > 1 response on touchscreens (eg, mobile phones or tablet computers) until partially through the survey period. Thus, this item was scored such that failure to select the anthropomorphic distractor was considered a correct response. On the postclass survey, this error was corrected and the item was scored as a series of independent binary response (true or false) questions with a maximum score of 3 points (ie, 2 points for selecting the 2 correct responses and 1 point for not selecting the distractor). Similarly, an SATP question regarding motivation for growling with anthropomorphic and dominance distractors was corrected partway through the preclass survey period. For the preclass survey, the most frequently selected distractor (dominance) was the only response option scored, and this item was treated as a true-or-false question. For the postclass survey, all responses were scored as described for other SATP questions. Consequently, the possible number of points for knowledge of cat and dog behavior was higher for the postclass survey than for the preclass survey; because of the SATP nature of the questions, the potential to select myth-based responses was lower in the preclass survey. Consequently, percentage scores were used to evaluate the differences in pre- and postclass behavior knowledge.
Statistical analysis
Responses from the pre- and postclass surveys were scored and tabulated for statistical analysis. For preclass knowledge assessments, 4 separate multivariable linear models were built.65 Variables from the demographic and self-assessment section of the survey, including age, gender, participation in an animal behavior course during undergraduate education, self-assessments of general behavior knowledge, and sources of knowledge about companion animal behavior, were evaluated as predictors of the preclass total score (model 1). Ownership of specific species (eg, dogs) and species-specific self-assessment items (eg, understanding dog behavior and motivation of dogs) were evaluated as predictors for each species-specific (dog [model 2], cat [model 3], and horse [model 4]) behavior knowledge subscore. Descriptive statistics were generated for all variables of interest to screen for missing data or entry errors and to ensure sufficient variability of the predictors for the modeling process. Numeric variables (age and items scored with a Likert-type scale) were checked for the assumption of linearity graphically and by introducing its quadratic term into a univariable model for situations in which the relationship appeared quadratic; variables without a linear or quadratic relationship with the outcome of interest were dichotomized around the mean for continuous variables and around the median for discrete variables. To screen potential predictor variables for multivariable model inclusion, univariable linear regression models were built. All potential predictors with values of P < 0.20 were included in initial multivariable regression models.65,66 Collinearity was evaluated by means of Spearman correlation, with an rs ≥ 0.60 considered to indicate collinearity.65 Final multivariable models were built with a backward stepwise approach, and values of P < 0.05 were considered significant. Variables were defined as confounders if removal from the model changed any coefficients by ≥ 20%.65 Finally, the model assumption of homoscedasticity was assessed with the Cook-Weisberg test.65
The distributions of gender and age of individuals who completed the postclass survey were compared with those for individuals who completed the preclass survey by use of κ2 and Kruskal-Wallis tests, respectively, owing to the lack of fully paired data (students who took the preclass survey but not the postclass survey as well as the reverse).65,67 For assessment of postclass knowledge and changes in knowledge, preclass and postclass distributions for total scores, species-specific knowledge (dog, cat, and horse) subscores, and topic-specific knowledge (learning or training, body language and motivation, and dominance index) subscores for the survey were assessed. Differences between the preclass and postclass scores for each data set were calculated. When the distributions for score differences were normal, paired t tests were performed (total, dog behavior, and body language and motivation scores); Wilcoxon signed rank tests were performed68 for analyses of all other data. Finally, a univariable linear regression model was created to evaluate the postclass survey total knowledge score as a predictor of the final examination score for the behavior course.
Results
Of 162 veterinary students enrolled in the Introduction to Animal Behavior course, 156 (96.3%) logged into the survey prior to the course. Two students logged into the survey (one on the preclass survey and another on the postclass survey) but provided no responses to any questions. Some students provided only demographic and self-assessment information (n = 3 for the preclass survey and 5 for the postclass survey). Of 152 students who responded to all sections of the preclass survey, 101 (66.4%) completed the postclass survey, for an overall (2-survey) response rate of 62.3%; 2 students who did not complete the preclass survey completed the postclass survey. Students who did not complete the postclass survey were nearly equally distributed between males (24/53 [45.3%]) and females (29/53 [54.7%]), indicating no bias in attrition. Postclass survey gender and age distributions (82/101 [81.2%] females and 19/101 [18.8%] males; median age of postclass participants, 22 years [range, 20 to 43 years]) did not differ significantly from preclass survey participants.
Preclass surveys
Many of the students at the institution where the study was performed had master's degrees or doctorates in fields tangentially related or unrelated to veterinary medicine (eg, economics or chemical engineering).a Although we included education level in the demographics from the original survey,b it was excluded a priori from the model for this reason. Responses to 2 of the self-assessment items rated with a Likert-type scale, methods to reduce undesired behavior in companion animals and understanding basic learning theory for training or how to train, met criteria for collinearity (rs = 0.638; P < 0.001). As these 2 items represented related principles46 and single-item analysis is prone to inaccuracy, results for the 2 items were summed69 to create a single item that represented understanding how to change behavior, with scores ranging from 2 to 14. Most students (104/155 [67.1%] for all responses; 103/152 [67.8%] for completed surveys) had a skill or comfort level rating on this combined scale between 6 and 10.
Predictors of total score—Five variables in the demographic and self-assessment section of the survey met cutoff criteria for prediction of the total score of the knowledge testing section of the preclass survey for univariable analyses. They were initially offered as fixed effects for multivariable model 1 (Table 1). The 5 variables included the combined self-assessment item of understanding how to change behavior and 4 sources of knowledge about companion animal behavior (previous undergraduate course, peer-reviewed journal articles, magazines or online pop-culture articles, and television shows). Having taken an undergraduate animal behavior course was suspected to be a confounding variable for selecting a previous animal behavior course and peer-reviewed journals as sources of knowledge in companion animal behavior and for the combined self-assessment item. However, its removal from the multivariable model resulted in a < 20% change in the coefficients of these variables, and it was therefore removed from the final model.
Results of univariable analyses of demographic data and self-assessed general animal behavior knowledge or comfort level scores for 156 first-year veterinary students (statistical model 1) to identify potential predictors of total score on the companion animal behavior knowledge (dog, cat, and horse body language; motivations for behavior; and basic learning theory definitions and applications) testing portion of a survey prior to participating in an introductory animal behavior course (ie, preclass survey).
| Variable | No. (% of total) or mean ± SD | Coefficient* (SE) | P value |
|---|---|---|---|
| Gender (n = 154) | |||
| Female | 111 (72.1) | Referent | — |
| Male | 43 (27.9) | 0.81 (2.22) | 0.717 |
| Age† (n = 155) | |||
| < 24 y | 115 (74.2) | Referent | — |
| ≥ 24 y | 40 (25.8) | 2.17 (2.32) | 0.352 |
| Undergraduate animal behavior course (n = 155) | |||
| No | 87 (56.1) | Referent | — |
| Yes | 68 (43.8) | 1.15 (1.99) | 0.564 |
| Sources of animal behavior knowledge (n = 155) | |||
| Previous animal behavior course | 57 (36.8) | 3.32 (2.06) | 0.107 |
| Peer-reviewed journal articles | 18 (11.6) | 7.80 (2.99) | 0.010 |
| Books | 34 (21.9) | 0.49 (2.43) | 0.842 |
| Magazines or online pop-culture articles | 14 (9) | −5.47 (3.64) | 0.135 |
| Television shows | 26 (16.8) | −4.04 (2.61) | 0.123 |
| Dog trainers | 48 (31) | 1.19 (2.15) | 0.580 |
| Family veterinarian | 68 (43.9) | 2.05 (1.98) | 0.304 |
| Personal experience | 134 (86.5) | −3.16 (2.91) | 0.279 |
| Understanding how to change behavior‡§ (n = 155) | 7.57 ± 2.41 | 0.85 (0.41) | 0.038 |
Not all students answered every question; multiple responses were allowed for the question of sources of animal behavior knowledge. Variables with values of P < 0.20 in univariable analysis were included in the multivariable model.
The coefficient in univariable analyses indicates potential percentage variation in total score attributable to a change of 1 unit in the variable if all other variability is presumed constant.
Dichotomized on the basis of mean age for all participating students.
Knowledge or comfort level was rated on a Likert-type scale from 1 (indicating no knowledge or not being comfortable) to 7 (indicating expert knowledge or being extremely comfortable) for individual behavior-related subjects
Two self-assessment items that were applicable to all species of interest (dogs, cats, and horses) were included in the survey: methods to reduce undesired behavior (normal or abnormal) in companion animals and basic learning theory for training or how to train. Results for these 2 related questions were combined69 to create a single item, understanding how to change behavior, with possible scores ranging from 2 to 14.
— = Not applicable.
The final multivariable model included 2 sources of companion animal behavior knowledge: peer-reviewed journal articles and magazines or online pop-culture articles (P < 0.001; Table 2). The model met the assumption of homoscedasticity (P = 0.68). Students who reported peer-reviewed journal articles as a source for previous companion animal behavior knowledge scored 9.0% higher on the knowledge testing portion of the preclass survey than students who did not cite this source. Students who reported magazines or online pop-culture articles as a source scored 7.6% lower than students who did not cite this source.
Results of multivariable analysis of demographic data and self-assessed general animal behavior knowledge or comfort level scores (statistical model 1) to identify potential predictors of total score on the companion animal behavior knowledge testing portion of the preclass survey for the students in Table 1.
| Variable | Coefficient* (SE) | 95% CI | P value |
|---|---|---|---|
| Intercept | 48.70 (1.04) | 46.64 to 50.75 | < 0.001 |
| Source of animal behavior knowledge | |||
| Peer-reviewed journal articles | 9.03 (3.02) | 3.07 to 15.00 | 0.038 |
| Magazines or online pop-culture articles | −7.57 (3.62) | −14.72 to −0.428 | 0.003 |
Values of P < 0.05 were considered significant.
The coefficient of the intercept represents the predicted score on preclass survey without influence of the 2 significant variables in the model (peer-reviewed journal articles and magazines or pop-culture articles). Agreement with mean preclass score (48.9) indicates no problems of scale. The coefficients of remaining variables are interpreted as for univariable analyses.
See Table 1 for remainder of key.
Predictors of species-specific subscores—Companion animal ownership (dog, cat, or horse) did not meet criteria for inclusion in models for species-specific behavior knowledge subscores on the preclass survey (ie, in models 2, 3, or 4). The only variable that was a significant predictor in any of these separate models was the self-assessed score for understanding motivation of behavior in dogs, which was a predictor for the dog behavior knowledge subscore (intercept, 36.28 [SE, 4.33]; 95% confidence interval, 0.30 to 3.99; P = 0.023). No other self-assessed species-specific companion animal behavior knowledge or comfort level score was a significant predictor of species-specific knowledge subscores on the survey; thus, there were no multivariable models for species-specific knowledge (Table 3).
Results of univariable analyses of companion animal ownership status and species-specific companion animal behavior knowledge or comfort level scores (statistical models 2 to 4) to identify potential predictors of behavior knowledge subscores for each respective species on the preclass companion animal behavior survey for the students in Table 1.
| Model No. | Variable | No. (% of total) or mean ± SD | Coefficient* (SE) | P value |
|---|---|---|---|---|
| 2 | Dog ownership (n = 154) | |||
| No | 35 (22.7) | Referent | — | |
| Yes | 119 (77.3) | 0.77 (2.89) | 0.791 | |
| Understanding motivation of behavior in dogs (n = 152)‡ | 4.46 ± 1.27 | 2.15 (0.93) | 0.023 | |
| 3 | Cat ownership (n = 150) | |||
| No | 81 (52.3) | Referent | — | |
| Yes | ||||
| Understanding motivation of behavior in cats (n = 149)‡‖ | 74 (47.7) | 1.06 (2.99) | 0.723 | |
| ≤ 3 | 86 (55.8) | Referent | — | |
| ≥ 4 | 68 (44.2) | 4.20 (3.01) | 0.164 | |
| 4 | Horse ownership (n = 155) | |||
| No | 138 (89.0) | Referent | — | |
| Yes | 17 (11.0) | −1.60 (3.66) | 0.662 | |
| Understanding motivation of behavior in horses (n = 152)‡ | 2.63 ± 1.65 | 0.15 (0.72) | 0.833 |
Responses for this self-assessment item were dichotomized around the median (3.5) of the scores for this item because there was no linear or quadratic relationship between the scores for this variable and cat score on the preclass survey. There were no significant multivariable (final) models for preclass species-specific scores.
See Table 1 for remainder of key.
Comparison of preclass and postclass survey scores and associations between postclass survey and final examination scores—Total score for the companion animal behavior knowledge testing section on the preclass survey (mean ± SD, 49 ± 12.7%) was lower than that on the postclass survey (84.3 ± 8%; P < 0.001). Species-specific (dog, cat, or horse) and topic-specific knowledge (learning or training and body language and motivation) subscores were also significantly lower (P < 0.001 for all comparisons) on the preclass survey, compared with postclass survey results (Figure 1). Dominance index, an inverse measure of knowledge on the topic of dominance scored from 0 (most knowledge) to 18 (least knowledge), was significantly (P < 0.001) lower on postclass surveys (mean ± SD, 0.4 ± 1.1) than on preclass surveys (6.9 ± 2.6).
Comparison of mean ± SD companion animal behavior knowledge scores of 152 first-year veterinary students before (preclass survey; gray bars) and after (postclass survey; black bars) participation in an introductory animal behavior course. Total scores and species-specific subscores are shown. Because of overlap in questions related to species-specific animal behavior, percentages for the subscores do not add up to 100. Paired t tests were used for comparison of normally distributed data (total, dog behavior, and body language and motivation scores), and Wilcoxon signed rank tests were used for comparison of nonnormally distributed data (all others). *Values indicated by straddle bars were significantly (P < 0.001) different.
Citation: Journal of the American Veterinary Medical Association 256, 10; 10.2460/javma.256.10.1153
The mean ± SD score for all students on the final examination for the Introduction to Animal Behavior course was 91.3 ± 5.7% (ie, 65.77 ± 4.13 of 72 possible points). The total score for companion animal behavior knowledge on postclass surveys was a significant (P < 0.001) predictor of the final examination score (coefficient, 0.121; 95% confidence interval, 0.051 to 0.192). Homoscedasticity requirements were met for this model (P = 0.14). Problems of scale were ruled out because only participants who did not take the postclass survey scored below the model's intercept of 56 points on the final examination. There was no significant difference in final examination scores for those who took the postclass survey (92.8 ± 4.1% [66.8 ± 2.98 points]), compared with scores for the class overall.
Discussion
The present study investigated incoming first-year veterinary students' knowledge about basic learning theory, body language, motivations for behavior, and dominance in horses, cats, and dogs at 1 college of veterinary medicine, with a specific focus on misleading and unsubstantiated information regarding animal behavior (ie, behavior myths). Potential predictors of incoming knowledge and the effects of an introductory animal behavior course on student knowledge were also investigated. The results revealed that incoming veterinary students had little knowledge regarding animal behavior, and their responses to items listed in the knowledge testing portion of the preclass survey were consistent with common behavior myths rather than information supported by scientific evidence. The mean total scores for behavior knowledge on preclass surveys reflected what was considered a failing grade (ie, < 50%). In the final multivariable analysis model, higher scores were associated with students reporting peer-reviewed journal articles as sources of previous companion animal behavior knowledge, and lower scores were associated with reporting of magazines or online pop-culture articles as sources for this knowledge. After a semester-long (32 contact hours), 2-credit Introduction to Animal Behavior course, students' total behavior knowledge scores as well as behavior knowledge scores for all 3 species of interest (dog, cat, and horse) and all specific topics evaluated (learning or training, dominance, and body language and motivation) were significantly improved.
The finding that reporting peer-reviewed journal articles as a source of previous knowledge predicted higher scores, whereas reporting magazines or online pop-culture articles as a source of previous knowledge predicted lower scores on the preclass survey, was not surprising. These 2 sources of information have previously been described to have widely differing qualities of behavior-related information.24,31 It is also interesting to note that the predictive score increase of peer-reviewed journal articles was somewhat greater than the predictive score decrease for magazines and pop-culture articles. This may have been attributable to wider variation in the quality of information in pop-culture sources or could have been reflective of the fact that as this option was selected more frequently, it accounted for less variation in students' poor scores. Citing culturally driven sources of behavior knowledge (eg, television shows and magazines or online pop-culture articles) was common among surveyed students. Behavior myths are culturally rampant and more easily accessed through modern media than reputable sources. As an example, internetd searches performed in December 2017 and March 2018 with the search terms “why do dogs hump?” and “why do dogs mount?” produced a mix of search results from those with no citations to websites that quoted DACVBs and cited scientific research for the information provided. Unfounded web pages also commonly appeared before evidence-based sources in searches performed with search terms of “dominance in dogs,” “dominance in cats,” and “dominance in horses.”
Other sources cited for previous animal behavior knowledge by first-year veterinary students (personal experience, books, dog trainers, and family veterinarian) were not predictive of preclass behavior knowledge scores. This finding may have resulted from a lack of properly defining these variables or from large variations in students' responses. For example, personal experience could have ranged from animal ownership to research or professional dog training. Books and trainers may have been individually valuable, but as a whole, neither significantly predicted students' incoming behavior knowledge. One possible explanation for the lack of association between books as a knowledge source and preclass knowledge scores was a failure to specify book type, specifically the inclusion or exclusion of pop-culture books. Popular press books range in the amount of evidence supporting them, even with publications limited to the last 15 years. The previously discussed range of educational experiences that even new veterinarians have14,15 may explain to some degree why noting the family veterinarian as a source of prior knowledge was also not a predictor of preclass behavior knowledge. Additionally, relationships between students and family veterinarians may have differed from those between students and mentoring veterinarians, and this was not specifically addressed. Variation among undergraduate behavior courses likely contributed to the lack of predictive value of such a course or reporting it as a source of prior knowledge.
In our sample of students, companion animal ownership did not predict an increase in preclass behavior knowledge scores. Other studies about companion animal ownership in developed countries have found knowledge gaps in a variety of topics ranging from basic care70,71 to zoonotic disease potential.72–74 However, in lay conversations, on social media, and on veterinary school admission applications, companion animal ownership may be touted as a valuable source of expertise. Given that knowledge gaps include important subjects that can impact not only the health and well-being of animals but also that of their owners,71,72 it is perhaps not surprising that the cognitive disconnect between ownership and knowledge75,76 extended to behavior knowledge in our students at the time of the preclass survey. Most students' self-assessments were minimally accurate; self-assessed skill in understanding the behavior and motivation of dogs predicted only a small increase (2%) in preclass scores for dog behavior knowledge, and no other self-assessments were significantly associated with respective preclass behavior knowledge scores. The use of single items with Likert-type scale ratings may have contributed to this finding, as single items tend to be less accurate than summed items.69 As first-year veterinary students, their initial self-rating may be classified as unconscious incompetence. This is considered the first level of skill competency, in which a learner is unaware that they do not perform a skill well.77,78 However, little if any correspondence between self-evaluation and actual competency is common, even among highly trained individuals and those who have completed medical training.79–82 Respondents from a sample of year-2016 veterinary school graduates were more likely to report feeling ready for day 1 of practice if they attended a full-semester class in the first year taught by a DACVB14 at an institution such as ours (and students from our institution likely contributed to that study). The question remains as to whether that assessment overestimates competency, owing to the disconnect between self-ratings and knowledge suggested by results of our study.
Students' performance on the knowledge testing portion of the survey improved substantially following the animal behavior course from a mean grade interpreted as failing academically to one considered academically acceptable (> 80%). Significant improvement was consistent across all scores (including species- and topic-specific subscores) between these time points. As the behavior class final examination scores of postclass survey takers were representative of final examination scores overall, the course successfully improved scientific behavior knowledge. Together, these findings suggested that the survey provided an accurate measurement of basic companion animal behavior knowledge for all 3 species (dog, cat, and horse) and all 3 subtopics (learning or training, body language and motivation, and dominance) to the exclusion of common myth-based information. Higher scores on the knowledge testing section of the postclass survey were predictive of slightly (< 1%) higher scores on the final examination for the behavior course. This small difference was likely attributable to 2 main effects: the difference in focus on animal behavior myths between the course (including the final examination) and survey questions, and the particularly narrow grade distribution for the final examination. Video evaluation of students' real-time responses to animal body language and mock case examples would strengthen the predictive validity of such a measure beyond the content and comparative variable validity examined in this study.
There were some potential sources of measurement error in our study, owing to technical difficulties with SATP questions in the preclass survey. Students may have been more likely to select myth-based feline behavior motivations if the > 1 answer choice had been properly enabled for that survey. Furthermore, students may have been more likely to select > 9 myth-based distractors related to canine dominance in questions about dog behavior on the preclass survey if that question had properly allowed multiple selection. It is disconcerting to consider this suggested the incoming students' scores might have been lower than those assessed. These methodological deviations were not ideal; however, the improved performance of students on the postclass survey overall and on these subtopics indicated that the option to select additional distractors did not decrease performance on the postclass survey. The potential bias in scores created by the technical error could have minimized the differences between the pre- and postclass survey scores.
There may also have been a component of selection bias for both surveys; students who believed they had substantial behavior knowledge may have been more likely to participate in the preclass survey, whereas students who believed they had gained much knowledge during the course may have been more likely to participate in the postclass survey. However, this bias likely would not have had a meaningful impact on students' scores in the 2 surveys and was not likely to affect the association between postclass survey and final examination scores. Additional investigation, including a larger number of participants over various class years and across institutions, is warranted to further assess the validity of the survey used in this study as well as the general state of animal behavior education. As a screening tool, the survey may be applicable for evaluation of less formalized educational programs in veterinary schools or for tracking students' behavior knowledge bases across changes in curricula or between curricula of different durations and depths. The disparity we found between self-assessed companion animal behavior knowledge and test performance for not-yet-educated students suggested that self-reports of competency might be inflated in the population that is least competent in this area. Coadministration of a day 1 readiness survey14 and the survey used in the study described here may provide the most complete picture of graduates' knowledge base and assessment for institutions to track efficacy.
If our incoming class was representative of incoming veterinary students overall, our findings emphasized the critical need for formalized behavior education for veterinary students. Incoming knowledge for the study sample of students was not only lacking but seemed consistent with behavior myths that are potentially detrimental to human and animal safety and welfare. Similar to previous research,14 results of the present study indicated that veterinary students were better prepared for day 1 behavior competencies after a full-semester animal behavior course than they were upon entry to this curriculum. Until full-semester animal behavior courses are available to every veterinary student, colleges of veterinary medicine are encouraged to focus on providing students with peer-reviewed, evidence-based journal information before and after graduation to aid animal behavior education and treatment. Graduates who meet day 1 competency in behavior have the potential to preserve more human-animal relationships, reduce loss-of-home and euthanasia rates attributable to behavior problems, and better uphold the oath they take to enter the profession.
ABBREVIATIONS
| DACVB | Diplomate of the American College of Veterinary Behaviorists |
| SATP | Select all that pertain or apply |
Acknowledgments
No third-party funding or support was provided for the study. The authors declare that there were no conflicts of interest.
The authors thank Shayna Moore for Qualtrics administration and Jenni Davids for additional editorial support.
Footnotes
Information on file with Ohio State University College of Veterinary Medicine.
Lilly ML, Watson B, Siracusa C. Educational and interventional program at a small shelter. In: Proceedings of the 2016 Veterinary Behavior Symposium. San Antonio, Tex: American College of Veterinary Behaviorists, 2016.
Qualtrics Survey Software, XM Online platform 2017, Qualtrics, Provo, Utah.
Google web search. Available at: www.google.com. Accessed Mar 2, 2018.
References
1. Holcomb KE, Stull CL, Kass PH. Unwanted horses: the role of nonprofit equine rescue and sanctuary organizations. J Anim Sci 2010;88:4142–4150.
2. Lenz TR. The unwanted horse in the United States: an overview of the issue. J Equine Vet Sci 2009;29:253–258.
3. New JC Jr, Salman MD, King M, et al. Characteristics of shelter-relinquished animals and their owners compared with animals and their owners in US pet-owning households. J Appl Anim Welf Sci 2000;3:179–201.
4. Patronek GJ, Dodman NH. Attitudes, procedures, and delivery of behavior services by veterinarians in small animal practice. J Am Vet Med Assoc 1999;215:1606–1611.
5. Salman MD, New JG Jr, Scarlett JM, et al. Human and animal factors related to relinquishment of dogs and cats in 12 selected animal shelters in the United States. J Appl Anim Welf Sci 1998;1:207–226.
6. Salman MD, Hutchison J, Ruch-Gallie R, et al. Behavioral reasons for relinquishment of dogs and cats to 12 shelters. J Appl Anim Welf Sci 2000;3:93–106.
7. Scarlett JM, Salman MD, New JG, et al. The role of veterinary practitioners in reducing dog and cat relinquishments and euthanasias. J Am Vet Med Assoc 2002;220:306–311.
8. Weiss E, Gramann S, Spain CV, et al. Goodbye to a good friend: an exploration of the re-homing of cats and dogs in the US. Open J Anim Sci 2015;5:435–456.
9. Lenz TR. The “unwanted” horse in the US: an overview of the issue. In: Revised Proceedings. The Unwanted Horse Issue: What Now? Forum. Washington, DC: USDA, 2008;24–28. Available at: pubs.nal.usda.gov/sites/pubs.nal.usda.gov/files/unwantedhorse2008.pdf. Accessed Jan 1, 2018.
10. Unwanted Horse Coalition. Unwanted horse survey. Available at: unitedhorsecoalition.org/wp-content/uploads/2019/04/unwanted-horse-survey.pdf. Accessed Sep 11, 2019.
11. ASPCA. Pet statistics. Available at: www.aspca.org/animal-homelessness/shelter-intake-and-surrender/pet-statistics. Accessed Dec 30, 2017.
12. Dinwoodie IR, Dwyer B, Zottola V, et al. Demographics and comorbidity of behavior problems in dogs. J Vet Behav 2019;32:62–71. Available at: www.sciencedirect.com/science/article/pii/S1558787818302727. Accessed Aug 26, 2019.
13. AVMA. Joint AVMA-FVE-CVMA statement on veterinary education. Available at: www.avma.org/KB/Policies/Pages/Joint-Statement-Veterinary-Education.aspx. Accessed Jul 14, 2017.
14. Calder CD, Albright JD, Koch C. Evaluating graduating veterinary students' perception of preparedness in clinical veterinary behavior for “Day-1” of practice and the factors which influence that perception: a questionnaire-based survey. J Vet Behav 2017;20:116–120.
15. Shivley CB, Garry FB, Kogan LR, et al. Survey of animal welfare, animal behavior, and animal ethics courses in the curricula of AVMA Council on Education–accredited veterinary colleges and schools. J Am Vet Med Assoc 2016;248:1165–1170.
16. Hammerle M, Horst C, Levine E, et al. 2015 AAHA canine and feline behavior management guidelines. J Am Anim Hosp Assoc 2015;51:205–221.
17. Knesl O, Hart BL, Fine AH, et al. Opportunities for incorporating the human-animal bond in companion animal practice. J Am Vet Med Assoc 2016;249:42–44.
18. Hart BL, Hart LA, Bain MJ. Canine and feline behavior therapy. 2nd ed. Ames, Iowa: Blackwell Publishing, 2006.
19. Horowitz A, Mills DS. Horowitz DF, Mills DS, eds. BSAVA manual of canine and feline behavioural medicine. 2nd ed. Gloucester, England: British Small Animal Veterinary Association, 2009.
20. Overall KL. Clinical behavioral medicine for small animals. St Louis: Mosby, 1997.
21. Bouton ME. Learning and behavior: a contemporary synthesis. Sunderland, Mass: Sinauer Associates, 2007;184–224.
22. Reisner IR. Differential diagnosis and management of human-directed aggression in dogs. Vet Clin North Am Small Anim Pract 2003;33:303–320.
23. Reisner IR, Shofer FS, Nance ML. Behavioral assessment of child-directed canine aggression. Inj Prev 2007;13:348–351.
24. Bain MJ. Myth busters: common misconceptions in behavior (proceedings). dvm360. Available at: veterinarycalendar.dvm360.com/myth-busters-common-misconceptions-behavior-proceedings. Accessed Dec 31, 2017.
25. Juarbe-Díaz SV. Behavioral medicine opportunities in North American colleges of veterinary medicine: a status report. J Vet Behav Clin Appl Res 2008;3:4–11.
26. Pike AL, Scheip J. Dog and cat behavior myths debunked. Veterinary Practice News. Available at: veterinarypracticenews.com/veterinary-behaviorist-debunks-dog-cat-behavior-myths/. Accessed Dec 25, 2017.
27. Horowitz A, Bekoff M. Naturalizing anthropomorphism: behavioral prompts to our humanizing of animals. Anthrozoös 2007;20:23–35. Post-print version. The Humane Society of the United States Animal Studies Repository. Available at: animal studiesrepository.org/acwp_habr/14. Accessed Dec 21, 2017.
28. Rajecki DW, Rasmussen JL, Modlin SJ, et al. Dog bites boy: judgments of blame and shame. Anthrozoos 1998;11:66–73.
29. Morris PH, Doe C, Godsell E. Secondary emotions in non-primate species? Behavioural reports and subjective claims by animal owners. Cogn Emotion 2008;22:3–20.
30. de Keuster T. Jung, Hildegard. Aggression toward familiar people and animals. In: Horwitz DF, Mills DS, eds. BSAVA manual of canine and feline behavioural medicine. 2nd ed. Gloucester, England: British Small Animal Veterinary Association, 2009;182–210.
31. Horowitz A. Disambiguating the “guilty look”: salient prompts to a familiar dog behaviour. Behav Processes 2009;81:447–452.
32. Mitchell RW, Hamm M. The interpretation of animal psychology: anthropomorphism or behavior reading? Behaviour 1997;134:173–204.
33. Wynne CDL. The perils of anthropomorphism. Nature 2004;428:606.
34. Cafazzo S, Valsecchi P, Fantini C, et al. Social dynamics of a group of free-ranging domestic dogs living in a suburban environment. J Vet Behav Clin Appl Res 2009;4:61.
35. Landsberg G, Hunthausen W, Ackerman L. Behavior problems of the dog and cat. 3rd ed. Philadelphia: Saunders, 2013;317–319.
36. Hartmann E, Christensen JW, McGreevy PD. Dominance and leadership: useful concepts in human–horse interactions? J Equine Vet Sci 2017;52:1–9.
37. Hetts S. Pet behavior protocols: what to say, what to do, when to refer. St Louis: American Animal Hospital Association Press, 1999;53–55, 105–198, 239–252.
38. Houpt KA. Terminology think tank: terminology of aggressive behavior. J Vet Behav Clin Appl Res 2006;1:39–41.
39. van Kerkhove W. A fresh look at the wolf-pack theory of companion-animal dog social behavior. J Appl Anim Welf Sci 2004;7:279–285.
40. Mech LD. Alpha status, dominance, and division of labor in wolf packs. Can J Zool 1999;77:1196–1203.
41. van Doorn GS, Hengeveld GM, Weissing FJ. The evolution of social dominance II: multi-player models. Behaviour 2003;140:1333–1358.
42. Zimen E. Social dynamics of the wolf pack. In: Fox MW, ed. The wild canids: their systematics, behavioral ecology and evolution. New York: Van Nostrand Reinhold, 1975;336–362.
43. Scott JP, Fuller JL. Genetics and the social behavior of the dog. Chicago: University of Chicago Press, 1965;18, 91, 109, 153–163.
44. Schenkel R. Expression studies in the wolf; captivity-observations [in German]. Behaviour 1948;1:81–129.
45. Herron ME, Shofer FS, Reisner IR. Survey of the use and outcome of confrontational and non-confrontational training methods in client-owned dogs showing undesired behaviors. Appl Anim Behav Sci 2009;117:47–54.
46. Mazur JE. Learning & behavior. 7th ed. New York: Routledge, 2013;127–153.
47. Bradshaw JWS, Blackwell EJ, Casey RA. Dominance in domestic dogs—useful construct or bad habit? J Vet Behav Clin Appl Res 2009;4:135–144.
48. Siracusa C, Provoost L, Reisner IR. Dog- and owner-related risk factors for consideration of euthanasia or rehoming before a referral behavioral consultation and for euthanizing or rehoming the dog after the consultation. J Vet Behav Clin Appl Res 2017;22:46–56.
49. Hess EH. Imprinting. An effect of early experience, imprinting determines later social behavior in animals. Science 1959;130:133–141.
50. Mal ME, McCall CA, Cummins KA, et al. Influence of pre-weaning handling methods on post-weaning learning ability and manageability of foals. Appl Anim Behav Sci 1994;40:187–195.
51. Simpson BS. Neonatal foal handling. Appl Anim Behav Sci 2002;78:303–317.
52. Søndergaard E, Jago J. The effect of early handling of foals on their reaction to handling, humans and novelty, and the foal–mare relationship. Appl Anim Behav Sci 2010;123:93–100.
53. Spier SJ, Berger Pusterla J, Villarroel A, et al. Outcome of tactile conditioning of neonates, or “imprint training” on selected handling measures in foals. Vet J 2004;168:252–258.
54. Williams JL, Friend TH, Toscano MJ, et al. The effects of early training sessions on the reactions of foals at 1, 2, and 3 months of age. Appl Anim Behav Sci 2002;77:105–114.
55. Williams JL, Friend TH, Collins MN, et al. Effects of imprint training procedure at birth on the reactions of foals at age six months. Equine Vet J 2003;35:127–132.
56. Fox MW, Beckoff M. The behaviour of dogs. In: Hafez ESE, ed. The behaviour of domestic animals. London: Bailliere Tindall Ltd, 1975;370–409.
57. Houpt KA. Domestic animal behavior for veterinarians and animal scientists. 6th ed. Ames, Iowa: Wiley-Blackwell, 2018;1–23, 34–39.
58. Wan M, Bolger N, Champagne FA. Human perception of fear in dogs varies according to experience with dogs. PLoS One 2012;7:e51775.
59. McDonnell SM. The equid ethogram: a practical field guide to horse behavior. E-book. London: Eclipse Press, 2003.
60. McGreevy P. Equine behavior, a guide for veterinarians and equine scientists. London: Elsevier Health Sciences, 2012;119–162, 313–321.
61. McGreevy PD, Henshall C, Starling MJ, et al. The importance of safety signals in animal handling and training. J Vet Behav Clin Appl Res 2014;9:382–387.
62. Overall KL, Tiira K, Broach D, et al. Genetics and behavior: a guide for practitioners. Vet Clin North Am Small Anim Pract 2014;44:483–505.
63. Scott JP. The evolution of social behavior in dogs and wolves. Integr Comp Biol 1967;7:373–381.
64. Drews C. The concept and definition of dominance in animal behaviour. Behaviour 1993;125:283–313.
65. Dohoo IR, Martin SW, Stryhn H. Veterinary epidemiologic research. Charlottetown, PE, Canada: VER Inc, 2009;365–445.
66. Blackwell EJ, Casey RA, Bradshaw JWS. Efficacy of written behavioral advice for separation-related behavior problems in dogs newly adopted from a rehoming center. J Vet Behav Clin Appl Res 2016;12:13–19.
67. Petrie A, Watson P. Statistics for veterinary and animal science. Chichester, England: John Wiley & Sons, 2013.
68. Feng C, Wang H, Lu N, et al. Log-transformation and its implications for data analysis. Shanghai Arch Psychiatry 2014;26:105–109.
69. Gliem JA, Gliem RR. Calculating, interpreting, and reporting Cronbach's alpha reliability coefficient for Likert-type scales. IUIPUI ScholarWorks Repository, 2003. Available at: scholarworks.iupui.edu/handle/1805/344. Accessed Dec 20, 2017.
70. Lesimple C, Poissonnet A, Hausberger M. How to keep your horse safe? An epidemiological study about management practices. Appl Anim Behav Sci 2016;181:105–114.
71. Ramón ME, Slater MR, Ward MP. Companion animal knowledge, attachment and pet cat care and their associations with household demographics for residents of a rural Texas town. Prev Vet Med 2010;94:251–263.
72. Steele SG, Mor SM. Client knowledge, attitudes and practices regarding zoonoses: a metropolitan experience. Aust Vet J 2015;93:439–444.
73. Sandhu GK, Singh D. Level of awareness regarding some zoonotic diseases, among dog owners of Ithaca, New York. J Family Med Prim Care 2014;3:418–423.
74. Fontaine RE, Schantz PM. Pet ownership and knowledge of zoonotic diseases in De Kalb County, Georgia. Anthrozoos 1989;3:45–49.
75. Tami G, Gallagher A. Description of the behaviour of domestic dog (Canis familiaris) by experienced and inexperienced people. Appl Anim Behav Sci 2009;120:159–169.
76. Demirbas YS, Ozturk H, Emre B, et al. Adults' ability to interpret canine body language during a dog–child interaction. Anthrozoos 2016;29:581–596.
77. Adams L. Learning a new skill is easier said than done. Gordon Training International. Available at: www.gordontraining.com/free-workplace-articles/learning-a-new-skill-is-easier-said-than-done//. Accessed Sep 6, 2018.
78. Maslow AH. Motivation and personality. New York: Harper & Row Publishers Inc, 1954;44–51, 78, 84, 261.
79. Mabe PA, West SG. Validity of self-evaluation of ability: a review and meta-analysis. J Appl Psychol 1982;67:280–296.
80. Marteau TM, Johnston M, Wynne G, et al. Cognitive factors in the explanation of the mismatch between confidence and competence in performing basic life support. Psychol Health 1989;3:173–182.
81. Risucci DA, Tortolani AJ, Ward RJ. Ratings of surgical residents by self, supervisors and peers. Surg Gynecol Obstet 1989;169:519–526.
82. Tracey JM, Arroll B, Richmond DE, et al. The validity of general practitioners' self assessment of knowledge: cross sectional study. BMJ 1997;315:1426–1428.
