A refinement and validation of the Monash Canine Personality Questionnaire (MCPQ)

https://doi.org/10.1016/j.applanim.2008.09.009Get rights and content

Abstract

The investigation of canine personality has failed to find strong agreement between studies as to its structure. The area has been hampered by a reliance on human personality models and by a tendency to limit the types of dogs used as subjects. These problems were recently addressed during the development of the Monash Canine Personality Questionnaire (MCPQ). In this follow-up study, over 450 participants provided demographic information about themselves and their dog and completed the MCPQ for their dog. Structural Equation Modelling results necessitated reassessing the original data and reanalysis suggested a more succinct questionnaire, the MCPQ-R, to measure five dimensions of canine personality very similar to those revealed in earlier work (extraversion, motivation, training focus, amicability and neuroticism). Owner reports of personality were generally consistent across demographic variables for all five canine personality subscales. There was no association between any subscale score and owner gender or education level, or dog sex or sexual status (desexed or not desexed). Significant, but generally weak associations were found for owner age and Extraversion (r = 0.17, P < 0.001), owner knowledge of their dog and Training Focus (r = 0.22, P < 0.001), time spent inside and Extraversion (r = −0.13, P = 0.007), dog age and Extraversion (r = 0.14, P = 0.004), Training Focus and Extraversion (r = 0.13, P = 0.007), dog height and Neuroticism (r = −0.20, P < 0.001) and dog height and Amicability (r = 0.2, P < 0.001), dog weight and Neuroticism (r = −0.17, P < 0.01) and dog weight and Amicability (r = 0.19, P < 0.01). There were also few differences in personality ratings across recognised purebred dog breed groups, although Working Dogs and Terriers scored significantly more highly on the Extraverted subscale than other groups (P < 0.001) and Working Dogs and Gundogs scored more highly on the Training Focus subscale (P < 0.001). These results are consistent with the view that the MCPQ-R assesses a construct, canine personality, which is relatively stable and comprised of five dimensions.

Introduction

Personality in dogs has been investigated by several groups, with existing literature showing mixed results as to the structure of canine personality. In particular, the personality dimensions identified have been varied in their number and nature (Hart and Miller, 1985, Svartberg and Forkman, 2002, Jones and Gosling, 2005). This may be partly because of two limitations in existing studies, identified by Jones and Gosling (2005). These are, first, that existing studies often rely on unique populations of dogs, such as guide dogs or police dogs, where heterogeneity of personality traits may be restricted and, second, that personality traits from the human psychological literature are typically applied to dogs without initially testing the assumption that dogs and humans have the same personality dimensions. Reconciliation of the variable results is made even more difficult by a lack of follow-up studies attempting to replicate initial findings in independent samples. In human personality research, constant revisions and modifications of personality models in response to ongoing studies has led to much stronger results.

In a previous paper (Ley et al., 2007) we described the development of a canine personality scale, the Monash Canine Personality Questionnaire (MCPQ), using an adjective-based methodology similar to that used in the formulation of a popular model of human personality, the Big Five Model (BFM) (John, 1990). A comprehensive list of personality adjectives (67), considered by experienced people to be applicable to dogs, was given to a large sample (1016) of companion dog owners for them to rate how well each word applied to their dog. The data were subjected to exploratory principal component analyses and five factors were identified and then confirmed in a number of follow-up analyses. These were well represented by just 41 adjectives (see Table 1) and were tentatively labelled extraversion, self-assuredness/motivation, training focus, amicability and neuroticism; although it was noted that additional research was required to establish whether these names were appropriate and whether the factors are both reliable and valid. In our conceptualisation (see Table 1) extraversion describes the perceived energy level of the dog; self-assuredness/motivation characterises a dog's perceived persistence in the face of distractions (for example begging for food, finding a particular toy); training focus describes the perceived trainability of the dog; amicability refers to how the dog is perceived to tolerate other individuals, be they humans, dogs or other animals; and the last factor, neuroticism, describes how cautiously or nervously a dog is perceived to behave.

The aim in this study was to use the 41-item MCPQ to further explore the personality of Australian companion dogs. We were interested in testing whether the five factors identified previously were sufficiently stable to reappear in a smaller sample, therefore supporting the construct validity of the five dimensions of personality described by the MCPQ. We were also interested in examining average ratings on the five personality subscales believed to measure these personality dimensions and the range of variability evident in a broad canine population. Most importantly we were interested in whether canine personality subscale scores might differ depending on owner characteristics, such as age, sex, education and location or type of residence, or depending on dog attributes, such as breed, size, age, sex, sexual status (desexed or not desexed), physical build, height and weight. Although the popular media tends to associate some canine characteristics, such as aggressiveness or friendliness, with certain breeds or types of dog, such as Rottweilers or Labrador Retrievers, the extent to which these stereotypes are accurate is unknown.

Section snippets

Participants

Participants were 588 adults, at least 18 years of age, who owned a dog aged at least 18 months of age. They were recruited through the Victorian Canine Association, veterinary clinics and dog training establishments around metropolitan Melbourne, and from around Australia through reports of the study in local and national media outlets. Following the protocol used in Ley et al. (2007), questionnaires were removed from analysis if the respondent was younger than 18 years old, spoke a language

Structural equation modelling

Structural equation modelling using the data from the previous study suggested that a stable 26-item solution could be extracted from the 41 items administered as part of the MCPQ. This solution was tested using the data from the current study. Four goodness-of-fit indices were calculated: Normed Chi Square (CI/DF); Root Mean Square Approximation of Error (RMSEA); Comparative Fit Index (CFI); and Root Mean Square Residual (RMR). The RMSEA, recommended to be 0.05–0.08 to be a reasonable fit (

Discussion

Analysis of data collected in this study using the MCPQ prompted a minor revision of the model of canine personality developed previously (Ley et al., 2007), and a corresponding revision of the MCPQ. The new model retains five personality dimensions, four (extraversion, training focus, amicability, neuroticism) virtually identical to those described previously. The fifth dimension identified in our previous study was self-assuredness/motivation, which has now been replaced by a personality

Conclusions

This study continues the characterisation of the personality of dogs kept as companions, using an instrument only recently developed using a large sample of Australian dogs and modified slightly on the basis of the data collected. The dogs in our sample were found to score, on average, moderately on personality subscales believed to measure extraversion and motivation, more strongly on subscales measuring training focus and amicability and more weakly on a subscale believed to assess

References (27)

  • A.L. Edwards

    Social desirability and personality test construction

  • J.A. Hair et al.
    (1998)
  • B.L. Hart et al.

    Behavioral profiles of dog breeds: a quantitative approach

    J. Am. Vet. Med. Assoc.

    (1985)
  • Cited by (0)

    View full text