The role of observer individual differences in personality assessments of the domesticated horse: A novel application of Kelly’s Repertory Grid Technique

Consistent personality differences (e.g. shyness, sociability) among animals have been frequently reported by scientists. Such information used to be considered unscientific, even though it was widely used to describe animals and predict their behaviour (Hebb, 1946). Over the past three decades however animal personality has been studied systematically in various species (Gosling, 2001), using provided lists of descriptors, consistent with the five-factor model of personality (Costa & McCrea, 1992). Few researchers have allowed raters to produce their own list of descriptors (Dutton et al., 1997; Wemelsfelder et al., 2000), and none have investigated the potential influence of human individual differences on ratings. The three main goals of this thesis were (i) to develop in-depth personality profiles of domesticated horses, (ii) to investigate the effect of personal backgrounds and attitudes of observers in their construction of these profiles, and (iii) to provide insight into the reliability and validity of the provided assessments of horse personality. The thesis starts with a literature review of personality psychology, psychometrics, animal and horse personality; followed by the introduction of the concept of animal-as-a-scientist. As methodology is an important part of the project, the repertory grid technique (RGT) originally developed by Kelly (1955), was employed as a novel method to assess horse personality. This method gives observers the freedom to generate their own descriptors, which allows them to integrate the totality of their experience with animals into personality constructs which are meaningful to them. RGT has previously been used for personality assessment in chimpanzees (Dutton et al., 1997) and for the assessment of individual styles of interaction in pigs (Grajfoner et al., 2002). The experimental part of the thesis consists of two studies. In both, horse personality was assessed by groups of human participants, either familiar or unfamiliar with the horses. The familiar groups assessed the horses based on their past experiences. The unfamiliar groups watched short videos of horses interacting with a human. In the first study 44 female observers rated 21 horses from two stables. The results show a significant degree of agreement within the observer groups for 95% of horses. However, correlation of horse personality scores between the observer groups was not significant. The degree of agreement was higher when the observers were familiar not only with the horses but also with the descriptors (Adams-Weber, 1970). In the second study we further investigated the observed incongruity between familiar and unfamiliar observers. A novel object test was added to provide the unfamiliar observers with more information about the horses. Thirty four female observers rated 38 horses from three stables. Again, the degree of agreement was highly significant within the observer groups; however, correlation of the results between familiar and unfamiliar observers was only marginally improved. The degree of agreement was not consistently higher when the horses were rated on elicited constructs. Both studies showed that two personality dimensions, neuroticism and extroversion, are fairly robust in horses. The third dimension, agreeableness, was generated only by familiar observers. Contrary to expectations, observers’ personality, empathy or emotional intelligence did not significantly affect the degree of agreement or how they rated the horses’ personalities. Assessment of individual differences in horses using qualitative descriptors generated by the observers themselves is therefore not observers’ self projection. These results make a significant contribution to the debate on anthropomorphism. The overall degrees of agreement between the observers indicates consistent reliability of the RGT throughout the observer groups in both studies. Finally, the academic and practical implications of the study are discussed. On an academic level, individual differences in animals are of pivotal importance for understanding personality in the contexts of evolutionary, comparative and social psychology. On a practical level horses are, according to their individual differences, selectively used for different purposes: horse assisted therapy, racing and the police. Avenues for investigating the relationships between animal personality, performance and welfare should therefore be further explored