Avaliação qualitativa do comportamento como indicador do temperamento em gatos domésticos (Felis silvestris catus)

Abstract

The Qualitative Behavior Assessment (QBA) is a method that builds the observation of animals' body language through descriptors, instead of focusing on isolated behavioral categories. As the QBA is a method subject to the evaluator's interpretation, it is necessary to evaluate the consistency between evaluators in order to enable its use as an indicator of temperament in animals. Thus, the aim of this study was to assess the validity and inter-observer reliability of QBA as an indication of cats’ temperament. Four behavioral tests were applied to 42 shelter cats: unfamiliar person (UP); novel object (NO); co-specific reaction (CO) and food offering (FO). The tests were filmed and assessed using an ethogram that included discrete behavioral categories and by a score of three categories created for each of the tests, generating quantitative information on the existing temperament dimensions. The videos were also assessed using the QBA method by 19 assessors of different profiles (i: experienced in animal behavior assessment; ii: inexperienced; iii: cat owners; iv: no-owners; v: all observers). The 20 descriptors used were scored in visual analog scales (active, affectionate, aggressive, agitated, attentive, alert, calm, confident, curious, fearful, friendly, indifferent, nervous, relaxed, sociable, stressed, suspicious, tense, vocal and greedy). The quantitative and QBA data were analyzed using Principal Component Analysis. One of the observers was designated as ‘gold’ because he already had experience with QBA, and its analysis was compared with the results extracted from the behavioral tests, by analysis of variance and Spearman's rank correlation. Kendall’s coefficient of concordance (W) was used to obtain inter-observer reliability. The Principal Components (PC) found, considered as the main dimensions of cats temperament, by using coding method were ‘fearfulness’ (PC1-UP), ‘neophobic’ (PC1-NO), ‘anticipation’ (PC1-FO), ‘sociability’ (PC1-CR), ‘indifference’ (PC2-UP, PC3-NO, PC3-FO) and ‘tolerance’ (PC3-UP, PC2-NO, PC2-FO, PC2-CR). While by using QBA we found the dimensions ‘calmness’ (PC1-QBA), ‘restless’ (PC2-QBA) and ‘aggressiveness’ (PC3-QBA), which identified four cat profiles: calm/active, calm/quiet, fearful/aggressive and fearful/flighty. The first PC obtained for each test were significantly correlated with the PC1-QBA (calmness) and PC3-QBA (aggressiveness) was correlated with the score that represented aggressive behaviors in th UP test. This correlations showing that ‘gold’ observer captured aspects evidenced by the first dimension of all behavioral tests and the aggressive behaviors expressed only by the score, validating the QBA as a method of identifying the temperament of domestic cats. The reliability of the dimensions for the 19 observers showed ‘high’ agreement on PC1 (0.71) and ‘low’ for PC2 (0.21) and PC3 (0.29). Thus, the QBA was considered a valid and feasible tool for identifying and differentiating cat temperaments, as well as reliable for a ‘calmness’ dimension. The dimensions found, regarded as valid and reliable, are important traits of cats temperament, being promising in practical shelter environments as a basis for establishment of effective adoptive processes and management strategies to domestic cats kept in shelters.