Evaluating Shelter Dogs: Insights into Their Behavior and Welfare

This dissertation explores resource-conservative methods for animal shelters to assess and modify dogs’ behaviors, aiming to provide tools for optimal care within limited resources.In Chapter 2, the Canine Behavioral Assessment and Research Questionnaire (C-BARQ) was adapted for shelter dogs. Factor analysis identified a five-factor structure (Fear, Arousal, Human Excitability, Dog Aggression, and Human Aggression) with 24 items. Psychometric evaluations, including item response theory and inter-factor correlations, demonstrated reliability and validity. This resulted in a validated questionnaire tailored for assessing shelter dog behavior, marking the first adaptation of the C-BARQ for shelter use. In Chapter 3, I examined whether the same items measure identical behavioral constructs in shelter and home environments and explored how shelter behavior predicts in-home behavior. Differences in factor loading magnitudes indicated the same behavioral constructs were not consistently measured across contexts. Differential Item Functioning analysis revealed functional similarities in some items from the Fear, Arousal, and Dog Aggression factors, but significant differences in responses for most items between shelter and home samples, indicating substantial contextual differences that impact the reliability of behavioral constructs measured in the home. In Chapter 4, a noncontingent reinforcement intervention was implemented with dogs across three shelters to reduce human-directed reactive behaviors. The intervention did not effectively decrease human-directed responses compared to baseline behavior. Behavior changes in the presence of the experimenter versus non-experimenters showed minimal evidence of generalization. Improvements were limited to the presence of the experimenter, with generalization observed only for behaviors associated with increased front-of-kennel orientation. Biting occurred most frequently on intervention days toward non-experimenters furthest removed in time from intervention. The intervention did not significantly reduce length of stay for dogs compared to control groups in two shelters. This dissertation aimed to identify resource-efficient methods for animal shelters to assess and modify dogs’ behaviors. Chapter 2 successfully developed a low-resource tool for assessing shelter dog behavior. However, Chapters 3 and 4 were less effective in achieving their aims compared to existing methods in the literature, which, although more resource-intensive, were also more efficacious. It is valuable to recognize that less resource-intensive interventions may also lead to lower efficacy.