Assistance dogs play a vital role in improving the lives of individuals with disabilities, yet training them is time-consuming and expensive. A 2019 study by Canine Companions for Independence (CCI) analyzed thousands of dogs using two behavioral tools to improve early selection for training programs.
The researchers evaluated data from more than 5,900 dogs, using two instruments: the Canine Behavioral Assessment and Research Questionnaire (C-BARQ) and a standardized In-For-Training (IFT) temperament test. These assessments were conducted early—at the start of formal training—offering a chance to spot likely dropouts before resources were heavily invested.
The study developed predictive models based on each test separately and in combination. All models performed significantly better than random chance, achieving 64–68% overall accuracy. Predictions were strongest at the extremes: dogs in the bottom 10% for success probability were correctly identified 85–92% of the time, while dogs in the top 10% were identified with 62–72% accuracy.
Interestingly, combining the C-BARQ and IFT scores did not improve general accuracy but did enhance identification of dogs at the very low end of success potential. This makes such screening tools highly valuable for early-stage decision-making in large assistance dog programs, allowing for better resource allocation and improved program outcomes.
The results suggest that objective, validated behavior tools can be effectively integrated into dog training programs to guide decisions long before the costly stages of specialized assistance work begin.
Source: Emily E. Bray, K. M. Levy, B. S. Kennedy, D. Duffy, J. Serpell, E. MacLean, Frontiers in Veterinary Science, February 27, 2019.
