Evaluating the generalizability of a standardized behavior test in predicting programmatic success of guide dog puppies

The ability to improve the selection process of working dogs offers significant advantages to organizations dedicated to breeding, raising, and training assistance and service dogs. Early identification of behavioral suitability has been proposed to allow for optimization of resource expenditure, resulting in not only more successful working dogs but in improved working dog welfare. To date, mixed results have been found in utilizing behavioral selection tests to predict working dog success. Here, we utilize the Guide Dogs for the Blind Association, United Kingdom (GDBA UK) Puppy Profiling Assessment (PPA) to evaluate whether this standardized test of behavior can be applied to predict success in another guide dog population at Guide Dogs for the Blind, United States of America (GDB). Three-hundred and one potential guide dog puppies (7 weeks of age), 249 which met final inclusion criteria, were presented with eight subtests: 1) follow; 2) retrieve; 3) restraint; 4) noise; 5) stroking; 6) a moving "squirrel" toy; 7) tunnel; and 8) ramp, in which their responses were scored on a sevenpoint scale. Unlike Asher et al., (2013) which identified that five of these subtests were associated with success in guide dog training, no subtest assessed at Guide Dogs for the Blind provided predictive value for a dogs ultimate success. This lack of generalization within guide dog programs highlights the importance of continuing to assess the suitability, predictability, and reliability of standardized behavior testing across working dog programs.