Abstract

Assistance Hearing Dogs provide independence to people with a hearing impairment by alerting to household and danger sounds, and provide companionship in a range of public places. Hearing Dogs experience thorough socialising and training in order to meet high Assistance Dog standards. Present procedures in use by Hearing Dogs For Deaf People, UK involve an 8 months behavioural test, with the aim to predict each dogs’ future training performance. Knowing the predictive validity of this test, with respect to task performance is of great importance. The sensitivity, specificity, and predictive validity of 8 month behaviour tests already in use by the UK charity, Hearing Dogs for Deaf People, for performance at 18 months were evaluated on 62 dogs over a two year period. Correlations between the two tests and significant differences in performance were assessed using Spearman’s Rho and Wilcoxon matched pairs tests. The following 13 behavioural factors were analysed: social behaviour with adults, children and dogs, environmental behaviour, recovery rate, adaptability, vocal reactivity, motivation, trainability, frustration, chase, hunt and distraction. Convergent validity of both the 8 and 18 month behaviour tests was further examined where possible using a CBARQ questionnaire completed at 8 and 18 months by the socialisers and trainers of 12 of the dogs. Positive correlations were observed between all behavioural factors with the exception of distractibility and recovery rate, however overall the behavioural tests demonstrated a high sensitivity and a low specificity, signifying a poor predictive validity with respect to task performance. While these results partly reflect the small number of failures in each test, the findings may also be symptomatic of the continued socialising and training that all dogs, irrelevant of pass rates, undergo between the 8 and 18 month behaviour tests. The value of the 8 month test and its individual elements is critically appraised.