Abstract
We evaluated a scheme for assessing shelter dog behaviour, which used 28 tests and rated responses from 0 (positive response) to 5 (fear, tonic immobility, or escape attempts). The assessment was evaluated for 236 dogs, and was repeated by a different assessor for 39 dogs approximately 80 days after rehoming to determine relevance of individual test components. A new owner survey evaluated satisfaction with the dog. A total of 130 of 236 dogs passed (score 80) failed. Scores were mainly unaffected by dog type and environmental variables, but decreased if dog faeces from a previous test was present in the arena during a test. Shelter tests only correlated with repeat tests if there was no direct contact with assessors. Adopters were satisfied with their dogs, despite reporting some behaviour problems. The shelter assessment was therefore robust against most outside influences but did not predict responses to people well.