November 16, 2015

Development and validation of method for evaluating behavior and temperment of guide dogs

In dog breeding for guide dogs one of the big issues is a dog’s temperament. If the dog has a bad temperament, then they will not make a good guide dog for a blind or partially blind person. People are constantly looking for new tests for dogs’ temperament. For this test, they have the breeder answer questions about how the puppy acts in the first 6-12 months of life. The approach assumes that the breeder has the most knowledge about the dogs since they spend the most time together. The questionnaire asks breeders for a ranking of the puppies’ responses to different stimuli on a scale. The list of questions was determined based on a list given by guide dog instructors and supervisors as well as time observing dogs in the field. Then a secondary questionnaire was created based on a panel of responses. The final questionnaire had 40 questions. The questionnaire compared 1067 animals from four breeds: Labrador Retriever, Golden Retriever, German Shepherd, and Labrador Golden Retriever mix. The questionnaire was compared to the already existing method of behavior analysis to check validity. Any dog that is rejected from training is given a single primary reason for being rejected.

There were 40 total questions. Of the questions two were often skipped over because of lack of stimuli for the dog to respond to. In the end one of these questions was dropped due to lack of responses. The remaining questions were placed into categories based on how the aggression was directed. Aggression was categorized as stranger directed, owner directed, dog directed or other categories leading to a total of eight factors. Some factors from the physical tests were excluded because their relation to behavior could not be proved. The reason for rejection should line up with the answers given on the questionnaire and the subgroup that most often matched up.  However some categories lined up better than others: fear of strangers lined up with lack of confidence as expected, distraction lined up with energy levels as expected as well. Of the 16 expected associations, 14 were statistically or marginally significant. This is true even though the reason for rejection occurred 2-12 months after the dogs had been surveyed. Of the 56 non associations, three were statistically significant. The results indicate that the survey can be predictive but should not be considered grounds for exclusion. They also found that the puppy raiser’s information is a useful tool in determining a dog's temperment; however, further information is necessary so final evaluation cannot occur at this time. As is further research into this specific questionnaire.


Serpell, James.a, and Yuying Hsu. "Development and Validation of a Novel Method for Evaluating Behavior and Temperament in Guide Dogs." Applied Animal Behaviour Science 72 (2001): 347-64. Print.