Researchers have developed a questionnaire that could help evaluate whether puppies will be a good match for guide dog programs.

Finding out whether dogs will be a good fit in service programs is a priority for organizations that train and place dogs because the cost in training them is so high. Being able to weed out dogs before spending the time and money to train them unsuccessfully would benefit programs and allow them to focus resources. Earlier this year, a study reported in NEWStat focused on conducting fMRIs to see whether there were any neurological clues to what makes dogs suited for service work.

The latest study aimed to see if a questionnaire could help identify dogs who are not suited to becoming guide dogs early on. The researchers, from the University of Nottingham, worked with an organization called Guide Dogs who applied the questionnaire to their own prospective puppies. The open-access article was published in the June 2017 issue of PLOS One.

Aside from the main purpose of the study—to develop an assessment that would help identify dogs suited to the training process—researchers also wanted to test the dogs at several different times to see what would happen to one dog’s score across months.

For this reason, the assessment was completed when the puppies used for the study were 5, 8, and 12 months old. Researchers selected a pool of 1,401 puppies and 54 training supervisors.

The questionnaire itself, dubbed the puppy training supervisor questionnaire (PTSQ), was developed from previous questionnaires that had demonstrated predictive validity. Additional items were added by researchers to ensure all important areas were covered. A draft of the questionnaire was then refined based on a feedback from a panel of puppy training supervisors.

The researchers were particularly interested in seven targeted traits covered by two or more items in the questionnaire: Adaptability, Body sensitivity, Distractibility, Excitability, General anxiety, Trainability, and Stair anxiety.

Thresholds were created to determine at what score the dogs were considered “at risk” for each trait. Then, the scores the dogs received were mapped out. For each trait, a dog could be assigned a “green” flag, meaning that they scored within the threshold that identified qualified dogs. A “red” flag indicated they scored within the threshold for identifying withdrawn dogs. If a dog had a higher than 50% likelihood of being withdrawn but didn’t fall inside the threshold for a red flag, they were assigned a “yellow” flag.

Looking at the flags given, a majority of green flags were assigned to qualified dogs at each age. A majority of the red flags were assigned to dogs withdrawn for behavior. Yellow flags were split about down the middle—of the 161 dogs flagged, 68 qualified and 72 were withdrawn for behavior. Using the assessments from 5, 8, and 12 months, green flags had a positive predictive value (PPV) of 8.4% and a sensitivity of 85%. Red flags had a PPV of 8.4% and a sensitivity of 83%. The high sensitivity percentage indicates good target discrimination.

By identifying dogs with red and yellow flags, trainers and supervisors could step in for interventions with those dogs earlier to see if the scores could be improved. This might prevent withdrawals for behavior, or it could pull the dogs before training to prevent costs. While the PPVs in this study are lower than standardized tests of working dog behavior, the researchers make the argument that those tests “are costly to implement and to evaluate in large organizations such as Guide Dogs, where questionnaire based assessments would be comparatively economical and feasible.” Even only targeting the most unsuitable dogs and conducting early interventions to improve qualification rates could save both time and costs.

Photo credit: © iStock/nycshooter