Using Model Dogs to Assess Aggression in Fight-Bred Dogs

Screening a dog for aggression toward other dogs (conspecific aggression) is a common practice in animal shelters before allowing dogs to attend a shelter playgroup or interact with other dogs. Shelters may also use it to assess whether a dog may pose a threat to other dogs in the community.

Using a model plush dog instead of a live “helper dog” reduces the risk of injury to the helper dog and the evaluator. It also enables shelters to screen for dog aggression even if helper dogs are not available.

Shelter-based research on the use of model dogs to predict behavior is equivocal – there is no strong evidence to support the use of model dogs as screening tools for normal, low-level aggression.

However, there hasn't been any research on the utility of using a model dog for detecting or predicting aggression in dogs bred to fight other dogs. Some people have suggested euthanizing all fight-bred dogs without consideration for individual behavior or traits. This research provides new insights into that assumption. 

Study Design

This study used a life-sized plush model dog as a screening tool for dog-directed aggression in a population of dogs bred for use in dogfighting. Researchers predicted that dogs who showed aggression to dogs would behave consistently toward both the model and real dog.

The study sample comprised 292 fight-bred pit bulls seized from illegal dogfighting operations in July 2009. The term "pit bull" is a description used by dog fighters. The dogs were not genotyped and were not necessarily American pit bull terriers or any other pure breed. 

Researchers screened each for dog aggression with 4 different stimuli: exposure to same- and opposite-sex dogs, a model dog, and a control object. Evaluators recorded the dogs' reactions to the various stimuli as:

• Aggressive

• Friendly

• Fearful

• Neutral

• Pushy

• Sexual

Study Results

Here are some of the findings:

• 90% of the dogs were classified as Neutral to the control object.
• Most dogs were not aggressive in any test scenario.
• Over 75% of test dogs aggressive to the same-sex helper dog scenario (86 dogs) were also aggressive to the model dog, making the model's sensitivity for conspecific aggression 81%.
• Of the 94 test dogs aggressive to the model dog, 70 were also aggressive to the same-sex helper dog, meaning the positive predictive value of the model dog was 75%.

• The remaining 24 test dogs were not aggressive to the same-sex helper dog for a false alarm rate of 26%.
• When Pushy responses to the same-sex helper dog were included, 75% of the false positives were removed, increasing the model dog's positive predictive value for conspecific aggression to 94%, dropping the false alarm rate to 6%, and increasing sensitivity to 85%.

Study Takeaways

Findings suggest the following:

• It's a misconception that all fighting dogs need to be euthanized for aggression. Most dogs in this study were not aggressive in any test scenario, which underscores the importance of considering the behavior of the individual animal as opposed to breed alone.  
• Screening for dog-directed aggression with model dogs is effective for fight-bred dogs and substantially reduces the risk to both helper dogs and evaluators by screening out the majority of extremely dog-aggressive dogs.  
• Based on the findings, the researchers recommend using a model dog first when evaluating fight-bred dogs. If the dog shows no physical aggression toward the model or the behavior is ambiguous, evaluators should test the dog with a same-sex helper dog. 

Read the full study.