Over the past 15 to 20 years, behavior evaluations have become the standard of care in shelters for identifying dangerous and potentially aggressive dogs that may not do well in a home as a pet. Today 28 percent of shelters use a formal evaluation and about 60 percent use some sort of modified behavior evaluation or other behavior observations to direct adoptions and to make euthanasia decisions. Often these euthanasia decisions are based upon signs of aggression during testing. The testing involves putting dogs in a variety of situations they might encounter in an adopter’s home and recording their responses, paying attention to any concerning behaviors that might indicate it would not be safe or humane to put the dog up for adoption to the public. Unfortunately, there is little standardization of these evaluations, with shelters using a wide range of different testing protocols.

The problem with behavior evaluations

Critics of behavior evaluations have become more and more vocal, questioning their validity and their usefulness in making placement decisions for shelter dogs. One recently published paper is highly critical of the practice of basing a euthanasia decision on potentially false positives for aggressive behavior, where a dog is flagged for aggressive behaviors in the shelter but would not display dangerous behaviors in a home setting. Many of the major criticisms lie in the fact that very little research has gone into developing a scientifically validated tool. Critics oppose reliance on non-validated tools to make life-or-death decisions in shelter dog populations. The argument can be made that using these evaluations is not a reliable way to make placement decisions because a dog may act very differently in a shelter setting than in a home, meaning the evaluation done at the shelter may not be a good predictor of how the dog behaves after being adopted.

The criticisms of this practice are warranted and deserve a great deal of discussion; we need more research and, I believe, standardization of evaluation protocols. The fact is, no person can predict the future behavior of any living animal with complete certainty; all we can do is try our best to make educated guesses at what will occur when a dog is in a different environment later in life.

When dealing with a population of animals that exceeds an organization’s adoption capacity, shelter workers need a tool to make decisions about the adoptability of dogs in their care, and behavior evaluations can be helpful in making decisions when you can’t adopt out all dogs. Using behavior evaluations may help to alleviate the compassion fatigue that is inevitable with such work because it gives shelter workers a clear system to use, so they may not feel as emotionally burdened by what can be very tough decisions.

Many shelters are encountering changing populations in their intake, and now have fewer dogs to adopt out. Fewer animals means that we now have the ability to work with and adopt to the public dogs with behaviors that in the past would have not been considered for placement.

Making decisions about placement, training, behavior modification, enrichment, and euthanasia is a daily tasks for shelter workers who are often not formally trained behaviorists. So how do they work with dogs, make sound decisions regarding placement, and do all they can to ensure public safety when placing dogs with behavior concerns? Today I am not sure that there is a perfect answer for this, but I would like to share how we at the Animal Rescue League of Boston have tackled this question.

Introducing MATCH-UP II

In 2011 we developed the MATCH-UP II Shelter Dog Rehoming Program.

The MATCH-UP II Shelter Dog Rehoming Program is a multi-part system designed to help shelters learn about the personalities and needs of their dogs so that behavioral interventions can be implemented and successful matches can be made. This web-based behavior evaluation is open for use by any shelter that finds it fits their needs. A great deal of research went into the development of MATCH-UP II, and in my opinion it the best evaluation currently available to help shelters understand, record, and communicate dog behavior.

Here is a video of the behavior evaluation component of MATCH-UP II.

Although the behavior evaluation—11 subtests that include food bowl testing, handling, high-value food guarding, history of training, dog to dog interaction, interaction with a strange-looking person and with the toddler doll—takes the most time in the rehoming process, it is not the only tool that is used to make pathway decisions for our dogs. Any available history and all observed behaviors in the shelter are also recorded and weighted with equal importance. All this information is calculated together to give the evaluator a triage report that includes personality scores, triage points, identified potential behavior concerns, and training protocols for any behavioral issues. Triage points correlate to concerning behaviors; an animal that gets a higher score in triage points may have a more pressing need for behavioral intervention. The entire system is designed to be objective because is based on whether a given behavior is observed or not observed, not on what the evaluator interprets the behavior to mean. Each behavior is scored independently, rather than combined into one score. Each specific behavior is a “building block” that we then combine into scores for a more comprehensive view into the dog’s overall behavior and, maybe, personality.

The use of history and shelter-observed behavior helps to get a better picture of a dog and takes the reliance off the evaluation itself as the sole decision-making tool. I often explain this to adoption staff in the light of two research papers—Mohan-Gibbons et al and Marder et al— on food bowl aggression in shelter dogs that have revealed that many dogs may show aggression over a food bowl during a shelter evaluation, but not in a home. We know, however, that dogs that have shown this behavior in a previous home are very likely to show it again in another home, which is why getting as accurate of a history as possible from previous owners is important. In our shelters, exhibiting food bowl aggression in any context will not automatically make a dog a candidate for euthanasia. We use the evaluation as a starting point to determine how easy this behavior will be to manage in a home. We then adopt out those dogs that are manageable, disclosing this information and demonstrating the behavior to the adopter prior to placement. We discuss what exactly has been seen and in what contexts with this particular dog, and counsel adopters on what that may mean in their home. Here is an example of one of our disclosure forms.

When looking at a dog’s performance on a behavior evaluation and triage report, we at the Animal Rescue League of Boston do not see MATCH-UP II as a set of pass or fail criteria but as a way to get to know our dogs as individuals with an objective tool that has been validated for inter-evaluator reliability. Our (as yet unpublished) research has shown that 82 percent of all recorded behaviors (approach person, wag tail, lip lick etc.) had a coefficient of reliability across evaluators of greater than 0.60, showing a strong reliability between evaluators. We used a special interrater reliability statistic and looking across all behaviors for all sub-tests, the overall average reliability coefficient was 0.77, with low of 0.15 and a high of 0.98.

Only in extreme and rare cases have we found a dog so dangerous that a euthanasia decision was made based solely upon the findings of the behavior evaluation itself. For most dogs, the information provided by MATCH-UP II is used to help identify play preferences, training protocols (generously provided by the Center for Shelter Dogs at Tufts University), best enrichment options (also found at the Center for Shelter Dogs website), and steps for further evaluation. For example, the handling subtest includes touching of all four paws.

This subtest only includes feet because our research showed a strong correlation between dogs that reacted negatively to paw handling and those that reacted to other types of handling such as muzzle holds or pushing into a sit. If a dog shows any signs of discomfort or aggression in this subtest, we will also look to the information our veterinary staff recorded during medical procedures, any history from a previous shelter or home, and our shelter staff and volunteers’ interactions with that dog. All of this will give us an idea of how uncomfortable the animal is with being handled, and the reliability of the results of this subtest. We can then further evaluate the dog to better understand what exactly is causing the discomfort or aggression when being handled. The next steps would be counter-conditioning and desensitization, if appropriate, and finally counseling adopters on exactly what we have seen in this dog and how to best train and manage that dog in a home.

How to use MATCH-UP II

The Animal Rescue League of Boston feels strongly that every dog needs to have their behavior observed, recorded, and communicated to staff, volunteers and adopters as objectively as possible without bias based on breed. Today we find that MATCH-UP II is the best tool for this purpose. The training and resources page on our website will help any new shelter staff member or volunteer get to know how the evaluation works, and each subtest has a video on how to conduct it so that consistency is easily achieved by evaluators. Technical and training support is also available to users who have signed up to be a shelter partner.

When first starting to use MATCH-UP II, it can seem a bit overwhelming. You may ask yourself, “What is a personality score or a triage point anyway?” The manual is a great way to start getting to know this system easily and is available online. Although entering and scoring the evaluation data using the online site is my first choice, worksheets like these, which include the intake profiles, evaluation and shelter-observed behavior can be printed out and used in the absence of computer access.

Personality scores are automatically calculated for you from the behavior evaluation, and give you an understanding of each dog’s friendliness to humans, overall fearfulness, aggression, playfulness, excitability, and cues that dog knows. The research conducted did show that three of these traits (friendliness, fearfulness, and aggressiveness) may enable identification of stable behavioral tendencies.

Triage points are also automatically calculated in the triage report from the history, evaluation, and shelter-observed behavior to identify potential behavior concerns. Here is an example Triage Report. As stated previously, animals with more triage points may be in greater need of intervention for their behavioral issues: Two dogs may have the same behavior concern identified—fear of adults, for example—but one will have a lower triage score, which means that individual may have less severe fear than another with a higher triage score. With a bit of training, shelter staff and volunteers with differing levels of experience and behavior knowledge can easily use this system to better understand the dogs they are serving and make well-informed placement decisions that include full disclosure of any and all potential concerns to adopters. Maybe most importantly, the shelter dog rehoming program automatically records this information in a way that makes it easily communicated to all involved.

When used properly, behavior evaluations—especially MATCH-UP II—can help educate staff, volunteers and adopters and make informed decisions about how to best serve the needs of their dogs. Since we have no perfect tool (I am uncertain that a perfect one will ever exist), this is the best way we have found today to move forward with our dogs, knowing the limitations of our ability to predict future behavior in any animal in all contexts, environments and states of well-being.

Dot Baisly, MS, CPDT-KA, Associate Shelter Member-IAABC is the behavior and enrichment manager of the Animal Rescue League of Boston, overseeing the behavioral care of all animals in 3 shelters. She has been working in animal welfare and behavior for over 17 years, both in animal welfare and rescue organizations in New York and New England, and with private clients as a consultant. When not working with shelter animals she also works with service dogs as a field representative for Paws With A Cause.