The impact of excluding food guarding from a standardized behavioral canine assessment in animal shelters

To best understand this article in the context of the behavior evaluation literature, please see National Canine Research Council’s complete analysis here.

Article Citation:

Mohan-Gibbons, H., Dolan, E. D., Reid, P., Slater, M. R., Mulligan, H., & Weiss, E. (2018). The Impact of Excluding Food Guarding from a Standardized Behavioral Canine Assessment in Animal Shelters. Animals8(2), doi:

National Canine Research Council Summary and Analysis:

The prospective experimental study by Mohan-Gibbons et al. (2018) has many strengths which contribute to the discussion on behavior evaluations in general, and food guarding specifically. This study builds upon the previous food guarding research and echoes the sentiments that scientists before them have suggested: food guarding assessments are unreliable and unnecessary. The paper is included because it contributes new knowledge from a different perspective (effects on the shelter versus effects on the adopter), and deepens our understanding of food guarding assessments. Though several other studies have examined food guarding and found them to be insufficient, they are still used daily in shelters across the country. Thus, the conversation continues.

Nine shelters—private and municipal—from seven states participated in the study. Inclusion criteria required that the shelters already screen for food guarding behavior, and were willing to cease assessments for the treatment period of the study. The study lasted 5 months which included two months of baseline, two months of treatment, and a final month of data collection only. In baseline, shelters were instructed to continue their assessments as normal, and report outcome variables on a monthly basis. Dependent variables included canine intake; returns; average length of stay; number of adoptions, transfers, and euthanasia; bites and injuries in the shelter; and bites and injuries in the home post adoption. It is important to note that bites and injuries were combined to obtain the most conservative estimate possible—any injury was counted against the dog, even if food was not involved. During treatment, shelters were asked to cease all formal food guarding assessments, but to continue as usual with all other programs and data collection.

In the shelter, food guarding was identified either by a formal assessment (baseline), owner history, or observations from staff, volunteers, or potential adopters. A major strength of this study is its external validity: for baseline, each shelter used their existing food guarding procedures.

Straightforward analyses (chi square) were used to identify outcome differences between baseline and treatment, and dogs with and without documented food guarding. Overall, occurrence of food guarding was low; only 5% of dogs exhibited food guarding, or were reported to from owner history. Bites and injuries were rare for both groups (food guarders and non-food guarders); in the home; there were 43 reported injuries with 35 of those coming from non-food guarders. The percentage of dogs exhibiting severe food guarding did not change, even when formal assessments were discontinued. This supports the arguments commonly made by opponents of formal assessments: problem behaviors can be identified without diverting resources to biased evaluations.

Across both study phases, dogs that showed food guarding behavior had a longer length of stay (M = 13.6 days, range = 4.7–44 days) than the general dog population (M = 9.9 days, range = 3.6–20 days). When the shelters stopped assessing for food guarding, there was a 3% increase in overall returns (from 10% (379/3867) to 13% (526/4094)). However, there was no difference in the rate of returns of food guarding dogs, even though more dogs were adopted because fewer were identified with food guarding behavior.

Approximately 11% of dogs (1592/14,180) were euthanized during the study, with dogs showing food guarding behavior being euthanized at a higher rate (15.4%, 120/778) than dogs who did not show that behavior (10.9%, 1472/13,402), p < 0.001. Of the 120 dogs euthanized in the food guarding group, 74% (n = 89) had behavior listed as a reason for euthanasia, and 45% of those had food guarding listed as a reason for euthanasia.

In sum, the authors argue that because formal assessments did not affect the percentage of dogs identified with food guarding behavior, and because removing the assessments did not increase the number of injuries or bites, formal food guarding assessments should be discontinued. These findings and recommendations are in line with other recent papers including Marder et al. (2013), Mohan-Gibbons et al. (2012), and Patronek and Bradley (2016).

One design weakness should be considered: all shelters were in baseline and treatment at the same time. In this paradigm, a multiple baseline design should have been employed. Alternatively, a reversal design would also have been appropriate, although less ethically sound given the results. Because of this design flaw, potential confounds, such as time and history effects, cannot be ruled out.

Finally, there was an interesting and unanticipated finding: during baseline, only 49% of dogs were evaluated for food guarding. The remaining 51% were either returned to owners before they could be evaluated, or bypassed the process because they were of desirable “appearance, size, or breed.” This reinforces what experts have long suspected: that shelters pick and choose when and how to evaluate dogs, further weakening their utility.

Link to Full Text of the Original Article: