In a recent study in the journal Current Biology, researchers assert that shelter dogs who show behavior indicative of separation distress tend to be pessimistic, compared with more optimistic dogs who are less likely to exhibit separation-related behavior. I’m going to explain briefly how the experiment was conducted and then discuss my concerns with the researchers’ conclusions.
In their experiment, 24 shelter dogs were taught that a bowl in one location had food in it, while a bowl in another location was empty. Once the dogs were trained to this paradigm, they were tested to determine whether or not they had a “pessimistic” cognitive bias, or an “optimistic” one. In the test, bowls were placed in locations other than the ones that the dogs had been trained to understand. These ambiguous locations were in the same room as the tests with bowls that had either been empty or containing food during training. The time it took for the dogs to approach the bowls in these new locations was recorded.
Dogs who went quickly to bowls in ambiguous locations were regarded as having an optimism that the bowl would contain food, while dogs who were slow to approach the bowl were considered to be pessimistic about the likelihood that the bowl would contain food.
In another part of the study, these same dogs were observed to determine how much time they spent exhibited separation-related behavior patterns such as vocalizing, destructive chewing, and inappropriate elimination. The researchers found that “pessimistic” dogs showed more separation-related distress than the “optimistic” dogs, and thus concluded that the negative affective state of these pessimistic dogs is correlated with separation distress.
My concern about this study is that I’m not convinced that the time until a dog approaches a bowl in an unknown location indicates optimism versus pessimism. What if degree of curiosity or tendency to fear new things is more relevant, rather than a cognitive decision about the likelihood of food being present? It is even possible that the dogs who were slower to approach the bowls were not as good at generalizing from the learning task or that they spent time considering what to do rather than acting impulsively. Or, perhaps the dogs who were slow to approach the bowls don’t tend to investigate things that are not theirs? (For dogs in home settings, we call this being “well-trained” or “well-behaved.”)
The authors say that the results of the experiment were “unlikely to be explained by running speed/motivation, learning ability, or other dog characteristics” but except for running speed, they did not control for them. The researchers have provided evidence that dogs who are slower to approach a bowl in an ambiguous location are more likely to exhibit signs of separation distress, but I don’t think they have made a strong case that they can conclude more than that. They have not demonstrated a correlation between separation related distress and a pessimistic cognitive bias. There are too many other possible explanations that need to be sorted through and tested for such a claim to be convincing.