Consistency of dogs’ reactions to threatening cues of an unfamiliar person

Judit Vas, József Topál, Borbála Győri, Ádám Miklósi

Consistency of dogs’ reactions to threatening cues of an unfamiliar person

Applied Animal Behaviour Science, Volume 112, Issues 3–4, August 2008, Pages 331–344

Abstract

Reliability is one of the most important aspects of the behaviour observations measuring personality traits in animals. The most fundamental way to test reliability is the assessment of the test–retest consistency. On the other hand, in situations where social interaction between a human participant and the animal subject is at the scope of the study, the behaviour of the humans in the test situation should be restricted by a set of rules. However, if more than one experimenter participates in the observations, the similarity of the behaviour of different persons is an important aspect of the reliability of the study.

In our first study we investigated the consistency of dogs’ behaviour during approach by a person in a friendly and in a threatening way, repeating the test either immediately after the first test or at least 6 months later. We found that the dogs’ sensitivity to the human's behavioural cues in this situation proved to be consistent over repetition when the second test was performed approximately 1 year after the first test, but it was not the case when the time elapsed between test occasions was a few seconds. The detailed analysis indicated that dogs performing extreme behaviour (friendly or threatening) tended to show more conservative responses than dogs showing intermediate reactions when the stranger approached threateningly. Nonetheless, the reaction of the dogs failed to be in accordance with the opinion of the owners about the dogs’ behaviour in similar situations in everyday life.

In the second study we examined the consistency of the dogs’ behaviour in the same situation when confronted with two different unfamiliar persons immediately after each other. The consistency of the dogs’ response to the two experimenters was found to be reliably high.

In sum, this test procedure proved to be reliable enough to be a valuable measure of a definite part of the personality characters of dogs.