To best understand this article in the context of the Breeds and Behavior literature, please see National Canine Research Council's complete analysis here.

Fadel, F. R., Driscoll, P., Pilot, M., Wright, H., Zulch, H., & Mills, D. (2016). Differences in trait impulsivity indicate diversification of dog breeds into working and show lines. Scientific Reports6(22162), 1-10. doi: 10.1038/srep22162

National Canine Research Council Summary and Analysis:

Fadel, Driscoll, Pilot, Wright, Zulch, and Mills (2016) studied impulsivity both between and within two breeds (Border Collies and Labrador Retrievers). It is included here because it investigates presumed selection for a complex personality trait (impulsivity) among working lines of 2 breeds (those lines presumed to be selected for the ability to perform the breed traditional tasks), but also because it provides an example of the difficulties in defining a phenotype for large personality traits. The authors hypothesized that the expression of the broad personality trait or group of behaviors collected under the label “impulsivity” would differ more between the working lines of the 2 breeds than between the show lines of the same breeds which are selected, presumably, primarily for appearance. They selected two popular breeds that have both working and show lines to compare and contrast—breeds whose working functions, the authors judged, call for contrasting levels of impulsivity. Fadel et al. (2016) reported small but statistically significant differences in impulsivity between the working lines of the 2 breeds but not between the show lines. Although, as others have found, there were greater differences within breeds as a whole than between them.

Fadel et al. (2016) employed the relatively new Dog Impulsivity Assessment Scale (DIAS) owner survey* to determine impulsivity levels in their sample of 1,161 dogs in the U.K. Each owner identified his or her dog as a purebred and as a member of 1 of 4 groups: Labrador Retriever from a working line; Labrador Retriever from a show line; Border Collie from a working line; or Border Collie from a show line. No documentation of these identifications was required. The majority of owners (75.73%) reported having acquired their dogs from breeders, which makes the claim for purebred status somewhat more credible among these dogs, although there is no information about how this method of acquisitions was distributed among the various lines. The lack of documentation of the type of line is more problematic as there is no information as to how the owners made this identification. Finally, there is no information about the proportion of dogs in any of this lines that were actually engaged in either the breed’s traditional work or in the behaviors expressed in the show ring, so there is no basis for even speculation as to the role of learning in the small behavioral differences found here.

DIAS consists of 18 items that can be grouped in several ways. In accordance with the original study, Fadel et al. (2016) used a 3-factor grouping to explain impulsivity, which included 1) behavioral regulation, 2) aggression threshold and response to novelty, and 3) responsiveness. They note that DIAS has been validated** and shown to correlate with behavior on a delayed-reward test (a measure of impulsivity).

Small significant differences were reported both between and within the two breeds and lines for impulsivity. When only show and working lines were included, the two breeds differed for impulsivity, behavior regulation, and aggression threshold and response to novelty. Within show lines, the two breeds differed for aggression threshold and response to novelty, while working lines showed breed differences for impulsivity, behavior regulation, and aggression threshold and response to novelty.

The results of this study are a good example of the difference between statistical significance and practical significance. Because of the very large sample size, statistical differences were detected, yet the effect sizes were very small, and the authors acknowledged that the differences could be a result of the sample size.

Overall, the authors’ reported their hypothesis as supported; working lines of different breeds were reported by their owners to have differences in impulsivity and show lines did not. It should also be noted that the sample could not account for the potential confound that dogs being used for working tasks are likely to have profoundly different and very task-specific husbandry and training – experiences that might account for some or all the differences found here. Similar to other studies based on owner surveys, potential owner biases are introduced. Owners of dogs believed to be of specific purebred lines may perceive their dog’s behavior through a lens of expectations, making it hard to separate out differences in lines or breeds from differences in preconceived notions about those lines or breeds. Additional caveats common to such an approach: the scant examination underlying claims of validation for owner survey canine behavior assessment tools (in this case, the DIAS) and the unverified owner reports of their dogs’ breeding background.

Abstract and Link to Full Text of the Original Article:


* On closer examination, however, that validation consisted primarily of one test retest reliability protocol and one small construct validity experiment comparing DIAS results with those on a delay discounting test (how long the dog will wait for 3 pieces of kibble when he can get one piece immediately; Wright et al., 2011 & 2012). On the latter procedure, however, the researchers do not appear to have adjusted the inter trial interval (making sure the dog can’t get more kibble by simply staying with the immediate delivery device). This adjustment is critical to a meaningful result on a delayed reinforcement test (Odum, 2011).

In any case, these two studies represent only a very small segment of the accepted range of reliability and validity measures expected when evaluating human psychometric instruments. A list of just the basic ones would include face validity, test-retest reliability, construct validity, content validity, criterion validity, etc., and nothing like this scope of examination has been done. Moreover, the trait of interest, impulsivity, while it has been the subject of some limited study in dogs, is much like aggression, in that there is no agreed upon definition in the canine behavior literature. It is unsurprising then that according to the authors of the DIAS, the items suggested by the panel of 32 animal behavior experts whose responses were used to compose the questions for the survey “revealed high variability in suggested items so an attempt to reach a consensus between the experts was not made” (Wright et al., 2011).

** As noted above, however, the delayed-reward test used in this validation might actually lead to paradoxical findings, given the reinforcement that the dogs experience during the test.

Additional References:

Wright, H. F., Mills, D. S., & Pollux, P. M. (2011). Development and Validation of a Psychometric Tool for Assessing Impulsivity in the Domestic Dog (Canis familiaris). International Journal of Comparative Psychology, 24(2).

Wright, H. F., Mills, D. S., & Pollux, P. M. (2012). Behavioural and physiological correlates of impulsivity in the domestic dog (Canis familiaris). Physiology & behavior, 105(3), 676-682.

Odum, A. L. (2011). Delay discounting: I'm a k, you're a k. Journal of the experimental analysis of behavior, 96(3), 427-439.

Page last updated July 22, 2019