"Behavioral Phenotyping of Golden Retrievers by Means of an Aggression Test"

This study investigated the threshold for aggressive behavior in Golden Retrievers. The researchers conducted behavioral testing on 83 Golden Retrievers and analyzed the results employing three different methods to create a phenotype of behavior. The various methods were then compared to claims made by the dog owners to determine their validity and whether this type of behavioral testing used could be viewed as an effective and objective way to identify dog behavioral phenotypes.

Of the 83 subjects, 59 were classified as aggressive and 24 as non-aggressive according to statements from the owners when interviewed. The behavioral test included 22 subtests; most adapted from a previous study conducted by Netto and Planta (1997). These subtests incorporated situations in which the owner was present or absent, and in which the stimulus was another dog or a person. An ethogram of aggressive dog behavior and one of fearful dog behavior were created and scoring was based off the number of times each dog exhibited each behavior. 

The three methods utilized to analyze this data were a “snap/attack score”, a “total aggression score”, and a principle factor analysis (PFA):

The “snap/attack score” is the frequency of snaps (defined as “mouth opens and closes, possibly accompanied by showing the teeth and/or growling and/or barking associated with a short lunge forward (not maximally) or a quick head movement”) and attacks (defined as “the dog quickly moves forward maximally and makes snapping movements or actually bites…possibly accompanied by showing the teeth and/or growling and/or barking”) in subtests 4 through 21. This method analyzes only the two most aggressive behaviors. 29 Golden Retrievers or 35% demonstrated attacking and snapping. Owner-acknowledged aggressive dogs had much greater scores than owner-acknowledged non-aggressive dogs. This method was found to be a decent aggressive measure for those dogs who had a history of biting; however, there were discrepancies with the owner’s reports and the scores for many of the other owner-acknowledged aggressive dogs.

The “total aggression score” accounts for all behaviors described in the aggressive behavior ethogram. Owner-acknowledged aggressive dogs once again exhibited lower total aggression scores. However, the total aggression score was found to be a better measure for aggression than the snap/attack score because it more accurately represents the owner’s impressions.

Both the aggressive and fearful dog behavior profiles were assessed in the PFA. The PFA the variables or factors that account for the correlations between the elements of behavior. The  dogs can then be classified using the scores for the factors found in the PFA. Six factors were found: threatening, attacking, active fear, startling, support seeking, and uncertainty. Factor scores for factor 1 (threatening) had significant differences between owner acknowledged aggressive and non-aggressive. No significant differences were found for the other five factors. There were two surprising results: that there was no significant difference for factor 2 as that also relies largely upon aggressive behaviors like factor 1 and that for factor 1, there was no significant differences between factor scores for the subtypes of owner-acknowledged aggressive dogs (towards people, towards dogs, or towards both). It was concluded that the PFA has potential, but the results should be treated with caution as it is not entirely reliable.

Their original hypothesis that aggressive Golden Retrievers have a lowered threshold for aggression was proved correct. The total aggression score was deemed the best method for analyzing behavioral test results.

 

Van den Berg, L., Schilder, M., Knol, B., 2003. Behavior Genetics of Canine Aggression: Behavioral Phenotyping of Golden Retrievers by Means of an Aggression Test. Behavior Genetics. 33(5):469-483

http://download.springer.com/static/pdf/870/art%253A10.1023%252FA%253A1025714431089.pdf?originUrl=http%3A%2F%2Flink.springer.com%2Farticle%2F10.1023%2FA%3A1025714431089&token2=exp=1445918437~acl=%2Fstatic%2Fpdf%2F870%2Fart%25253A10.1023%25252FA%25253A1025714431089.pdf%3ForiginUrl%3Dhttp%253A%252F%252Flink.springer.com%252Farticle%252F10.1023%252FA%253A1025714431089*~hmac=1951b25138d94432e03f446a29393d2adfa3a2b4097c835b262b414a891a6203