Aggression in the domestic dog – Part 1

Materials and methods


Unfortunately biting incidents with dogs are a problem in various countries, not only in The Netherlands (Hanna and Selby, 1981; Morton, 1973; Beck et al., 1975; Sacks et al., 1989). In The Netherlands in recent years, aggression in dogs has become a considerable problem. Results of a study by Mulder (1991) indicate that annually 50000 people have to be treated in hospital for bite wounds caused by dogs. A number of very severe incidents caused primarily by Pit Bull Terriers prompted the Ministry of Agriculture, Conservation and Fisheries to set up an Advisory Committee concerning Aggression in Dogs. The committee made several recommendations, one of which was to develop a behavioural test for aggression in dogs. We were asked to develop and validate such a test. The primary goal of the test at the moment is to select highly aggressive individuals in three “potentially aggressive breeds” indicated by the Ministry: the fila Brasileiro, the dogo Argentino and the American Staffordshire Terrier. To prevent dogs of these breeds from being used as substitutes for pit bulls, breeding has to be controlled. Measures are being prepared which will prevent the breeding of dogs that have failed an aggression test. Selection in the breeding program is a means of decreasing aggressive behaviour. Aggressive behaviour is part of the normal behaviour of dogs. However, the level of aggression in individuals may be so high that it is unacceptable to the direct environment or to society in general. As in all types of behaviour, both genetic and environmental factors play a role. Important for the goal of this study is the fact that aggression has a genetic component. Selection for guard and police dog functions has resulted in breeds with a lower threshold for aggressive behaviour (Beaver, 1981). Other authors report that biting incidents are more frequent in certain breeds (e.g. van Gorp et al., 1990). Males are more often involved in biting incidents than are females (e.g. Beck et al., 1975; Borchelt, 1983; Wright, 1991). Clearly there is a genetic basis for aggressive behaviours with an originally biological function. In this connection, authors mention various classes of aggression (e.g., Borchelt, 1983; Houpt, 1983; Polsky, 1984). The contexts in which aggressive behaviour is observed are the basis for the development of sub-tests for aggression. The literature provides some information about behavioural testing of aggression in dogs (Seiferle and Leonhardt, 1984). The information on tests available in the literature, experiences with tests used by dog clubs and our own research (van der Borg et al., 1991 and unpublished internal reports, 1989, 1990, 1992) were used to develop sub-tests. Additional sub-tests were developed, with a variety of stimuli, presented in contexts in which aggressive problem behaviour is likely to occur (van der Wijk and Klasen, 1981; Wright, 1985; Winkler, 1977; MOSS and Wright, 1987; Sacks et al., 1989; van Gorp et al., 1990; Podberscek and Blackshaw, 1990). Non-threatening sub-tests were added so that the test would be more acceptable to the owner. The purpose of our study is to develop a behavioural test that has been scientifically validated and that can be used in dog clubs by laypersons as an instrument for instituting selective breeding. The test should not only be applicable to the breeds indicated by the Ministry, it should also have more general applicability so that it can be used to test levels of aggressiveness in dogs that are a danger to society. In this paper we describe the set of sub-tests we finally used. Furthermore we investigate the reliability and the validity of the test as a whole.

Over 80 sub-tests were designed in two pilot studies. In these two pilot studies 171 dogs were tested with two different test configurations. On the basis of the results we made a final selection of sub-tests and some new sub-tests were developed. The final test is described in this paper.

All dogs (n = 112: 59 males and 53 females) in the final test were owned by private persons. Among the dogs were 75 dogs belonging to the three potentially aggressive breeds (PAB: 18 fila Brasileiros; 26 dogo Argentinos; 3 1 American Staffordshire Terriers). Of these dogs, 60% had bitten before they participated in the test (referred to hereafter as dogs with “biting history”). 37 dogs (78% with biting history) of different breeds and some mongrels were used as “controls” (CG). Some of these dogs (23 “aggressive controls”) belonged to breeds that had some aggressive characteristics in the breed properties (e.g. Rottweiler, Dobennann Pinscher). A second control group of 9 dogs consisted of “non-aggressive breeds” (e.g. Labrador Retriever, Golden Retriever, German Pointer). A group of 37 dogs was re-tested to establish the reliability of the test. Before the test we made use of an extensive questionnaire to gather information about dog and owner (details will be reported elsewhere) and in particular about the aggressive history of the dog.