National Canine Research Council Summary and Analysis:
Sacks, Sinclair, Gilchrist, Golab, and Lockwood (2000) examined 238 DBRFs over a 20-year period (1979-1998) in the United States to determine breed involvement and assess relevant policy. Ninety cases (38%) in which no breed was reported were excluded. They concluded that fatal attacks were too rare to drive public policy regarding dangerous dogs and recommended against breed specific ordinances as an approach to dog bite prevention. Although the study predates the documentation of the unreliability of visual breed identification by 2 decades, the authors note the difficulty of accurate breed identification.
They note other weaknesses in this attempt to categorize dogs involved in DBRFs by breed: 1) there may be a sampling bias as only a percentage of the dog bite-related fatalities are included in the report; 2) specific breeds may be more media-worthy and thus more likely to be identified in media searches of dog bite fatalities; 3) assignment of dog breed is not reliable and after a dog bite incident a subject may be more likely to be identified as a stereotypically aggressive breed; 4) cross-breed dogs are difficult to account for and depending on the approach taken a given breed may be over or under represented; 5) reliable breed-specific population data does not currently exist because only a portion of pet dogs are registered and licensed and there is likely a breed bias for those that are registered; and 6) even if accurate dog bite breed ratios could be calculated they will not account for owner-related factors such as neglect, training, or encouraging aggressive behavior.
Surprisingly, despite their recognition of the above weaknesses, Sacks et al. still attempt to make breed-specific claims regarding DBRFs, specifically, Rottweilers and “pit bull-type dogs.” They do not define “pit bull-type.” For all of the reasons rightly enumerated above by the authors themselves, however, these conclusions are not warranted.