In dog bite lawsuits, caution is warranted if the breed of the dog is based solely on physical characteristics. In other words, what witnesses or the owner say the dog looks like. In other 
words, based on visual criteria. For example, take the German Shepherd, or a German shepherd mix. There are many dogs that looks similar to the German Shepherd but have different genotypes. Likewise, this holds true for another breed of dog frequently involved in dog bite lawsuits; namely, the pit bull terrier.
In recent years, animal behaviorists have become increasingly aware of the difficulties in making accurate breed identification based solely on visual inspection of a dog. The first detailed, data-based study which examine this issue was published by Victoria Voith. The study of Voith is discussed elsewhere on this website. Dr. Voith found disparity between the breed of a dog owners thought they had and the genotype of the dog based on DNA analysis.
Judith Blackshaw, Ph.D. was one of the first animal behavior scientists to note the difficulties in correctly identifying the breed of the dog based on its physical characteristics. Blackwell published her findings in the Australian Veterinary Journal. The English bull terrier was the breed under investigation. Blackwell developed interest in the problem because members of the Bull terrier association in Australia felt that Bull terriers were unjustifiably being named as the breed of dog involved in a spate of dog attacks in that country. Blackshaw hypothesize that if the dogs involved were in fact purebred bull terriers, then owner should be in possession of the dog’s breeding papers specifying lineage, or have a valid registration for the dog with a kennel club. Her belief was that if verification of breed could not be established through these methods, then uncertainty justifiably remained about the breed of the dog.
Blackwell identifying 52 dogs advertised for sale identified as bull terriers in the classified ads of a Brisbane newspaper. Blackwell contacted the advertisers and discovered that only 23% of the dogs were papered or registered with the kennel club as a Bull terrier. Customarily, owners of purebred dogs, especially an uncommon breed like the bull Terrier, make the effort to register the dog. Moreover, bull terriers are usually the product of a breeder. Random or backyard breeding of bull terriers unlikely happens to the extent that it does in more popular breeds like the Golden retriever.
The upshot of the study: breed identification based on visual inspection of a dog’s physical features is problematic if it is not supported by independent verification through registration with a kennel club, through paperwork from a reputable breeder, or through commercially available DNA test kits. Note, however, currently DNA results using these test kits are limited because they generally do not specify the specific breed of a dog. Rather, results give broad genetic groupings to which the dog may belong. DNA test kits will likely produce results with greater specificity in the future, however. Read more about the problems in making accurate breed identification based solely on the physical features of a dog. 
Also see: Pit bull identification test
