Small Pieces of Fact Can Misinform

I was thinking about this on my walk this morning. I met someone who owns a mastiff. He told me about how people are afraid of the dog even though it only ever tries to love people. This made me think of the stereotype around pit bulls, and how people walk out into the road to avoid walking past Kelly when she’s walking one of our pit fosters. I couldn’t think of why people would think of him as threatening. He has a gimp leg, and one look at his face will tell you that he doesn’t have a brain in his head. I thought about this while I was finishing my walk, and realized it’s groups that take small bits of incomplete information and pass it off as fact. Take for instance. They publish a dog bite study that they pass off as fact even though I find it to discredit itself in its own introduction. I only mention because it is the group I was thinking of this morning. This can be found other places especially in politics.

Dog attack deaths and maimings, U.S. & Canada September 1982 to June 25, 2010

Compiled by the editor of ANIMAL PEOPLE from press accounts since 1982, this table covers only attacks by dogs of clearly identified breed type or ancestry, as designated by animal control officers or others with evident expertise, who have been kept as pets.

First off I’ve never heard of any valid study that is taken as a complete truth based on media reports. They also say that it only covers dogs of clearly identified breed type or ancestry determined by animal control officers or others with evident expertise. Yet these experts failed to identify some of the breeds which means they could be wrong about the others. It also means that the breed determinations weren’t based on DNA testing. This for me is a large red flag. I personally have not met a animal control officer, shelter worker, rescue worker/volunteer, dog trainer, or any other person that claims they can identify every breed in a dog. The words I hear most associated with a dogs breed are ” Our best guess is…”

Due to the exclusion of dogs whose breed type may be uncertain, this is not a complete list of fatal and otherwise serious dog  attacks; but there have been very few qualifying attacks by dogs of uncertain ancestry in recent decades.

First off they admit to not including all dogs in the study. Even if there are few qualifying attacks not including these attacks makes the data set incomplete, and less valid.

The ‘%/dogs” column states the percentage of each breed of dog among 3.2 million classified ads listing dogs for sale at web sites during the first half of 2010. Similar data has been collected in many previous years, but has not previously been included in this table. If a percentage is not listed for a breed or mix, it either appears to be too low to calculate or too difficult to isolate from other variants of the breed or mix.

I also can’t understand the idea of basing breed percentages on classified listings; because these dogs aren’t in their “permanent” homes which means they may not be an accurate representation of dogs actually in people’s homes. Would you base population statistics on the members of dating websites? They’re also based on ads only for the first half of 2010 which means that dog bites from 1982 are based on unclear population percentages from the first half of 2010. If this is the case how do they control for popularity of breeds over the years before 2010? It also doesn’t include mixes that they couldn’t identify thereby making the data set even more incomplete.

There is a persistent allegation by pit bull terrier advocates that pit bulls are overrepresented among reported dog attack deaths and maimings because of misidentifications or because “pit bull” is, according to them, a generic term covering several similar types of dog. However, the frequency of pit bull attacks among these worst-in-10,000 cases is so disproportionate that even if half of the attacks in the pit bull category were misattributed, or even if the pit bull category was split three ways, attacks by pit bulls and their closest relatives would still outnumber attacks by any other breed.

The report and their website both show an obvious bias toward pit bull. This bias is also apparent in the fact that they include the breed “Pit mix unknown” which by their standard shouldn’t have been included. I have yet to see a mix that doesn’t have some dominant traits of one breed or another. So the inclusion of this even though it’s only 4 attacks skews the data even further.

Another issue I noticed while writing this is that the “dogs X victims” column and the “individuals” columns don’t always match. Take the pit mix unknown breed that I mentioned before the columns are Number of Victims: 4  Children: 2  Adults: 1 Deaths: 0 which doesn’t add up. You can’t add 2+1 and get 4.