Over the past several months multiple news outlets have quoted the Puppycide Database Project as "estimating" US police shoot 500 dogs a day.

So far we are aware of one major news outlet making this claim, and several dozen smaller blogs and magazines made the claim go somewhat viral in July. We have requested corrections from three of these news sites; two of them have simply ignored us, the third outlet we just found and contacted today.

The attribution to us is based on a mis-reading of an interview with the online news magazine DailyDot. Here is what they published:

“We are nowhere close to having complete dog killing records for all 18,000 police departments,” Wieder told the Daily Dot. Puppycide DB looked at the dog-killing records of just 40 police departments. The data showed a mean killing rate of 10 dogs per department. Based on just that tiny sample, Wieder said, that rate would add up to 500 dogs a day nationwide.

“Do we think the numbers are that high? We aren't sure,” Wieder said. “The point is we don't have enough evidence to make authoritative statements about national rates, and neither does the Department of Justice.”

Reading both of those paragraphs, it is obvious the point we were making is that no one currently has enough data to make any sort of authoritative claim about the national rate at which US police shoot dogs. When we spoke to the Daily Dot, we repeatedly stated that the sample that produced the figure was much too small to determine anything from. At the time, the sample was 40 police departments. There are between 12 and 18,000 law enforcement offices in the US depending on what you consider a law enforcement office. So the sample was 1/3rd of 1% of the total law enforcement offices.

Our Daily Dot interview was given at the same time we began publishing a lengthy four-part series of articles that investigated and debunked misleading and dishonest statistical claims like this.

Here is an example of what our quote has been turned into:

The Puppycide Database Project estimates the number of dogs being killed by police to be closer to 500 dogs a day (which translates to 182,000 dogs a year).

Mistakes happen. Certainly, we have made mistakes from time to time. When that happens the Puppycide Database Project's policy varies based on where the mistake took place (blog or social media). A mistake in a blog post warrants a new blog post linking to the original describing the error, and an update to the original describing the error. We leave the error and the correction in place. We then send out Tweets & Facebook posts so that people who already read the article are aware of what happened. So far we have had to do this twice: once when we mistook the Boston Police Department for the Baltimore Police Department and another time when we repeated claims made by several major media outlets about proposed changes to the Department of Justice's Asset Forfeiture Program that turned out to be overblown and inaccurate.

Unfortunately we are almost alone in this approach to corrections, and some of the most trusted news sources in America have proved to be incredibly cavalier about updating mistakes - particularly to online content. Often mistakes are simply deleted without informing readers. Sometimes entire articles are removed.

Its our hope that the news outlets involved with this error have a higher ethical standard. Perhaps our attempts to contact two of the three sources making this claim that have gone unanswered were simply "lost in the mail".

There is no fault in making a mistake. However, leaving this mistake online does our project and the work of our volunteers an extreme disservice. Puppycide Database Project was founded in the hopes of providing journalists with access to accurate information about police shootings that simply did not exist before we got started. We remain the only searchable database of police shootings of animals, despite the enormous amount of public attention directed towards these incidents. Bringing transparency to police violence allows the public and policy-makers to make informed decisions about these incidents. Our project is founded in the belief that providing people with access to the truth is our best hope for making this problem better. Publishing untruths undoes this work. It means that we have to spend that time repairing the damage. It means that people will stop trusting us when they see obviously false statements attributed to us.

So - if any of the writers or editors involved in this are reading this - please publish a correction or please contact us either on our website, Facebook or Twitter so that we can discuss this.