They say numbers don't lie. That's not necessarily true. Statistics are numbers subject to interpretation, influenced by the biases of the people who compile them, and skewed by the goals of the people who analyze them. There are multiple examples of how the biases against police officers held by the people calculating recent police statistics affected their analysis of the numbers. Let's look at two.

At the end of last year, the results of a law enforcement body-worn camera study were released. In that study, 1,112 officers of the Metropolitan Police Department of Washington, DC, were given body cameras and the same number were not. Then the researchers compiled data for a set period of time and reported their results.

They didn't like their findings. Neither did the mainstream media. What they discovered is that the officers wearing the cameras and the officers without cameras showed no statistical difference in use-of-force incidents or civilian complaints. "These results suggest that we should recalibrate our expectations of BWCs' ability to induce large-scale behavioral changes in policing," the researchers wrote.

Talk about bias. The researchers went into their study thinking it would show that the body cameras would change police behavior and decided their results weren't significant because they didn't change police behavior at all. But the results were very significant. What they showed is that the deciding factor in police behavior toward "civilians" is the behavior of the civilians themselves.

It's important to note here that previous studies of the effects of body-worn cameras in policing have shown that the cameras have a substantial effect. A study on body camera use by the Las Vegas Metropolitan Police Department reported 37% fewer use-of-force incidents and 30% fewer complaints against officers.

So does this mean the Vegas officers were unnecessarily using force before they wore the cameras in contrast to their DC counterparts? Absolutely not. The one variable these studies never consider is the public. I would dare say the Vegas officers have to deal with many more drunken belligerents than the officers in DC. And maybe, just maybe, the sight of that camera now on the Vegas officers makes those people think twice before acting the fool.

Example number two of bias in interpreting statistics is much more heinous. It involves the number of officers killed in the line of duty last year.

Every December the National Law Enforcement Officers Memorial Foundation sends out a press release that reports the number of officers killed in the line of duty for that year. The total of officers killed by gunfire in 2017 was 44. That's a 33% decrease when compared to 2016 when 66 officers were shot and killed.

The mainstream media picked up that story. And some reported it as just a straight numbers story, good news that fewer officers were being feloniously killed. But some in the press made the specious comparison of the 44 officers killed to the 987 people the Washington Post says were shot and killed by police officers last year and essentially implied that the dangers faced by law enforcement professionals are exaggerated.

They brought their biases to the numbers and didn't dig deeper. Here's what they would have found if they had looked at the numbers objectively. The National Fraternal Order of Police reports that 271 officers were shot in 2017, including the 44 who were shot and killed. Many of those officers were saved from fatal wounds by body armor. When you substitute 271 for 44 in the ratio of how many people were killed by officers, you get a fairer comparison.

And if you want to be even more accurate, then we need a number that isn't available: how many officers were targeted by gunfire. No one keeps that number. But it's a lot higher than 271. A young officer who works a particularly dangerous part of a major city recently told me he had taken rifle fire 20 times in his career.

Thankfully, that young officer hasn't been hit. Which means you won't find his experience recorded in the statistics. And until we have such numbers accurately represented in the assaults on officers statistics and until researchers take into consideration the role civilians play in officer use-of-force incidents, any analysis of the numbers used to condemn police for the way they carry about their duties is at best incomplete and at worst a damned lie.