Gun Statistics

Gun Statistics

Guns don't kill people, and statistics don't lie.

I always tell my students: statistics don’t lie, just like guns don’t kill people.

It is a grave responsibility, using one, but too often, they’re found in irresponsible hands – waved around as though it imparts authority, used as ‘protection.’ Protection from what? I don’t know. Protection from listening? Protection from intrusion? He who holds the weapon speaks and makes the rules.

But it’s a dangerous weapon that deserves respect.

Too often, people use these weapons without taking the time to learn how. They never learn the intricacies of how it works, how to care for it, and when it’s appropriate to use it. They never learn to be responsible with it.

As a result, people use it as a kludge to prove their own righteousness, tossing it around casually, thinking they know more than they know, and have more power than they actually have. Truly, using one can make you feel powerful. It gives you the illusion of authority.

And people get hurt.

While such a weapon in the hands of a skilled professional can be extremely effective, misuse of the same will often achieve a most undesirable effect. And when two people, equally unskilled and unknowledgeable, start using such a weapon against one another, the results can be quite chaotic.

Here’s one example. #BlackLivesMatter activists often point out that 40% of police fatalities are black men (source: the internet) while only 6% of the nation fits this demographic (source: 2010 US Census). Clearly, proof that police are racist.

This particular weapon gets waved around from time to time by activists, and as with any unskilled, irresponsible use of a weapon, it backfires. Despite the illusion of authority this gun imparts, it’s a banner for your opponent, or anyone who’s listening, to discredit you. It shows that you a) don’t have a basic understanding of statistical knowledge, and therefore b) don’t have any credibility to anything you say.

If you want a weapon you can use, it helps to know a little about what you’re doing.

It doesn’t help to make blanket protests against general themes such as police racism and police brutality, we can get surgical, and determine the type and extent of racial profiling within a given jurisdiction. And that is useful, because it helps us to be able to target the more problematic areas. In order to do that, you need to know:

  • The race/gender demographics within a given police jurisdiction. This you can get easily.
  • The total number of police interactions initiated by the officer. This you can get from the police department. You may have to get all the police reports, and spend some time doing a little sifting.
  • The total number of (the same) in which the suspect was black. While in reality, race is an ambiguous construct, in this case, what matters is the officer’s perception. So we use the way it’s identified on the report.

The first one is simple. If you want to compare numbers in a group, you have to stay within that group. One of the biggest fallacies of the movement is to go after all police, when some are the problem – after all, how is that different from racist cops who target all black men when some are the problem? Say, for example, you want to see if the cops in Coral Springs are racist, you wouldn’t waste your time looking at Oakland Park. And it wouldn’t do to say “the cops in Oakland Park are racist because they only ever arrest black people,” if the only people in Oakland Park are black (this is hypothetical).

The second one is a little less obvious. You will need to get the reports from interactions in which the police officer initiated the interaction. Say, for example, you’re looking at Coral Springs, and you see black people make up 20% of the city, but 40% of the suspects in police reports. This may seem like it tells a story, but there could be a whole lot of confounding variables at play. For example, the neighborhood itself might be racist, and the police are only responding to 911 calls – which we WANT them to do. If you want to see if the police are racist, you will need to go through the reports (should be public information) and dig through them.

What we’re looking to do is compare the percentage of people in the jurisdiction who identify as black, to the percentage of police interactions initiated by the officer in which the suspect was identified as black. In order to do this, we’re going to use what’s called a two-proportion z-test. For this, you actually don’t need all of the reports; a random sample of adequate size should suffice. Here are the variables you will need to identify:

  • p1 is the proportion of black people in the police district
  • p2 is the proportion of police initiated interactions with black suspects in the report sample
  • n1 is the total number of people in the police district
  • n2 is the total number of police initiated interactions in the report sample

Once you have identified these numbers, you can plug them into this formula:
(p1 n1 + p2n2)/(n1 + n2)
In other words, take each proportion, and multiply it times the number of people that proportion is from, then add those together, and divide by the total number of both sets. This number, p, is called the combined sample proportion

By now, most of you will have gotten bored stopped reading, just like you did back in college. I suppose it’s easier to regurgitate numbers than it is to learn the truth.

Be social, please share!


Leave a Reply

Your email address will not be published. Required fields are marked *