Here is PDF original source for Author's statistics is referencing [ http://www.start.umd.edu/start/announcements/BackgroundRepor... ]. Of a particular interest is bar graph on the bottom of the second page. Ironically it comes out to about: US citizens are
eight and a half times
less likely to die due to terrorism on US soil (Before and after war on terror). That is just what statistics can do if you take any two numbers and start doing "science" with them. As far as I can tell the statistics author quotes actually compiled from different sources, which makes it even more questionable. I do not doubt the number provided by author is correct, however I just demonstrated that correctness is not enough, you need to treat your data and questions you ask of that data carefully in order to come up with something useful.
Original post below:
_________________________
There is a lie, there is a bold lie, and there is statistics.
These are only a few variables that might contribute to that number:
1. There are actually vastly more cops compared to terrorists.
2. Terrorists do not discriminate targets. Cops usually have a very specific set of people they are targeting (I.E. carrying guns and shooting at other people/gangs).
3. How many of those people killed by cops, were trying to shoot back? Terrorists usually kill people who do not fight back.
I am all up for this kind of statistics, however this one seems a bit sensationalist.
* - a robbery suspect lied about a gun.
- mistook cell phone for a gun.
- pulled out toy submachinegun
a. This adds up to 20, so I missed one. I got no idea which one.
b. Allegedly every mistake killing had at least one mitigating circumstance (And only case of a guy just fleeing and being shot got cop charged with manslaughter).
I did a quick perusal of other months in 2012, and they look similar. According to wikipedia cops killed 587 people in 2012 [1], at the same time 120-127 cops were killed [2]. If pattern established in my quick count holds, then a U.S. cop is more likely to be killed than kill somebody who: does not flee from or point things at cops (especially things that might look like guns).
I am not saying anything. What I did is try to get some numbers together and was astounded by a result. I am not sure how I am making a point that cops are bad.
I think you and I are operating with different definitions of word bad. I took your comment to mean, that the fact that cops are about* as likely to die as to kill someone by mistake, means we got badly trained cops. Now I see there is more than way to read that comment.
* I made up number that was used for that based on a very small sample. Please let me know if I messed up somewhere, I am quite curious about it.
I like this breakdown as it shows the problem with the title of the article "You’re Eight Times More Likely to be Killed by a Police Officer than a Terrorist".
For me at least I'm very unlikely to be attacking cops any time soon...
Was going through the summaries too. Problem is the mitigating circumstances stats are useless. An officer literally has their life on the line when writing their report. They'd have to be just about insane to not say they thought the person had a gun, moved their hands towards their waistbands, made a furtive movement, etc. Maybe the data is good for historical trends but that's it.
True, but the whole point is that the cops are a bigger danger to life and limb (and perhaps liberty) than terrorists. But ask 100 people on the street which is a bigger danger and you'd probably conclude (at least from their answers) terrorism.
That means that society is improperly calibrated to actually reducing mortality and there's no better way to recalibrate than some kind of sensationalist headline that really makes the point.
Two things that people are horrible at reasoning about, probably for biological reasons: Really massive numbers. Really really small numbers.
It only gets worse when you throw in something emotionally charged, like risk. Humans are absolutely awful at evaluating risk rationally.
Consider the standard HN example of robotic cars that are safer by an order or magnitude or two than human-piloted cars. The first time somebody will be killed by a robotic car (and it will happen eventually) the public will be calling for heads on a platter, even though the robotic cars are safer than human-piloted cars that kill hundreds every day.
Personally I cannot wait for my own robot car, so much free time will be given back to me.
However the irrationality that comes into play with a robot car killing a human means that the creators of such cars are going to go above and beyond to do what they can to prevent it. By virtue of human irrationality we get safer robot cars.
And the counterpoint is that they may not actually be the bigger danger, unless you are involved in the commission of a crime. Well. It depends on what you mean by "danger to life and limb". Do you mean, "For the average person?" Or do you mean, "For a particular person?"
The problem with "the average person" is that such a thing does not really exist. So the answer to the question may not be informative. I think most people, when reading this headline, probably think that it is claiming they are more likely to die from the police than terrorism. Unless you are a criminal, that may not be true (depending on P(criminal|killed by cop)), and it's not possible to estimate that based on data in the article, if it's the same article I've seen before.
True, but the whole point is that the cops are a bigger danger to life and limb (and perhaps liberty) than terrorists.
Only if you look at it simplistically. Terrorists on the streets don't really stop other crime, whereas cops do. Some drunks having a fight outside a bar at closing time? Having terrorists around won't help.
Also note that the cops are in much more danger of life and limb than the rest of us. That is, the cops get to confront the creeps so that most of us don't have to most of the time. Some of those creeps are armed, violent, and not inclined to cooperate with the cops. This leads to cops often being in situations where, if they're a half-second late in shooting, they're dead. If they're a half-second early, they're villified, suspended and charged with manslaughter.
We're asking them to do the impossible, and they're coming close to doing it. I get the bit about trying to recalibrate the public's perception of risk, but don't make them freak out about the cops to make your point.
If you go by deaths (admittedly, a terrible metric since cops wear body armor and typically can get on the radio for an ambulance pretty quickly), then cops are not in more danger of life and limb than the rest of us. There are plenty of jobs that you are more likely to be killed, even murdered, in.
Regardless, you are missing the point of the comparison. The comparison should not be taken to imply "You should fear the police". That would be an irrational takeaway from a comparison between police and terrorists. What you are suppose to take away from the comparison is "Gee, my fear of terrorists is pretty damn irrational, since plenty of other things are more likely to kill me, but I don't particularly fear any of them."
The fact that so many people get hung up on this comparison demonstrates the comparison's relative uselessness in practice. The comparison has too much emotional baggage to be properly digested without careful spoon-feeding.
The fact that I'm personally extremely unlikely to be killed by either terrorists or cops isn't the reason the statistic is important.
The statistic is important for public policy debate because there is 8 times more progress to be made in terms of saving lives by dealing with the problem of police killings.
It's interesting, but it isn't the only needed info. You might have 8 times as much opportunity for progress, but it might be 8 times harder to make progress.
eight and a half times
less likely to die due to terrorism on US soil (Before and after war on terror). That is just what statistics can do if you take any two numbers and start doing "science" with them. As far as I can tell the statistics author quotes actually compiled from different sources, which makes it even more questionable. I do not doubt the number provided by author is correct, however I just demonstrated that correctness is not enough, you need to treat your data and questions you ask of that data carefully in order to come up with something useful.
Original post below:
_________________________
There is a lie, there is a bold lie, and there is statistics.
These are only a few variables that might contribute to that number:
1. There are actually vastly more cops compared to terrorists.
2. Terrorists do not discriminate targets. Cops usually have a very specific set of people they are targeting (I.E. carrying guns and shooting at other people/gangs).
3. How many of those people killed by cops, were trying to shoot back? Terrorists usually kill people who do not fight back.
I am all up for this kind of statistics, however this one seems a bit sensationalist.