On the subject of statistics...

The big problem with stats is the average person does not understand them at all. Everyone thinks they understand them just fine, but unless you have received some form of training or instruction in them, your comprehension of them is vague at best.

This is where the concept of massaging statistics in the general view comes into play. The statistics (numbers) themselves cannot be manipulated unless you intentionally falsify or manipulate the data itself. This means that if the data is valid then the statistic generated (assuming no calculation errors which is very unusual as its all done by computer now) is also valid.

The problems come when statistics are presented to the public. They don't know how to interpret statistical results and typically will pay more attention to the material behind the statistic (what is said in relation to the statistic) then the statistic itself. This is of course counter productive as its the statistic and how you interpret it that matters, not how the guy presenting it interprets it. This is made worse by the habit of most people presenting made up statistics (like the probability of a bad guy getting shot comment someone made earlier).

The nice thing about statistics, if you really understand them, is that you can often tell when manipulation is going on (such as low sample sizes, odd population sampling, material presented as being statistically significant when its not, etc). I for example see tons of bad stats on display every time I turn on the TV. People make careers out of presenting any given stats in the most favorable light possible.

Anyhow as for this study, the stats (and therefore the results) themselves are pretty solid. The study was done at a reputable university, it was peer reviewed and published in one of the larger journals. The data was definitely gathered totally separate from the research team, most likely by the city or police force. It is very unlikely that the team massaged any of the stats (that would be career ending if caught, and not to hard to catch). They also didn't, as Skybird said, try to generalize the results beyond the scope of the research. Further these stats probably do generalize fairly well, as they suggest, to other locations in the US with similar environments as to where the stats were gathered (though further research in other cities would have to be done to verify that hypothesis).

Furthermore the statistics are statistically significant (lower than 0.05 chance or P<0.05 which is the norm in psychology), the sample size used is easily large enough to be considered representative of the population, and their conclusion is based entirely off the results. I therefore don't see any reason to doubt the stats, or the results.

I don't disagree with the results.

In fact - common sense SUPPORTS the results AS GIVEN.

What is laughable is that to use a NARROW scope result - that of the chances of being shot if your armed and are victimized by an armed criminal - as some sort of proof that guns in the hands of law abiding citizens is a negative thing.

The study itself is solid. I have no issue with it. However, it is but one very small facet of the total equation. That equation is more than just your chance of being shot in such a situation. The total question must account for:

#1 - Number of crimes not committed due to the intended victim being armed. (A totally UNQUANTIFIABLE piece of data)
#2 - Number of crimes committed by a criminal against an armed individual while said criminal was either unarmed, or armed with a weapon OTHER than a firearm. (Not taken into account in this study)
#3 - Number of crimes in #2 that resulted in the victim being injured by the criminal.
#4 - The number of attempted crimes that were thwarted due to the intended victim using necessary force to stop it.

And even this is not all the data needed. The conclusions of the study themselves are perfectly fine. They are right - if you go armed, you better be prepared to deal with the level of responsibility that requires. You should also insure you have the necessary skills to react properly should a situation demand it. Not a single thing wrong with those conclusions. Its also correct to state that introducing another firearm in such a situation increases the danger tremendously for most people. This is because MOST civilians still try to difuse the situation with the threat of "payback" violence. Any threat at that point increases the danger. Any logical person will realize that. The problem is not in going armed, it is that most civilians react by pulling the gun first, without already having made the decision to use it. This creates the classic "standoff" - which virtually assures someone is going to take lead. Not to mention most civilians don't have the training they should in how to make the needed decision, and how to carry it out. Everything in that conclusion is reasonable.

However - the comment of "Such users should reconsider their possession of guns..." is what sets most people off. The context is talking about personal carry at all times - but is easily construed as possession in total. Therein lies why people will react. Some would say that the study is further proof that the average person should not be allowed to "POSSESS" a firearm, when that is not the conclusion in context.

I am not saying Skybird was suggesting this as a "bad guns" stance, but the current climate is one where people react without clear thought on certain issues - and gun control is one of them.

Gun Control = One shot, One kill.

That isn't unquantifiable!
Just take a large sample of people with and without guns and see how
often each group is a victim of violent crime after adjusting for confoundments.

http://www.amazon.com/How-Lie-Statis.../dp/0393310728

:D

It would be pretty difficult to gain solid data on such things and there would be many possible confounds that would be hard to control against.

Number of people shot and if they were armed is easier to gather data on. First off gunshot victims either go to the hospital or the morgue in most cases, so you don't have to deal with the problem of crimes going unreported. Second as for unreported crime, this will be far more common for failed criminal attempts then successful ones, and there will also probably be a difference in failing to report a crime if the criminal is armed with a gun, a knife, a blunt object, or nothing. Shootings also bring police involvement which means records and police reports which can be handy for sorting out data, this may not exist if an attack does not occur.

Also how would you determine which group experiences more crime other then by asking each member of the group questions about it? As by doing so you will almost certainly create even more confounds (there is a good reason why most research tries to avoid involving the participants directly in the research as much as they can).


@Platypus,
Cute book and more or less right from the blurb on amazon about it. One should always look at the numbers, particularly the numbers behind the statistics or graph. If you understand the numbers you can see what is really going on, no mater how they may try to obfuscate it with trickery like playing with the scales in a graph, or other forms of distortion.

I use that book a lot at work. Statistics, when given to the uninformed, can be a powerful manipulation tool.

I cringe when ever a politician shows statistics. :damn:

A, nightmare literature. :D

For dreamers on our side of the border, this:

http://www.amazon.de/Statistik-Human...4900291&sr=8-1

I still get a stomach ache when remembering it. :D