Beagle-zebub
Member
Fairly strong correlations there, I say I say.
My never-before-seen graphs relating violent crime rate and gun owning % by state!
- Thread starter jlbraun
- Start date
- Status
FieroCDSP
Member
Interesting charts. I think it primarily shows that there isn't any way to group it all together (pro vs anti gun) on one (or even three) graphs. There are just too many variables. Legal and illegal ownership, local, state, federal laws, age of victims, age of assailants, population density...it all factors into the real scheme, and would require years of intensive study and work to even scratch the surface of an intelligible answer.
They say the human mind can do calculus problems simply through experience. Example is throwing a football to a running reciever. Lots of variables, lots of unknowns, but it can be calculated instinctually if not numericly, in a short amount of time with practice.
I think the gun control/crime rate issue is such. You just look at the fact that unarmed people are vulnerable, the fact that criminals prefer the vulnerable, and that leads simply, despite the other variables, into criminals dislike less-vulnerable victims. Less vulnerable is easiest achieved through arming the victims.
But I'm no math major
They say the human mind can do calculus problems simply through experience. Example is throwing a football to a running reciever. Lots of variables, lots of unknowns, but it can be calculated instinctually if not numericly, in a short amount of time with practice.
I think the gun control/crime rate issue is such. You just look at the fact that unarmed people are vulnerable, the fact that criminals prefer the vulnerable, and that leads simply, despite the other variables, into criminals dislike less-vulnerable victims. Less vulnerable is easiest achieved through arming the victims.
But I'm no math major
Excellent work!
What I see is no strong correlation between firearms and violent crime, which is what the CDC study showed.
What I see is no strong correlation between firearms and violent crime, which is what the CDC study showed.
ArfinGreebly
Moderator Emeritus
Correlation
One thing apparent from this snapshot is that, just using raw numbers, ignoring the local dynamics of any given segment, one cannot correlate guns to crime -- either way.
For every example of "see, more guns is more crime" there is a balancing "see, more guns is less crime" example.
More informative (and much harder to do) would be a graph of trends.
Take a given state, and run the numbers for a sequence of ten years or twenty years. You'll get a trend for that state. Take another state and do the same thing. If the preponderance of trends shows a decrease in violent crime as guns increase for several states then you have a homogeneous enough dataset that you can derive meaning from it.
Population density is clearly a meaningful factor.
Also important is the legal/enforcement climate. Giuliani takes credit for reducing "gun crime" in New York by virtue of "gun control" but in fact he created an enforcement climate where jaywalking could get you a busted skull. When looking cross-eyed at a cop will get you a "nightstick tan" then a lot of "out in the open" crime will dry up. It wasn't gun control, it was punk control.
Because these state-specific influences will bend the otherwise "pure" numbers, you kind of have to build your trends by state.
So, for Florida, when shall-issue CCW was introduced, what were the trends?
For Illinois (or is it Wisconsin?), where only the king and his entourage may carry, what are the trends?
Another meaningful comparison might be side-by-side with states having similar gun laws and carry restrictions. That is, can one fairly compare Arizona with Texas, given that one allows open carry and the other does not? Arizona along with, say, Nevada and Idaho might be more meaningful. And with what other state could you compare California? Or do you have to go to Europe for that?
That would be an interesting graph set, but surely a lot of work.
One thing apparent from this snapshot is that, just using raw numbers, ignoring the local dynamics of any given segment, one cannot correlate guns to crime -- either way.
For every example of "see, more guns is more crime" there is a balancing "see, more guns is less crime" example.
More informative (and much harder to do) would be a graph of trends.
Take a given state, and run the numbers for a sequence of ten years or twenty years. You'll get a trend for that state. Take another state and do the same thing. If the preponderance of trends shows a decrease in violent crime as guns increase for several states then you have a homogeneous enough dataset that you can derive meaning from it.
Population density is clearly a meaningful factor.
Also important is the legal/enforcement climate. Giuliani takes credit for reducing "gun crime" in New York by virtue of "gun control" but in fact he created an enforcement climate where jaywalking could get you a busted skull. When looking cross-eyed at a cop will get you a "nightstick tan" then a lot of "out in the open" crime will dry up. It wasn't gun control, it was punk control.
Because these state-specific influences will bend the otherwise "pure" numbers, you kind of have to build your trends by state.
So, for Florida, when shall-issue CCW was introduced, what were the trends?
For Illinois (or is it Wisconsin?), where only the king and his entourage may carry, what are the trends?
Another meaningful comparison might be side-by-side with states having similar gun laws and carry restrictions. That is, can one fairly compare Arizona with Texas, given that one allows open carry and the other does not? Arizona along with, say, Nevada and Idaho might be more meaningful. And with what other state could you compare California? Or do you have to go to Europe for that?
That would be an interesting graph set, but surely a lot of work.
LightningJoe
Member
I don't have data, but my hypothesis is that the incidence of crime is proportional to the presence of criminals. That's why incarceration works so well.
Gun ownership has anecdotal-style effects, probably not much effect on statistics below some threshold. 100% gun-toting would probably show up in the statistics, but just a few here and there probably doesn't affect the overall numbers.
Gun ownership has anecdotal-style effects, probably not much effect on statistics below some threshold. 100% gun-toting would probably show up in the statistics, but just a few here and there probably doesn't affect the overall numbers.
svtruth
Member
Correlation
is a word that appears in many of the posts. I don't see any correlation statistics.
A couple of years ago I regressed gun violence on Brady grades, there was a weak but not significant inverse trend. worse Brady grades, less violence.
Can you do a regression on your data?
is a word that appears in many of the posts. I don't see any correlation statistics.
A couple of years ago I regressed gun violence on Brady grades, there was a weak but not significant inverse trend. worse Brady grades, less violence.
Can you do a regression on your data?
One of Many
Member
The population density is listed for the entire state, so where 95% of the state population is confined to a major metropolitan area, and the area of the less populated rural areas dilutes the true population density in the metropolitan area, the result is skewed.
Look at the percentage of the state population living in high density metropolitan areas, and the actual size of that metropolitan area; compare that to the remaining population in the state, and the remaining area when the metropolitan area is subtracted.
I expect that what you will find is that the higher the true population density, the higher the violent crime rate, regardless of the number of firearms owned.
Look at the percentage of the state population living in high density metropolitan areas, and the actual size of that metropolitan area; compare that to the remaining population in the state, and the remaining area when the metropolitan area is subtracted.
I expect that what you will find is that the higher the true population density, the higher the violent crime rate, regardless of the number of firearms owned.
Librarian
Member
It's possible to get crime rates from cities or 'metropolitan statistical areas' - FBI - but that's not all that helpful; some cities have different laws from the rest of the state, and outside of the obvious Boston and New York City examples, one would have to get all the city detail separately.
What's happening here is essentially a recreation of Lott's numbers; he had a lot more variables available to correlate and he used county-level crime rates. It's an entertaining exercise, and very educational, but the results are unlikely to sway any opinions.
ETA: Lott will share his data set - go here and sign up (I just did); he'll release it to a non-academic. His attitude seems to be sort of 'more replications, less doubt', or something like that.
Username and password are case sensitive; the data is in LARGE files, formatted for a stat package called STATA.
What's happening here is essentially a recreation of Lott's numbers; he had a lot more variables available to correlate and he used county-level crime rates. It's an entertaining exercise, and very educational, but the results are unlikely to sway any opinions.
ETA: Lott will share his data set - go here and sign up (I just did); he'll release it to a non-academic. His attitude seems to be sort of 'more replications, less doubt', or something like that.
Username and password are case sensitive; the data is in LARGE files, formatted for a stat package called STATA.
Last edited:
ConfuseUs
Member
It may be the case that gun ownership and violent crime are not related in a linear fashion. However, outside of DC, the data are fairly noisy. A regression line won't have a very steep slope either way if DC is eliminated. It is more likely (and not politically correct) that graphing minority population as a % of total population and violent crime will show some interesting correlations. DC is mostly African American but low gun ownership and the Dakotas are mostly white and high gun ownership.
Adding a factor to a statistical model to account for minorities as well as gun ownership rates and violent crime might show a few interesting relationships. I would guess that a multivariate model correcting for factors such as minority demographics, gun ownership, violent crime, income, and educational levels would indicate a more clear relationship between gun ownership and violent crime than just gun ownership and violent crime on a graph would. It may even turn out that gun ownership rates are not statistically significant for modeling violent crime, which would really make the case that guns do not cause crime.
Alas, it does not make for a nice pretty Excel graph. You only get a big equation with that method.
This is quite a large pile of crap. Or quite a pile consisting largely, of crap, should I say.
However, notice it says % of murders committed with a gun. It should be murders per capita instead, that would be conclusive.
Duh. Most murders occur with guns, as they are easier to use than knives.
However, if one murder occurs in an area, and it was done with a gun, suddenly you have 100% of murders committed with guns. It's ridiculous. Anyone with half a brain can see that.
However, notice it says % of murders committed with a gun. It should be murders per capita instead, that would be conclusive.
Duh. Most murders occur with guns, as they are easier to use than knives.
However, if one murder occurs in an area, and it was done with a gun, suddenly you have 100% of murders committed with guns. It's ridiculous. Anyone with half a brain can see that.
fletcher
Member
I plotted Brady grade vs. violent crime rate last year (using a scale similar to GPA) and found there really wasn't much of a correlation.
It's not his fault. It's the Brady Campaign's fault. He just made the graph, showing Brady's lack of care for the facts, and willingness to skew information.
I think the Brady Campaign in its entirety is one pile of crap.
I was under the impression he was trying to show the Brady statistics are a pile of crap, by graphing them out and showing little or no correlation.
I think the Brady Campaign in its entirety is one pile of crap.
I was under the impression he was trying to show the Brady statistics are a pile of crap, by graphing them out and showing little or no correlation.
- Status
Similar threads
- Locked
