Look: I am eager to learn stuff I don't know--which requires actively courting and posting smart disagreement.

But as you will understand, I don't like to post things that mischaracterize and are aimed to mislead.

-- Brad Delong

Copyright Notice

Everything that appears on this blog is the copyrighted property of somebody. Often, but not always, that somebody is me. For things that are not mine, I either have obtained permission, or claim fair use. Feel free to quote me, but attribute, please. My photos and poetry are dear to my heart, and may not be used without permission. Ditto, my other intellectual property, such as charts and graphs. I'm probably willing to share. Let's talk. Violators will be damned for all eternity to the circle of hell populated by Rosanne Barr, Mrs Miller [look her up], and trombonists who are unable play in tune. You cannot possibly imagine the agony. If you have a question, email me: jazzbumpa@gmail.com. I'll answer when I feel like it. Cheers!

Tuesday, January 18, 2011

Yet Another Look at Gun Violence

For your convenience, earlier posts on this subject, with - you know - actual data and stuff - can be found here and hereEd has a post up talking about how the Regressive meme is to turn the focus from guns in gun violence to mental health in gun violence.  Well, guess what - there's a bunch of sociological data available that can be compared with gun death data.   This gives me a chance to play with my new favorite toy - cloud charts - in the Open Office spreadsheet application.

We'll have a look at how gun fatalities correlate to variety of social phenomena, and I'll just let the cat out of the bag early by saying that the one thing that it has absolutely no correlation with is the incidence of poor mental health.  Let's work our way down.

For these charts I've plotted the numbers for a variety of social phenomena on the X-axis vs gun deaths on the Y-axis.  Each point represents the paired data for a single state.  In each graph, the best 10 (in green) and worst 10 (in red) States for gun deaths are indicated by an oversized dot.  The best of best and worst of worst States are in the biggest square dots.

Here are the worst States, ranked by Gun death incidence, 2007 data.

Louisiana         20.2
Mississippi      18.5
Alaska              17.8
Alabama           17.5
Nevada             16.3
Arkansas          15.2
Arizona             15.1
Wyoming          15
New Mexico     15

And the best.

Hawaii                    2.6
Rhode Island         3.4
Massachusetts      3.6
Connecticut           4.2
Iowa                       5
New York              5
New Jersey           5.2
New Hampshire    5.5
South Dakota        6.1
Minnesota             6.5

The strongest correlation among all the things I looked at was with Vehicle Deaths, where there is a strong correlation with an R of .8, and R^2 of .63. 

The best way I can rationalize this is by grasping for a common cause - perhaps substance abuse plays a part.   Note the tendency of best states to cluster at the bottom left and worst states to cluster at the top right.  What we will see quite frequently in these graphs is the best States will cluster more tightly, and the worst states will spread out more, right to left.

The chart for the Total Death Rate looks pretty similar.  I guess it makes some sense for these two to move together, though they differ in quantity by about two orders of magnitude.  At R = .7 and  R^2 = .49,  there is reasonably strong correlation.

Suicide rate is where I expected the strongest correlation.  Alas, twas not to be.  Still, with R^2 at .42 (indicating a correlation coefficient at a quite respectable .65)  the correlation is significant. 

Now, this one is a surprise.  There is identical correlation (R = .65; R^2 = .42) to the infant mortality rate.

Correlating gun deaths to the incidence of violent crime seems intuitively right.  But R is only .5, with R^2 at .26, so this is bordering on weak correlation.

You get the same correlation to the instance of low birth weight.

What about other health-related events?


Colorectal Cancer Death Rate.

Not particularly strong correlations, but something, anyway.

This is what a total lack of correlation looks like: an essentially horizontal trend line (i.e. no trend at all) and correlation near ZERO.

I looked at a number of things that would not be expected to correlate with gun deaths, and found various levels of correlation, from strong to weak.  What these things may indicate is up for speculation.  Thom Hartman has mentioned studies that correlated all manner of social ills to wealth disparity, and that is certainly one sitation that has gotten much worse in this country since 1980.

It's pretty clear that Louisianna, Mississippi, and Alabama are lousy places to live, but great places to die.

The only social variable in this array that has absolutely no correlation with gun deaths is poor mental health.


BadTux said...

The question is, how do you define "poor mental health"? Psychiatric visits? One thing I'll point out is that most people in Louisiana and Mississippi (two states I'm quite familiar with) have no access to mental health professionals at all, either because there simply are no mental health professionals in their communities, or because they still subscribe to the notion that poor mental health is a moral weakness rather than a medical issue and thus simply won't go unless involuntarily committed. Meanwhile, in New York City it's trendy to have your very own psychotherapist. Does that mean that New Yorkers are crazier or have poorer mental health than the crazed gun-stroking hillbillies of Mississippi and Louisiana(*)? I think not... but I also suspect that your "poor mental health" statistic would say yes, New Yorkers *are* crazier than Louisianians. At which point, having been both places, I call bull**** :).

- Badtux the "Definitions matter!" Penguin

(*)Note that contrary to popular belief, the top 3/4 of Louisiana is hills, not swamps, and most of Mississippi is hills, not the famous Mississippi delta prairies where cotton is grown.

Jazzbumpa said...

Tux -

You raise a number of good points.

I didn't check into the way any of these variables are defined or measured. This is just a number crunching exercise, at this point.

That's why I'm not eager to draw any conclusions. OTOH, I really expected non-correlation to things like colon cancer deaths and infant mortality. The existence of ANY correlation suggests that States have differing approaches to issues of public health/common welfare that cut across a number of social resultants.

And, I think Hartman is on to something.