"Guns are Bad News for Women?" Huh ... - Part VII

Well, after six parts this is getting long so hopefully I can wrap this up with some fun. To begin I want to address mikeb's assertions directly.

"Guns are bad news for women." Up to this point you may have noticed that I have purposely not actually addressed this point. Basically, there is nothing in this statement that can be refuted. It is a meaningless statement. "Bad news" is not some universal term that we all agree on, so how can one refute something that no clear definition is given. Based on his posts I am guessing that it is because women are killed by guns. OK... Women are killed by heart disease, cars, poison, drowning, falling, fire, etc, etc, etc at a higher rate than they are killed by firearms. Men are too.

Pretend for a moment the 2nd Amendment didn't exist. To advocate limiting an inanimate object, one must weigh the benefits and the risks. The gun control movement has done a fine job of over exaggerating the risks. Rarely if ever are the benefits presented. People do use firearms for self defense. Every day. Every hour and every minute (that would only be 525,000 if one defensive gun use happened every minute). These are crimes that are prevented. PREVENTED! Meaning, the crime doesn't happen, or at least it doesn't succeed in its goals. Target shooting provides relaxation and enjoyment for millions of Americans. Hunting is used by many to put food on their table AND wildlife management. Collector's assemble pieces of history. All of these are benefits to society. The gun control movement never weighs their benefit to the cost of implementing whatever the limitation du jure is.

Frankly, all of those are meaningless. We have the 2nd Amendment, it has not been repealed.

Now for the fun part. I alluded to this before, and I thought that I would do a little Hemenway type experiment. I picked some things that may or may not contribute to gun deaths and did a Hemenway/mikeb style analysis on them. I will post a table (with the 10 highest states and 10 lowest states) for each factor and give a brief synopsis of whether the factor is bad news for women. I did take a look at what all 50 states show and if there is something interesting, then I included it. Here goes:

First up, poverty rate (from the US Census Bureau):

2006 Death Rate of Women (per 100,000)
	High Poverty States	Low Poverty States
Firearm homicide	1.82	0.83
Other homicide	1.44	1.01
Total	3.26	1.84
Firearm suicides	2.04	0.79
Other suicides	2.60	3.21
Total	5.66	4.00
Firearm accidents	0.12	0.02
Other accidents	33.18	23.92
Total	40.68	23.94

Well, this is showing that poverty is bad news for women. Where there is poverty, more women are killed with firearms (and if you notice from the total numbers, more women are killed overall). It also looks as if people in poverty have significantly more accidents. Since I (and others) have said that poverty is a cause for crime, lets look at the states individually:

The suicide rate is looking somewhat more like there is a trend, however, the homicide rate clearly has a direct correlation to poverty. There are a few outliers, but nothing like the randomness that was seen when comparing gun ownership and firearm homicides.

Next, percentage of the population that live in urban areas:

2006 Death Rate of Women (per 100,000)
	High Urban States	Low Urban States
Firearm homicide	0.91	1.69
Other homicide	1.19	1.46
Total	2.10	3.15
Firearm suicides	0.55	2.12
Other suicides	2.88	2.91
Total	3.43	5.03
Firearm accidents	0.02	0.16
Other accidents	20.58	36.38
Total	20.6	36.53

So low urban population states have higher homicide and suicide rates with firearms. I 'd have to go back and check the definition that was used for the numbers but I think the datasource I used includes everything in cities/towns greater than 5000 people. (This would be nice to analyze further based on metro areas, large cities, etc.) What is most interesting from this chart is that low urban areas have vastly more accidental deaths than highly urbanized area (perhaps as a result of medical care not being right around the corner?). The graphs showed random chaos (more so than gun ownership, so I didn't include them).

Next, violent crime rates:

2006 Death Rate of Women (per 100,000)
	High Violent Crime States	Low Violent Crime States
Firearm homicide	1.78	0.78
Other homicide	1.63	0.85
Total	3.40	1.63
Firearm suicides	2.08	1.54
Other suicides	3.61	4.09
Total	5.68	5.63
Firearm accidents	0.08	0.04
Other accidents	32.09	30.20
Total	32.17	30.23

This is very interesting. I think it should be common sense that high violent crime states would use firearms more than low violent crime states (homicide being a violent crime) although FBI statistics indicate that only 9% of all violent crime has a firearm involved. But it appears in all three categories, violent crime is bad news for women. Maybe we should limit the availability of violent crime. The homicide rate showed a slight indication of a direct correlation between firearm homicides and urbanization (the opposite of what the table shows), but not enough that I would make a definitive statement from it, and therfore didn't include them.

Next depression, I have put this forward (as have numerous social scientists) as the primary factor in suicides so I expect there to be a good correlation.

2006 Death Rate of Women (per 100,000)
	High Depress States	Low Depress States
Firearm homicide	1.17	1.28
Other homicide	1.21	1.25
Total	2.38	2.54
Firearm suicides	1.85	1.18
Other suicides	4.34	3.08
Total	6.19	4.26
Firearm accidents	0.06	0.04
Other accidents	30.09	24.52
Total	30.16	24.56

Firearm suicides is showing a positive correlation, while homicides is very neutral (of course, I don't know that depression has ever been put forward as a cause of crime). The chart for suicide tells a different story though:

Notice how the values are scattered around just like gun ownership. This indicates on the surface that there is no correlation between depression and suicide. However, studies have shown that 80%+ of people who commit suicide had been depressed. In this dataset, there really isn't that much variation between the lowest depression state 6.74% and the highest depression state 10.14%. For the next few I chose some things at random.

High school graduation rates:

2006 Death Rate of Women (per 100,000)
	High HS Grad States	Low HS Grad States
Firearm homicide	0.92	1.82
Other homicide	1.04	1.60
Total	1.96	3.42
Firearm suicides	1.11	2.31
Other suicides	3.34	3.32
Total	4.45	5.62
Firearm accidents	0.05	0.10
Other accidents	32.24	33.72
Total	32.29	33.82

Looks like not graduating from high school is bad news for women. About twice as many women are killed by firearms in states with low high school graduation rates. Maybe we should give away more diplomas? Looking at the graphs, there is a weak inverse correlation for high school graduation to firearm homicides. More outliers appear than the poverty graph but not near the chatter of the gun ownership graph.

Next, percent of Catholics per state:

2006 Death Rate of Women (per 100,000)
	High Catholic States	Low Catholic States
Firearm homicide	1.13	1.21
Other homicide	1.16	1.37
Total	2.29	2.58
Firearm suicides	1.12	1.26
Other suicides	3.41	3.14
Total	4.53	4.40
Firearm accidents	0.02	0.06
Other accidents	26.09	28.26
Total	26.11	28.32

Good, this one did exactly what my impression would have been, nothing. So being Catholic makes no difference to whether a woman will be killed by firearms.

The next is my favorite. In one of my comments I alluded to cheese production, but I couldn't find quick numbers for that so I chose Milk Production per capita instead.

2006 Death Rate of Women (per 100,000)
	High Milk States	Low Milk States
Firearm homicide	1.06	1.30
Other homicide	1.10	1.22
Total	2.16	2.52
Firearm suicides	0.98	1.04
Other suicides	3.56	2.85
Total	4.54	3.89
Firearm accidents	0.02	0.12
Other accidents	24.99	27.93
Total	25.01	28.06

At first glance it appears that milk production per capita has no bearing on whether you will be killed by a firearm, but as I looked at these numbers I realized that they are some of the lowest values for any of the groupings I have made so I had to check the graphs:

These are interesting. There are indications of channeling occuring for both homicides and suicides. Does milking goes have some sort of weird cosmic effect where a minimun and maximum rate exist? That would be a research project. Maybe I could get stimulus money to fund that.

Next, cars per capita.

2006 Death Rate of Women (per 100,000)
	High Car States	Low Car States
Firearm homicide	1.27	1.11
Other homicide	1.33	1.40
Total	2.60	2.51
Firearm suicides	1.57	1.18
Other suicides	3.26	2.94
Total	4.83	4.12
Firearm accidents	0.08	0.05
Other accidents	32.79	25.61
Total	32.88	25.66

Well, this shows that cars are bad news for women, not as bad as poverty or high school graduation rates, but bad nonetheless. Of course, owning a car may just be a low grade inverse indicator of poverty.

Next, cubic miles of rainfall per capita.

2006 Death Rate of Women (per 100,000)
	High Rain States	Low Rain States
Firearm homicide	1.05	0.92
Other homicide	1.24	1.17
Total	2.28	2.08
Firearm suicides	1.70	0.55
Other suicides	3.94	2.94
Total	5.64	3.49
Firearm accidents	0.05	0.02
Other accidents	33.10	20.32
Total	33.15	20.34

While it doesn't seem to matter, it does appear that rain is bad news for women who are going to commit suicide. For some reason it is also especially bad for accidents in general (think cars slipping on roads, people slipping off wet roofs, etc.)

Finally, birthrate.

2006 Death Rate of Women (per 100,000)
	High Birth States	Low Birth States
Firearm homicide	1.36	1.12
Other homicide	1.26	1.15
Total	2.62	2.27
Firearm suicides	1.58	1.31
Other suicides	3.21	3.54
Total	4.79	4.86
Firearm accidents	0.04	0.05
Other accidents	25.35	30.29
Total	25.39	30.34

Similar to cars there appears to be a slight correlation. Giving birth might be bad news for women. The graphs show no such thing.

Do I believe there is any connection between milk production, cars per capita, Catholics, or rainfall and firearm deaths? No, and you shouldn't either. In fact, what you should realize from this part is that any analysis of two datapoints is nothing but fantasy (even when there actually is a correlation).

Now, while I may not have the credentials of Mr. Hemenway, I would certainly be amenable to some organization paying me for my analysis. So, if there are any organizations that need a shill, I am available for the right amount of money. Until such a time, I'll continue to do things like this because in some sick way I find it fun. If anyone desires to have the spreadsheet of data that was used to create this, email me and I'll send it to you (that way you can check my work and if it is wrong, I'm more than happy to provide a correction). If you have some other question or factoid that you think might make an interesting research project for me, send it too me. I might take it up the next time I have some spare time!