Wednesday, June 9, 2010

Gun Sales Up - Murder Rate Down

HolyCoast.com made a post on "gun control," or at least that was the category. I don't agree with Rev. Rick on much of anything, but I do like his blog. This post was no exception to either of those things. As I was reading I thought it sounded familiar. It became clear that the point was BECAUSE so many people bought guns last year the murder rate went down. What's that "causation" thing they're always criticizing me about? Anyway, when I clicked on the link to Big Government, which the Rev. was liberally quoting, I discovered why it was sounding so familiar. Who's face do you think I saw there?

At the same time gun sales were soaring, there was an unusually large drop in murder rates. The 7.4 percent drop in the murder rate was the largest drop in murder rates since the 1999. For those who don’t remember, 1999, when President Bill Clinton and Columbine occurred, was another time when gun sales soared. With people such as Elena Kagan serving as Mr. Clinton’s deputy domestic policy adviser were pushing hard for more gun control, Americans were worried that more gun bans were coming. And in response gun sales soared.

What's your opinion? Is it really as simple as that? Can you really say, 1. gun sales went up, 2. murders declined, the first one caused the second one?

Please leave a comment.

16 comments:

  1. If gun sales had gone up by the same amount, but murders had instead gone UP by the the same amount by which they had gone down....

    1) What would gun control advocates be saying about that? Do you think that they just might, maybe, somehow, imply causation? Naah.

    2) What would mikeb have to say about such claims? Doubt? Of course not! Or would he instead say:

    "That's it. The argument is over. The gun advocates can just go home now and wait for the government truck to come by to collect their guns."

    ReplyDelete
  2. This comment has been removed by the author.

    ReplyDelete
  3. they can be used to prove that the higher availability of guns has not caused an increase in crimes.

    Yup, and that definitively proves MikeB, Jade & the rest of their ilk wrong.

    ReplyDelete
  4. What RuffRidr said!

    This is the exact point I've been trying to get across in every one of my posts:
    Raw statistics are meaningless, especially if they're being used as propaganda.

    I would be so happy if you got interested in real statistical analysis. I could recommend a few good text books if you'd like?

    ReplyDelete
  5. I have, for a long time, expressed my skepticism regarding the idea of a causal relationship between dramatically spiking gun ownership and precipitously plummeting violent crime rate.

    That being said, I agree with FishyJay that, had the numbers gone the other way, the forcible citizen disarmament lobby would be loudly trumpeting them as "proof" of the need to fast track their agenda, and I suspect our gracious host would be cheering them on.

    I also agree with RuffRidr, that although there is not--John Lott (and Mary Rosh) notwithstanding--compelling evidence for "more guns = less crime," we are seeing a more and more compelling case for "more guns ≠ more crime."

    All of which takes a distant second place to the rather obvious fact that "fewer gun laws = more liberty."

    ReplyDelete
  6. I am in agreement with the majority here. No causation, but puts to bed the gun control argument that gun availability causes more crime.

    ReplyDelete
  7. RuffRidr said, "That being said, while you can't use the statistics to proves that guns reduce crimes, they can be used to prove that the higher availability of guns has not caused an increase in crimes."

    The rest of you agreed. In all fairness there are a few regular commenters absent.

    Sometimes I think you guys, RuffRidr and company, are starting to believe your own bullshit.

    How could the one correlation fail because it's "a large animal controlled by many different factors," and the other correlation work?

    Couldn't the same "many different factors" enter into the second correlation? In other words, the number of guns increased, crime did not increase, therefore we conclude that more guns does not cause more crime. What about the factor of law enforcement techniques, the economy, legislative changes, or even the weather or the alignment of the planets, for crying out loud.

    My point is, if you want to play the correlation/causation card, play it across the board.

    ReplyDelete
  8. This comment has been removed by the author.

    ReplyDelete
  9. n other words, the number of guns increased, crime did not increase . . .

    "Crime did not increase" is one way to put it, but my preference for expressing the situation would be to say that violent crime has fallen, again, and by a rather large amount.

    . . . therefore we conclude that more guns does not cause more crime. What about the factor of law enforcement techniques, the economy, legislative changes, or even the weather or the alignment of the planets, for crying out loud.

    What "changes in law enforcement techniques" would those be--what have cops suddenly discovered, that they didn't know before?

    "The economy"? The tanking economy, causing a dramatic decrease in crime? That's why poverty stricken nations tend to be peaceful paradises, eh? And going back to your "changes in law enforcement techniques," police departments must be pulling off this miracle with fewer man-hours, stemming from strained budgets caused by the economic problems. This article is from 2008, but few would say the economy has gotten much better since then. There's also the issue of police departments suffering because of experienced police officers deploying to Iraq and Afghanistan, with National Guard and Reserve units.

    "Legislative changes"? Like what? More castle doctrine laws? More states with concealed carry?

    "The weather"? I hear violence goes up when the weather gets warmer, and Al Gore says the Republicans are making the world burn up.

    C'mon--the case, such as it ever was, for "more guns = more violence" is in deep trouble, and none of your dancing around with amusing attempts at excuse-making can hide that.

    ReplyDelete
  10. Mike, this is why you should read a textbook.

    "has not caused an increase in crimes" is a negative hypothesis, or 'null.'

    You test your hypothesis (guns increase crime) against the null: if there is a correlation with all other factors included at a given confidence interval, you reject the null. Otherwise, you "cannot reject the null."

    If you cannot prove that something is happening, you say "nothing appears to be happening."
    That is what is being said in this instance.

    Are you being deliberately dense?

    ReplyDelete
  11. RuffRidr's last post is spot on, and more proof of why you need to read a text book before you talk about stats.
    In true anonymous fashion, I must advise you to LURK MOAR instead of arguing about statistics.

    To preempt you saying "I don't believe you," here's my proof:

    Make a very simple linear regression model.
    C=log of Crimes
    L=log of # of cops
    G=log of # of guns
    X,Y,Z=log of other factors previously established to affect crime rates (economy, immigration, whatever). Be prepared to defend your model with ample citations!
    So
    C=b1*L+b2*G+b3*X+b4*Y+b5*Z

    Plug in the data and run the regression to get estimates for the "b" coefficients. You also get an 'R' value that says how much of the changes in C can be explained by your model, and how much is the (assumed random) error term.

    Since the effect of gun ownership would have to be pretty huge to make it past the noise, what we will get in this case is "nothing seems to be happening"
    (b2 not >0 or <0 at a 95% confidence interval).

    Formal conclusion: "we cannot reject the null hypothesis that a rise in gun ownership has no effect on the crime rate for the given sample."
    In english: "having more guns doesn't seem to be doing anything, at least in this case."

    The italicized bit is important, because samples only tell you about themselves: you can't use this study to argue about gun violence in Pakistan, for example. That's another big sin of Lott.

    ReplyDelete
  12. MikeB, correlation does not mean causation, but in your case you don’t even have correlation. That is the least you need before trying to make a causation argument.

    ReplyDelete
  13. TS, I'm not the one making the bogus causation argument. Anonymous says RuffRidr has it right and I need to study up in order to understand. That may be the case, but still, I'm not making any conclusion-type arguments. Those have come from your side.

    ReplyDelete
  14. "I'm not making any conclusion-type arguments. Those have come from your side."

    Right, except for
    http://mikeb302000.blogspot.com/2010/06/more-guns-more-gun-deaths.html
    and
    http://mikeb302000.blogspot.com/2010/06/another-place-where-gun-control-works.html

    Lay off the reefer, man. It's really hurting your short-term memory.

    ReplyDelete
  15. Wait, let me clarify. I meant I was not making any conclusion-type arguments in this post. I didn't mean I've never done such a thing. But, this parcitular post was about that silly nonsense the conservative Rev. Rick said over at HolyCoast and about RuffRidr's double application of the correlation/causation thing.

    ReplyDelete
  16. This comment has been removed by the author.

    ReplyDelete