John Lott strikes back

Via the hated Instapundit, we came across John Lott's response to the National Academy of Sciences study on gun laws and crime rates on the Volokh Conspiracy.

To many in the gun rights community, John Lott needs no further introduction, but to the casual observer, he seems a rather polarizing figure.

The Posse takes no position on the bitter debate swirling about him and the often nit-picking details used to try to pick apart his work.

Certain people (among them Tim Lambert) have an obsession with Lott that is unhealthy to say the least.  It is convenient for Lambert to helpfully point out that his entire motivation for blogging was to destroy Lott and his research.

Again, much of the debate boils down to thinly-disguised character assassination couched in arcane debates about data sets, dummy variables and such.

This is of course what sets Lott apart from such obvious academic frauds as Michael Bellesiles, who was doggedly run to ground by Clayton Cramer, among others.

Bellesiles, of course, simply made stuff up.  One didn't have to have a strong background in statistics and math to notice his glaring omissions, misquotations and outright fabrications.

But we digress.

The point of this post is to emphasize the fundamental truth in Lott's response to the NAH study:

Based on 253 journal articles, 99 books, 43 government publications, and some of its own empirical work, the panel couldn't identify a single gun control regulation that reduced violent crime, suicide or accidents.

Whether or not Lott forgot to carry the one in his data sets, this finding is really all that matters in the real world.

For decades academics have looked down their noses at gun owners and held forth on how guns are simply too dangerous for the common folks to own.

Why just owning a gun was an invitation to be robbed, killed, or have a family member shot.  The methodologies behind these "studies" were so pathetic that even laymen could see through them.

The other avenue of attack was to argue that the Framers of the Constitution and Bill of Rights didn't know how to speak English.

When they wrote "people" in the Second Amendment, they were referring to "states."

Everywhere else, however, the word "people" actually meant "people."

No one ever accused gun controllers of being honest or reasonable.

If nothing else, the NAS study blasts a gaping hole in the articles of gun control faith: that gun control does anything to reduce crime.

It doesn't.

Therefore, one of the key arguments in favor of gun control - that gun rights must be denied or eliminated in order to assure public safety - has now been completely discredited.

It was always lame, but now it's worthless.

We view this development with great pleasure.  Whether Lott is eventually vindicated in full or retires to begin a second career in network television news (we hear CBS has some vacancies), gun control has just been dealt a crippling blow.

Comments

[Kevin Baker]

I'm going to disagree with you here regarding Lott. I believe that Lott's data is unreliable, and that the conclusions drawn from it are highly questionable. I have a severe problem with his Mary Roush alter ego. I have a severe problem with his "lost data." I have a problem with his miscoding and his cherry-picking. I believe that Lambert is, while obsessive, not incorrect about Lott.

I believe that the data shows NOT that "More Guns = Less Crime" but that "More Guns = No Discernable Effect on Crime" but that "Less Guns Among the Law Abiding = More Crime". I think Lott, like Bellisiles, had an agenda and pushed that agenda. Lott has been more successful at it because disproving a statistician is much harder than disproving an historian.

That's my 2¢, anyway.

[Posse Incitatus]

Fair enough.

The point is that data encoding is not something most people can look at and say "Oh yeah, that's bogus."

Actually misquoting from historical sources and making up sources that don't exist ARE things we can lay our hands on and see as false.

As for pseudonyms, well, we at the Posse can't be too critical of that for obvious reasons.

[Tim Lambert]

So first you say that you take "no position" on Lott, and then you dismiss all the criticism of him as just "character assassination" and then attack my character? I also like the way you praise Cramer's "dogged" pursuit of Bellesiles, whiling smearing me for my "unhealthy" "obsession" with Lott. Double standard?

And you don't need a strong background in maths to notice Lott's "glaring omissions, misquotations and outright fabrications", either. You just have to look at the evidence. But you won't.

[RL]

Lambert is pretty funny. He gets upset when someone says he is engaged in character assasination and Lambert proves it by visiting every single site that mentions Lott's name. Talk about Tim "glaring omissions, misquotations and outright fabrications" Lambert, this gets funnier all the time. Just go to Lott's website and look at his discussions on the Baghdad murder rate, the Appalachian Law School attack, the Merced Pitchfork killings, and others to see some of Lambert's inaccurate statements.

[Carl Jarrett]

You write -"Bellesiles, of course, simply made stuff up."

In Lott's response, he simply makes stuff up.

For example, Lott claims that "The panel was set up during the Clinton administration, and of its members whose views on guns were publicly known before their appointments all but one had favored gun control."

Will John Lott, or any of his supporters, produce the prior public statements by all of the panel members to back up that claim?

I doubt it.

Nagin's statement demonstrates that John Lott simply made up the supposed conversation between Phil Cook and Dan Nagin.

Will John Lott produce the date, time, and location of this supposed conversation?

I doubt it.

Note how John Lott writes "Wilson said that that panel’s conclusion raises concerns given that “virtually every reanalysis done by the committee” confirmed right-to-carry laws reduced crime."

But Wilson didn't write that it reduced crime, his statement notes that it reduces murder. Note how John Lott selectively quotes from Wilson and then changes what Wilson actually wrote.

[RL]

A) Lott's Los Angeles Times piece and his book The Bias Against Guns gives the names of four people and he mentions that there are others.
B) Why don't you write him about the details on Nagin.
C) Wilson's discussion is not using the term "murder" with that quote. I look at the panel's own regressions and for the ones that Wilson was talking about they show drops in ALL the violent crime rates.

[RL]

A) Lott's Los Angeles Times piece and his book The Bias Against Guns gives the names of four people and he mentions that there are others.
B) Why don't you write him about the details on Nagin.
C) Wilson's discussion is not using the term "murder" with that quote. I look at the panel's own regressions and for the ones that Wilson was talking about they show drops in ALL the violent crime rates.

[RL]

A) Lott's Los Angeles Times piece and his book The Bias Against Guns gives the names of four people and he mentions that there are others.
B) Why don't you write him about the details on Nagin.
C) Wilson's discussion is not using the term "murder" with that quote. I look at the panel's own regressions and for the ones that Wilson was talking about they show drops in ALL the violent crime rates.

[RL]

A) Lott's Los Angeles Times piece and his book The Bias Against Guns gives the names of four people and he mentions that there are others.
B) Why don't you write him about the details on Nagin.
C) Wilson's discussion is not using the term "murder" with that quote. I look at the panel's own regressions and for the ones that Wilson was talking about they show drops in ALL the violent crime rates.

[Carl Jarrett]

You are stuttering, RL. Listing four names is not the entire panel nor is it evidence. Where are the quotes for "all but one"? Lott claimed they were public. He should produce the data, if he has it. I doubt he does and I'm sure he's simply making it up.

Wilson is most certainly referring to murder in the snippet Lott cherry-picked. On Page 269, Wilson writes "If this analysis of Lott's work showed that his findings are not supported by his data and models, then the conclusion that his results are fragile might be sufficient. But my reading of this chapter suggests that some of his results survive virtually every reanalysis done by the committee. Lott argued that murder rates decline after the adoption of RTC laws even after allowing for the effect of other variables that affect crime rates."

Lott is dishonest when he writes: "Wilson said that that panel's conclusion raises concerns given that "virtually every reanalysis done by the committee" confirmed right-to-carry laws reduced crime."

Wilson sums up his dissent with "In sum, I find that the evidence presented by Lott and his supporters suggests that RTC laws do in fact help drive down the murder rate, though their effect on other crimes is ambiguous."

Lott: "confirmed right-to-carry laws reduced crime."

Wilson: "their effect on other crimes is ambiguous."

If you can't see the difference between what Wilson wrote and Lott's inaccurate and selective rephrasing, there isn't much that can be done with you.

Why don't you write Lott and ask for the time, place, details of the Nagin and Cook discussion he claims to have overheard? Nagin has already denied he ever talked to Cook on the subject. Where's Lott's evidence?

You don't need to understand math and statistics to know that Lott is blowing smoke. This isn't about arcane debates about data sets, dummy variables and such, it's about Lott's lack of integrity.

Where's his evidence for his claims?

[Tom]

Dear Confused Carl:

Lott is referencing two different parts of WIlson's discussion. Lott clearly quotes Wilson as saying: "confirmation of the findings that shall-issue laws drive down the murder rate . . . " On the other hand, virtually ever reanalysis of the committee showed that violent crime fell. Only the cases where Wilson objected to them not accounting for other factors did they not find an effect.

You are getting as bad as Lambert in sellectively quoting things.

[Carl Jarrett]

Confused Tom - The NAS report does not show that "virtually ever reanalysis of the committee showed that violent crime fell."

Lott's statement clearly misrepresents Wilson's dissent.

The only person doing the selective quoting is Lott.

[Tom]

Dear Confused Carl:

Even if you are unable to read what Wilson wrote, possibly you can look at the actual regression tables in Chapter 6 where the panel did its own work. If you leave out the regressions that do not control for other factors that can effect crime (what Wilson was saying wouldn't get published in a refereed journal), the violent crime regression show that VIOLENT CRIME IS FALLING! Not just murder, but rape, robbery, and aggravated assault. Look at all the regressions that use the longest period of data up to 2000. Even if you pretend to be dense in reading Wilson and find it necessary to confuse two different paragraphs that he has written, even you can not deny this point.

[Tim Lambert]

Dear Tom, you seem to be having some difficultties reading the NAS report. p140 row 1 shows a statistically significant 4.12% increase in violent crime when the regression uses data to 2000. This is why Wilson said "murder" rather than "crime".

[Tom]

A) One of the two sentences being debated here by Wilson says: "his results survive virtually every reanalysis done by the committee."
B) If you read More Guns, Less Crime, there is a powerful discussion about how the simply dummy approach can be misleading and this appears to be such an example.

[Tim Lambert]

Nice try, Tom. Wilson's sentence reads "some of his results survive virtually every reanalysis done by the committee." You left out the words "some of". Why did you do that?

Lott originally used the dummy variable model and only switched when it started to give results that he did not like. Neither the committee nor Wilson seem to have found Lott's reasons for discounting the dummy variable model persuasive.

[Tom]

Are you serious? No matter how you try to stretch things it is obvious that Wilson's argument in that sentence was much broader than just murder. Your discussion of the dummy variable issue is just childish. The discussion and graphs regarding the "Inverted V" argument in his book make it very clear that the simple dummy variable approach can be very misleading. I could quote the passages, but you would try to avoid the issue. Why don't you deal with the substance? If it is a V shape, how can the averages on either side be different? THis is discussed in more depth in his book.

[Tim Lambert]

Wilson's conclusion: "In sum, I find that the evidence presented by Lott and his supporters suggests that RTC laws do in fact help drive down the murder rate, though their effect on other crimes is ambiguous."

Are you going to try to tell us that Lott stopped reading Wilson's dissent before he got to the conclusion?

If it's a V shape then the dummy variable model is misspecified. If there is an abrupt increase or decrease when the law is passed, the trend model is misspecified. Raising a possibility of misspecification does not mean that that possibility is true. What is required is a statistical test for misspecification. Hoorowitz performed such tests (see Appendix D). ALL of the models were misspecified. That means that NONE of the models fit the data and they should all be thrown out, not just the dummy variable one.