I looked at some of Lott’s anonymous reviews of other people’s books on Amazon.com in this posting. Yesterday I found that Washingtonian had made a five-star review of More guns, Less Crime, just like Mary Rosh did. Today we are going to see just how many other five-star review.Lott gave his book.

To understand what follows you need to know three things:

  1. Lott obsessively checks the reviews of his books.
  2. Lott just has to reply to any criticism. Whenever anyone posted something critical on Usenet, Mary Rosh would leap to Lott’s defence. For pretty well every paper critical of Lott’s work, Lott has written a reply. Back in January when the Lott affair burst into blogspace, Lott (and Mary) would email bloggers who criticized him.
  3. Amazon includes a few recent customer reviews on the page about each book. The most recent review is listed first. So if there is a critical review in first position and someone posts a favourable review, it pushes the critical review to a less prominent position. A few more reviews and the critical review gets pushed off the front page and few people will be likely to see it.

Given that, if he saw a powerful negative review of his book you have to ask youself, WWLD? (What Would Lott Do?) Let’s roll the tape:

On October 29, 2001, I posted my review of the 2nd edition of More Guns, Less Crime:

i-6bfd3a36efd20db21f4af173d22351ee-stars-2-0.gifLott has not fixed the problems with the first edition
Reviewer: Tim Lambert (see more about me) from Maroubra, NSW Australia
The only significant difference between this edition and the first edition is the addition of chapter 9, where Lott updates some of his results and responds to the extensive criticism of the first edition.

I’ve already posted a review of the first edition, and mentioned two serious flaws, so to get a feel for how Lott does in his new chapter, we can look at how he addresses these two criticisms.

First, that Lott used two exit polls to estimate gun ownership even though dozens of other polls and guns sales statistics contradicted his results. Lott responds in section 9.9, but fails to give any reason why his polls should be regarded as better than all the other ones.

Second, Kleck’s critique that the change in the number of people carrying was too small to have produced the result Lott observed. Lott respond to this one twice, in section 9.7 and again in section 9.14 (misattributing the Kleck quote to me in that section). Even with two goes at it, Lott does not have a good response. In 9.7 he offers an “explanation” of Kleck’s position that completely misses the point. Things get worse in section 9.14. Lott asserts that the survey results (given in Kleck’s “Point Blank”) on gun carrying include mere transportation and not just carrying for protection. Lott’s assertion can easily be seen to be false by anyone who looks at Kleck’s book. Then Lott makes the bizarre assertion that it is misleading to look at current permit rates because more people might get permits in the future. However, Lott believes that the current low permit rate caused relatively large decreases in crime, so the future rates, whatever they might be are not relevant.

In summary, Lott has not fixed the problems with the first edition, so my rating has not changed.

A few days later, Washingtonian posted a review replying to mine and accusing me of dishonesty:

i-6da7393c8ff6de1c28d52199d33a151e-stars-5-0.gifImportant accurate info that Opponents constantly distort, November 8, 2001
Reviewer: washingtonian2 (see more about me) from Swarthmore, PA USA

This is by far the most comprehensive study ever done on guns. It provides extensive evidence on waiting periods, the Brady Act, one-gun-a-month rules, concealed handgun laws. For some gun laws this is the only study available and it is important to note how many academics have tired to challenge his work on concealed handgun laws and failed and that no one has even bothered to try and challenge his work on one-gun-a-month laws and other gun control laws.

I am constantly amused the lengths to which reviewers here will go to distort Lott’s research. Take the one by the Australian who claims that Lott doesn’t explain why he uses the polling data that he does on gun ownership rates. If he was honest, he would note that Lott talks about these being the largest surveys on gun ownership rates available and that it is necessary to have such a large survey to get detailed information at the state level. A survey of 1,000 or even 1,500 people nationally is not enough to allow you to make comparisons across individual states.

These guys will do anything to keep people from reading Lott’s work.

All right, we knew about that one already, but now we know the pattern.

Our next exhibit is this critical review:

i-ca5366d20b995c41c5e842b2d1c03ab0-stars-1-0.gifPoor Research Cannot Be A Basis For Policy Changes, October 9, 2000
Reviewer: Mark Wylie (see more about me) from Spokane, WA United States
John Lott’s provocatively titled “More Guns: Less Crime” has attracted enormous media attention and has become a powerful rhetorical weapon in the hands of gun control opponents. However, his work is marred by poor statistical analysis and sloppy reasoning, and should not be a basis for public policy.

Lott argues that if most adult Americans were able to get a permit to carry a concealed weapon in public, violent crime would fall. His reasoning is that criminals are calculating people who sit down and work out the benefits and costs of a potential criminal act; therefore, if they know that potential victims are likely to be carrying guns, they will be deterred from committing crimes against persons, and will switch to committing property crimes where they are less likely to encounter an armed victim. Lott then presents some extensive statistical analysis purporting to show that when jurisdictions around the US pass “shall issue” laws that enable most adults to carry concealed weapons in public, violent crime falls in those jurisdictions.

There are errors both in Lott’s reasoning and in his statistical analysis. Even those lacking the technical training to evaluate the latter can see the error in the former. Lott looks at criminals from the perspective of the neoclassical economist, viewing people as always rational–they always act in their own self-interest, carefully weighing the costs and benefits of any possible action. But anyone familiar with violent crime knows that much of it is extremely impulsive. For example, nearly half of all murders take place as a result of arguments or brawls, situations where people are hardly taking rational account of costs and benefits of their actions.

Turning to Lott’s evidence, you see that he tries to calculate the effect of passing “shall-issue” type laws on crime by using data from every county in the US to conduct what is called a multivariate regression analysis–a technique used by social scientists to analyze complex sets of data. It is this regression analysis that is the basis for the “more guns, less crime” conclusion. Such conclusions, however, are valid only if the analysis they are based on is valid, and there are several problems with Lott’s analysis, three of which I want to outline.

1. The Robbery Problem. Lott is studying the effects of “shall-issue” laws, which affect the ability of people to carry guns in public, not to have them in their homes. Logically, the greatest impact of these laws should be on robbery, which of the four main categories of violent crimes (the others are murder, rape and assault) is the most likely to occur away from the home in a public place, and which is also the most likely to occur as a result of advance planning. In fact, Lott finds that the impact of shall-issue laws on robbery rates is far smaller than on other violent crimes. This result is inconsistent with the deterrence theory he is proposing and suggests that there is something seriously wrong with his statistical analysis.

2. The Adult/Juvenile Homicide Problem. Lott’s analysis makes no distinction between murders of adults and juveniles. His logic, however, suggests that the effect of shall-issue laws should be greater on adult homicides than on juvenile homicides. Since the laws allow adults only to carry concealed weapons, from the criminal’s viewpoint, any adult is potentially defended, but juveniles are protected only when in the company of adults. However, when Professor Jens Ludwig redid Lott’s analysis looking at adult and juvenile homicides separately, he found that shall-issue laws lead to an increase in adult homicides.

3. The Stranger Homicide Problem. A similar problem exists when homicides are broken down by the relation of offender and victim. Lott’s logic suggests that “shall issue” laws should have a greater effect on “stranger” homicides, than on homicides where offender and victim are related, because someone intending to kill a family member will likely know whether their victim is armed, while someone killing a perfect stranger will not. But Albert Alschuler has found that this is not the case; once again, the data contradict Lott’s logic.

It is true, as Lott’s defenders will protest, that he attempts to respond in his book to some of these criticisms as well as others that I did not have space to review. However, like his analysis, his counterarguments are unconvincing. For example, on the adult/juvenile issue, he merely argues that juveniles are protected by the presence of adults in public, ignoring the fact that, as I noted, juveniles are protected only when in the company of a (potentially armed) adult, while adults are protected all the time.

Readers may ask why, if Lott’s work is so badly flawed, have so many people accepted it. Part of the reason seems to me to be the psychological effect it has on gun control opponents. The strongest argument in favor of gun control is, clearly, that it would save countless lives. Lott’s book allows opponents of gun control to pose as the true “life-savers,” enabling them to feel better about themselves. This attitude can be detected in many of the favorable reviews of Lott’s book posted here.

Sure enough, a few days later a review replying to Mark Wylie’s appeared:

i-6da7393c8ff6de1c28d52199d33a151e-stars-5-0.gifSecond edition is even more powerful than the first edition, October 26, 2000
Reviewer: maximcl from Philadelphia
As an academic I can honestly say that I have never seen so much data assembled to study a topic. There is no doubt that Lott has done by far the most comprehensive study on crime ever conducted. Not only is it the largest number of counties and cities studied over the longest period of time, but he accounts for more factors than anyone else has even come close to accounting for. It is an impressive accomplishment, but just as amazingly Lott provides this information in a readable and interesting manner. Most academics are unable to explain what they have done in common sense easily understandable terms. Not Lott. He eliminates the useless jargon that fills academic discussions and discusses what he has done in readily accessible terms. If you understand percentages, you will understand this work.

I also have to comment on some of the critical comments made by other readers here. I can only conclude that they have not read the book. As someone who has seen and been involved in academic debates, this is a particularly strange discussion. People repeat claims that they must have heard others make, but they are not correct. Take the review below by Mark Wylie.

1) The Robbery Problem. “In fact, Lott finds that the impact of shall-issue laws on robbery rates is far smaller than on other violent crimes.” This is false. Here is just a fraction of the pages that argue that the effect on robbery is very large (indeed, the largest single effect): 78, 133, 137, 173, and 215-217.

2) The Adult/Juvenile Homicide Problem. Wyle writes: “Lott’s analysis makes no distinction between murders of adults and juveniles.” Again, this is completely false. Lott not only discusses this possibility early on but he goes through a discussion to explain how the results for these different age groups all fit together. See for example Lott’s discussion on pages 98 and 147-148.

3) The Stranger Homicide Problem. Lott explains that gun ownership can also stop attacks when the attacker knows the victim. For example, see his discussion on pages 148-150.

Possibly, these attacks will work as long as they keep people from reading this book, but once people read it they will be amazed about how much that they have heard is so completely false. It is amazing that those like Wylie, who make claims such as Lott “makes no distinction” about the impacts of the law by age, make attacks that are so easily disproved once someone reads the book. Either he did or did not make this distinction. Any reader will clearly see that he spends substantial time on this point.

This one is from Philadelphia, just like Mary Rosh’s review. Did you notice the name of the reviewer? maximcl would be Maxim C. Lott, John Lott’s son, the “Ma” in MaRy RoSh. The same son who Lott tried to blame for the Mary Rosh review. Now it is obvious from the writing style that John and not fourteen-year-old Maxim wrote it, but anyone who wants to argue that Maxim did write it has to face the fact that it begins “As an academic” and fourteen is too young to be an academic.

I also reviewed the first edition of More Guns, Less Crime:

i-6bfd3a36efd20db21f4af173d22351ee-stars-2-0.gifKleck’s book is better, June 10, 2000
Reviewer: Tim Lambert (see more about me) from Maroubra, NSW Australia
This book is a greatly expanded version of a paper published by Lott and Mustard in 1997. In that paper they claimed that laws allowing the concealed carry of handguns caused significant decreases in violent crime. The paper was Lott’s first publication on firearms policy and unfortunately Lott’s inexperience with this subject shows.

Contrast the treatment of the topic of gun ownership in Lott’s book with that in another book by a pro-gun scholar, Gary Kleck’s “Targeting Guns”. Lott looks at the results of two exit polls, conducted in 1988 and 1996 and concludes that the percentage of the population that owned guns increased by 50% in just eight years. Kleck looks at 86 different surveys, going back to 1959, as well as half a century of gun sales data, which show that the gun ownership percentage has not changed since 1959. Kleck’s conclusion is obviously the better supported one.

Lott does a better job in his statistical analysis which found the the introduction of concealed carry laws was associated with declines in violent crime rates. That is because this analysis is in his area of expertise, econometrics. Unfortunately, his unfamiliarity with firearms research betrays him when he interprets this result to mean that the laws caused the decrease in crime. Kleck’s book contains details of surveys of gun carrying that show that the number of people that get permits for concealed carry is much less than the number of people who carry illegally, that is, the laws did not make a significant difference to a criminal’s chance of encountering an armed victim. Kleck concludes that the crime decreases were probably caused by some factor other than the carry laws.

There are many more errors of fact and interpretation in Lott’s book, too many to list here.

So why does the pro-gun book by Kleck have a sales rank of 72,000 while the pro-gun book by Lott have a rank of 1264? I think the reason is that Lott goes well beyond what the data supports to claim that more guns cause less crime. Kleck sticks with a position that is supportable by the data – that the bad and good uses of guns mostly cancel out, leaving little net effect on crime. Pro-gun readers would rather hear Lott’s message, even if it’s wrong.

Readers looking for a pro-gun book should buy Kleck’s book, rather than Lott’s. Lott’s book is only useful for those readers who are interested in the details of Lott’s multivariate statistical analysis. –This text refers to the Hardcover edition

Just two days later this other review appeared, pushing my review down the page:

i-6da7393c8ff6de1c28d52199d33a151e-stars-5-0.gifWell-written, important, powerful book, June 12, 2000
Reviewer: A reader from Philadelphia, PA
Here are just a few of the academics who have expressed admiration for Lott’s pathbreaking book. Few people have both the real world law enforcement experience and extenstive research background to take on this explosive issue of guns and crime. This is what the experts in the field think of Lott’s book. (The quotes are from the paperback version of the book.)

“John Lott has done the most extensive, thorough, and sophisticated study we have on the effects of loosening gun control laws.” — Gary Kleck, Professor, Florida State University

[Other quotes omitted]

The book has gotten similar positive comments from those working in law enforcement. This is a great book.

This isn’t as obvious a reply to my negative review as the other two, but the quote from Kleck is meant to rebut the mention of Kleck in my review. Lott used the same tactic in Usenet discussion when Kleck’s conclusions were brought up. This one is from Philadelphia, just like the one from maximcl. The writing style of the previous two is obviously Lott’s, but since this one is mainly quotes there isn’t enough text to judge. However, it is Lott’s style to post a review containing the blurbs, since that is what he did for the first edition. Here is the very first review of all:

i-2e986fbaaece70b42594ca82a7e7a543-stars-0-0.gifComments from the book’s reviewers:, March 10, 1998
Reviewer: john_lott@law.uchicago.edu from Madison, Wisconsin
“This sophisticated analysis yields a well established conclusion that supports the wisdom of the Second Amendment to the United States Constitution rather than of those who would limit the right of law-abiding citizens to own and carry guns. The general reader may find of most interest chapter 7 which documents how far ‘politically correct’ vested interests are willing to go denigrate anyone who dares disagree with them. John Lott has done us all a service by his thorough, thoughtful scholarly approach to a highly controversial issue.”-Milton Friedman

[Other quotes omitted]

Note that for this review Lott did the right thing—he signed his own name to it and told Amazon that it was from the author and hence it does not include a star rating. Would that he had done that for all the others.

I am highly confident that all three of the five-star reviews above are by Lott since there is at least three pieces of evidence pointing towards him for each review—the style, the timing and the location of the reviewer. I found three more reviews that I think are likely also from Lott, though I am not sure. Judge these ones yourself. All three are written in Lott’s style and make the same points that Lott likes to make. The first is from Madison, Wisconsin, the same location as the very first Lott review above. The next two have the tell tale phrase “as an academic” that was used in the maximcl review:

i-6da7393c8ff6de1c28d52199d33a151e-stars-5-0.gifRequired reading for those interested in what causes crime!, August 8, 1998
Reviewer: A reader from Madison, Wisconsin
More Guns, Less Crime is a very readable and thorough examination of the relationship between gun ownership and its deterrent effect on criminal activity. Professor Lott explodes myths about crime that go unquestioned in the press everyday. Should people behave passively when confronted by a criminal? Are people that we know a threat to us? What are the real risks in having a gun in the home? Do guns save more lives than they endanger? This book answers these questions and many, many more. I will never listen to the news media reports on crime the same way again. –This text refers to the Hardcover edition

i-6da7393c8ff6de1c28d52199d33a151e-stars-5-0.gifA MUST BUY! DEMOLISHES GUN CONTROL MYTHS THAT ENDANGER LIVES, July 30, 1999
Reviewer: A reader
Great book. As an academic, with all the garabage research that gets covered by the press, I can’t believe that this book hasn’t gotten more news coverage. If everyone actually read this book, we would have a lot fewer deaths. It is extremely well written and explains to people where the different claims that they hear come from. I have never seen such a careful indepth study of any issue. This guy really sets the standard for research! Despite what might be good intentions (though after reading chapter 7 I have real doubts about their intentions) gun control advocates are endangering people’s lives. The press really needs to think twice about the impact that their newscoverage has on people’s safety. I can understand why bad events get so much news coverage relative to good events, but this lopsided coverage creates some real misimpressions. –This text refers to the Hardcover edition

i-6da7393c8ff6de1c28d52199d33a151e-stars-5-0.gifIf you are interested in the facts, read this book, July 10, 2000
Reviewer: A reader from Miami, Florida

A couple of friends of mine have been nagging me to read this book for a couple of years. When the second edition came out I finally gave in and got it (for $9.60 I couldn’t argue that the price was too high). Anyway, I am only sorry that I didn’t read this book earlier. As an academic and a person who has been somewhat anti-gun, I had two reactions to the book.

1) I was amazed by how much research went into this book. I have never seen so many different data sets been used so comprehensively. State level, county level, and city level data is used, and not just from a few jurisdictions but for the entire country. Just the work that he has done on the impact of police on crime is amazing by itself.

2) The attacks on Lott disgust me. I confess that I have seen some of these claims in the press (about things like whether he accounts for certain factors or not), but one thing is obvious — the point of these attacks is to keep people from even reading the book. It is a high risk strategy because anyone who spends a half hour with this book will realize that those attacking Lott are lying constantly. How amazingly false these attacks surely made me wonder about other things that the media tries to push.

Finally, let me just say this is a book that shows you how research should be done. It is also written in a way that I wish other research was written. It is very clearly written and accessible to a very wide range of readers. To much of what academics write these days is filled with jargon. This book has changed my views when I didn’t think that they could be changed.

Summary: I believe that Lott reviewed his own book at least four times (Mary Rosh, Washingtonian, maximcl, and “a reader from Philadelphia”). It is probable that he reviewed it seven times. Each review gave it five stars.

There will be more tomorrow. If you can’t wait, I’ve given you the secret of finding Lott’s reviews. Think: What Would Lott Do?

Comments

  1. #1 Mark
    December 10, 2003

    Tim, thanks for taking note of my review. I of course saw the later review that blasted mine, but that it might have been written by Lott himself never crossed my mind at the time. I received an email from Lott, under his own name, after my review was posted. Needless to say, it was not a friendly email. I don’t remember all the specific points he made, but it was, unsurprisingly, a very petulant, whiny message. I do recall one point very clearly. In my review, I mentioned Jens Ludwig’s paper showing the differences in effects on adult and juvenile homicides. I had read Ludwig’s paper some time before posting my review, and understood his conclusions quite clearly. Amazingly, Lott denied that Ludwig had reached the conclusion that I mentioned, even though the paper was published long before I posted my review, and even though he had to know I would see through his lie quite easily (my reviewer’s profile at Amazon identifies me as a eollege econ professor).

    So when “Mary Rosh” was exposed earlier this year, I was not surprised in the least.

  2. #2 The...Ummm...Philly Man...I Mean Philly Girl
    December 10, 2003

    Hogwash. John Lott’s writing is as factual as it is provocative and sexy. Oh God it’s sexy, he is such a MAN, I just… ohhhhhhhh…. ahhhhh…. AH!

    For the record, I am an academic.

  3. #3 Tim Lambert
    December 10, 2003

    Do recall when you got the email from Lott? Was it, oh, about two weeks after you posted the review?

  4. #4 SullyWatch
    December 10, 2003

    Lott named his son after a lad’s magazine?

  5. #5 Abashed Chicago Grad
    December 10, 2003

    So why on earth does anyone expect the AEI to kick him out?

  6. #6 Mark
    December 10, 2003

    Tim–I don’t recall the exact date, but it was within a few weeks of my posting the review.

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.