A new article has recently appeared in a peer-reviewed journal on the issue of whether "shall issue" right-to-carry concealed weapons laws (reqwuiring authorities to issue concealed-weapons weapons to anyone who applies without a criminal record or history of mental illness). The article, found here, concludes that such laws are generally beneficial.
The article, written by Carlisle Moody and Thomas Marvell, rebuts the 2003 article in the Stanford Law Review by Ian Ayres and John Donohue. Ayres and Donohue found (contrary to the seminal work of John Lott and David Mustard) that shall-issue laws actual lead to an overall increase in crime. Here is how Moody and Marvell describe their findings:
While reading Ayres and Donohue’s 2003 article in the Stanford Law Review, we noticed that their analysis did not prove what they said it proved. They claimed that their model proved that shall-issue laws increased crime. Our conclusions are as follows.
Many articles have been published finding that shall-issue laws reduce crime. Only one article, by Ayres and Donohue who employ a model that combines a dummy variable with a post-law trend, claims to find that shall-issue laws increase crime. However, the only way that they can produce the result that shall-issue laws increase crime is to confine the span of analysis to five years. We show, using their own estimates, that if they had extended their analysis by one more year, they would have concluded that these laws reduce crime. Since most states with shall-issue laws have had these laws on the books for more than five years, and the law will presumably remain on the books for some time, the only relevant analysis extends beyond five years.
We extend their analysis by adding three more years of data, control for the effects of crack cocaine, control for dynamic effects, and correct the standard errors for clustering. We find that there is an initial increase in crime due to passage of the shall-issue law that is dwarfed over time by the decrease in crime associated with the post-law trend. These results are very similar to those of Ayres and Donohue, properly interpreted.
Moody and Marvell's findings seem plausible to me.
Related Posts (on one page):
- Update on the More Guns, Less Crime Debate
- Rebuttal of Ayres and Donohue Claim of "More Guns, More Crime"
As I read the article, it found that shall carry laws were associated with more assault but less burglary. Of course, that is exactly opposite to what theory should suggest. Weapon carrying should deter assault but have no effect on burglary (where by definition no one is present).
I'm a big fan of statistical analyses, but it looks to me like these laws have fairly little effect on crime (positive or negative), and the studies both ways are just picking up significance because they have large ns.
Moron Alert!!
The usual statistics for determining whether a given model is valid are only meaningful when the model is framed before the data is examined. If you get to the point of "let's throw in a dummy variable and see what happens", then the whole intellectual structure of the hypothesis test falls apart. It's not unusual to be able to change a model slightly and have the sigh of a coefficient change sign, i.e. a positive effect becomes a negative effect, or vice versa.
Common and understandable as it is, choosing the model that gives the result you want is intellectually dishonest. Hence, "lies, damn lies, and statistics."
I think our society is too data-oriented. We tend to think you can sift out the fourth or fifth most important influence out of a mass of population data. Usually, you can't.
Object to the use of the word "seminal" on the grounds that (1) the sense that the paper has not generated discussion beyond a modicum of debate about whether the particular conclusions of the paper are correct, which is fine but doesn't warrant referring to the paper as "seminal"; and (2) yuck.
There's a certain crowd, identifiable by a common political persuasion, that predictably tars those who disagree as not just wrong but morally suspect. The attack on Chagnon (accusing him of no less than genocide, yes I am serious) is another example of this happening. Lott got off easy.
And taking off from what SeaDrive said, if the effects are robust, we should be able to see them regardless of what model we choose, whether we look at 4 years, 5 years, 6 years, 10 years, etc. While cherrypicking one's model can be a sign of dishonesty, it's more of a sign of weakness.
That may be, but that doesn't mean they're wrong in this case, it just means we should be skeptical of their claims.
What I've read has been pretty convincing, with Lott's most egregious action being that he probably faked or lied about a study he conducted, and definitely made mathematically incorrect statements about it afterward.
Do you have reason to dispute Tim Lambert's claims? He was pretty thorough about addressing each specific point he raises. Also, Michelle Malkin is an example of a pro-gun conservative who has called Lott out on these shenanigans.
Also, although the Ayers and Donohue findings do not show that crime rates went down after shall-issue laws were enacted they also do not show that they go up over the longer term. Their method only showed a barely statistically significant and slight initial increase and then a leveling off. In other words, no significant increase in crime rates.
Even if the Ayers and Donohue methodology is the more proper one, their results still support making shall-issue laws the norm everywhere since they are consistent with greater freedom and greater personal security without significantly increasing criminal activity.
And your point is what? Lott has been attacked on those grounds, but the attacks appear much overblown. He probably did engage in sloppy analysis -- which appears not uncommon in academe, even in peer reviewed articles published in the best journals in subjects involving science and medicine (where more objectivity exists). See, e.g.
Whatever one thinks of John Lott, this need not be an issue of ad hominems. I'm a little troubled that his extensive body of empirical research very consistently seems to support relatively controversial conservative policy positions. My research experience seems to produce outcomes that are sometimes supportive of the conservative and sometimes the liberal position. But any work should ultimately be judged on its merits.
Pardon the echo, but 'Hear, hear!'. This point is not made nearly enough (IANAL, but I majored in statistics).
As Mr. Nieporent says, it is possible to bound the effect as small. That is not surprising given that the percentage of people carrying at any given time is very small.
And I think that is the point - liberal carry laws won't cure crime in general, or cause main street to become the OK Corral. They can make a huge difference for the stalking victim, the elderly forced by economics to live in a bad neighborhood, etc., and don't have any measurable societal downside.
BobDoyle :
TNeloms, I think you have grossly misrepresented what Jim Lindgren said on the subject. Lindgren's position as reported in the Chronicle of Higher Education:
Mr. Lindgren adds that he believes it extremely unlikely that any coding errors were the result of a conscious intent to distort the study's findings. He notes that Mr. Lott has not only shared his data sets with other scholars, but has made them generally available to the public on his Web site. "You tend not to do that if you've intentionally miscoded your variables," he says.
</blockquote>
I couldn't find that article, but what was the context of that quote? It seems to address Lott's supposed coding errors, and specifically whether they were intentional. Lott himself admitted that there were coding errors (http://johnrlott.tripod.com/attack.html) and corrected them.
I was referring to claims that Lott made up a survey, and I was genuinely asking a question ("Didn't Lindgren show...") because I heard about this but didn't know the details.
I looked up some details now, and Lindgren's final position seems to be inconclusive, but he found surprisingly little evidence that the survey was ever performed. I think this is the most recent recap of his position: http://volokh.com/archives/ archive_2006_06_11-2006_06_17.shtml#1150212677
Absence of evidence is at best weak evidence of absence.
And, indeed, the statistical analysis should perhaps be done by state, or by city vs. rural.
That's not true. If I claim that a bomb just went off in the building, but there is no evidence of any explosion, that is very strong evidence that my claim is false.
In general, if you expect there to be a large amount of evidence from an event, absence of evidence is strong evidence of absence of the event.
In this case, Lindgren says "I had never heard of a professor doing anything of that size with no funding, paid support, paid staff, phone reimbursements, or records (though there are probably precedents for such an unusual method)." To me, that is medium evidence of absence.
I agree that one should judge any work on its merits, and More Guns Less Crime has been criticized by many on its merits. That's why I cited Lott's own admission that there were coding errors above.
But the reason this may be an issue of ad hominems as well is that Lott's credibility has been seriously challenged, both concerning fabricating data and concerning making honest attempts to interpret data. It is not easy to discern those two things just from reading a work itself. I think it's fair to say that if a researcher has a history of fabrication and misinterpretation (which you can argue that Lott does not), then one should take new works with a grain of salt, even if you can't find anything specifically wrong with them.
Mexico has a handgun ban.
How is it affecting the drug lords there?
However.
I attended a weeklong seminar for a non profit and we hosted many historical martial artists. These are people who know everthing about spanish rapier, or german broadswoard and dagger,or 19th century dueling. I learned a lot from talking with them.
We all know that the 18th and 19th centuries were noted for politeness and manners. Why is that? Turns out duels had a lot to do with it. If I said anything that might offend you, you might challenge me to a duel. Or if I behaved improperly towards your sister or wife, or somehow challenged your honor, I might find myself in a duel.
Once challenged, I could not back off. that itself would bring dishonor to myself. Everyone knew the rules of duels, even if they didn't actually engage in them. As a result, it was, as Napolean said, "a dueling society is a polite society."
Now, that didn't mean that because everyone carried swords on their person, fights would break out all the time. they were actually fairly rare. (Also, the late 18th century in France, a law prohibited anyone from carrying a sword except for the Hussars. Therefore, everyone started carrying a cane. And as I learned, a cane was every good a defense as a sword, providing you knew how to use it.)
So, we have a situation when arms were banned,and everyone just switched to a cane, which on the surface was a mere fashion statement, but in reality was the weapon of choice. Also, when you might be challenged for your inappropriate actions or words, you watched what you said. you were, naturally, as polite as can be.
Now, I haven't totally reversed my thinking. People were thoroughly trained inthe use of rapier or cane, or dagger. There was a code of honor in place. Dueling had its rules. Few people were actually carrying around guns even then. So many differences, so I don't know how or if we can make any comparisons.
Still, it gives me pause,and it puts the 2nd amendment in much better context for me. In fact, it strengthens my argument that arms were considered more swords than they were guns back in the 18th century, and that the framers didn't believe that people would be carrying guns all the time. Swords, yes, guns, no.
But still.....it just goes to show, the more you know, the more confused you are!
I'd rather live in a rude society myself.
2) The coding errors were in a paper by Plassmann and WHitley. The errros involving the data that they added on to my original study from 1977 to 1996 in the second edition of More Guns, Less Crime where for their new data from 1997 to 2000. The errors were small (180 cells out of 7.5 million) and did not alter their main results at all. Plassmann's discussion can be found here. Plassmann and Whitley do thank me in their paper for my help, but the paper was their paper.
3) The Plassmann and Whitley paper does an excellent job in explaining that there is no temporary initial increase in crime after the law. If you look at the year by year changes in crime rates in either their figures or Ayres and Donohue's figures, there is no initial increase. It only looks like their might be an increase when one fits a straight line to nonlinear data.
4) I guess that I am amazed that anyone gives credibility to Lambert's claims. For discussions on false claims that he has made see here, here, here, and here. These links are from a site put together by Professor Jim Purtilo.
Is there evidence this never existed, or just that John Lott is not willing to travel a thousand miles back to Chicago in the hopes of finding it and force some internet critics move on to another line of personal attack.
Where would it be if he did go back to look? Would not the students be long gone by now? My understanding (If I remember correctly) is that this is about one tiny line in a large book having nothing to do with the main data. It is a claim about what % of self defence cases the gun is fired.
Questions: What was John Lott's claim? Is this claim countered by other surveys measuring the same statistic? Has anyone tried to do an independent survey and compare the result to Lott's reported result? If so, is the result consistent with or counter to the point Lott was trying to make by quoting the original statistic?
You miscited me. See my post above. I have not had coding errors in any paper that I have published nor in More Guns, Less Crime (in either edition). If you have evidence to the contrary please provide it. There were coding errors in a data set that I had, as a favor to others, place on my website, but that was corrected as soon as I became aware of the problem. The Ayres and Donohue work, however, has been filled with errors in their multiple papers, errors that did alter their results, and they have yet to admit publicly to them.
1) Most basic is that the survey results were replicated. That data has been made readily available to others. The survey questions have long been made available to others also.
2) I used the survey results that were biased against the point that I was making. My point was that defensive gun uses only get media attention only when the criminals are killed or wounded. The more that you think that the defensive gun uses involved brandishing or even warning shots the less likely they are to be newsworthy. The more common you think brandishing is, the less that you have to rely on media bias to explain the lack of news coverage. Since I believe that there is media bias, this is biased against the result that I believe to be correct.
3) If you look at the Kleck and Gertz paper, you will find that over 90 percent of their cases involved brandishing or warning shots. Their are only four surveys that are on point to this issue over the last 20 years. Two by me and two that Kleck was involved in. Generally, the difference in the results are in at least part explained by the fact that I ask people to recall events over the last year and Kleck's surveys ask people to recall events over the preceding five years. People tend to not mention more minor events over that longer time horizon.
3) Two people who were surveyed came forward and have been interviewed. Others such as David Mustard and John Whitley have also made statements that support me. Many others have confirmed other points.
What armed people do, however, are change the outcome of crimes. In a report from the National Crime Victimization Survey, it was shown that
While you hear of stories of gang bangers knocking each other out of the gene pool all the time, there are literally hundreds of stories a week of people defending their lives and property with a firearm.
So, while crime may not change at all, that is not the end all be all argument to use for liberalizing gun ownership laws. It is quite easy to see that legitimate gun ownership poses no threat to the populace and actually provides a benefit.
As to the fear of more guns = more criminals with guns, consider that there are already 270,000,000 firearms in the hands of civilians. That's 9 guns for every 10 people (I'm doing my best to increase that number personally ;) If simply increasing the number of guns actually caused crime, we'd have wiped ourselves off the globe already.
I remember more about a poll I took of classmates for the 1968 election (when I was in something like sixth grade) than Lott does about this so-called survey he "supervised".
The salience of this study, which was conducted in the same room as Bellesiles' perusal of nineteenth-century San Francisco probate records is that Lott's strident refusal to admit the obvious and move on suggests to me that he is more stubborn than honest.
Quibbling over my expression. It's a standard point (absence of evidence is not evidence of absence). I am aware that it is not an absolute which is why I tried to soften it, as you can see. So I didn't soften it to your satisfaction. How about: typically, absence of evidence is not very good evidence of absence. Or is that not good enough?
So, no, you don't get to brush aside a major point against your argument, merely because somebody declares that in this case in his personal assessment the absence of evidence actually is good evidence of absence.
The world is not as neat and tidy as a lot of people seem to think. This is what lets 9/11 truthers and UFO crazies convince themselves and win converts - there is enough messiness in the world for people to find, here and there, a narrow line of inquiry that produces extremely compelling evidence (compelling in its own narrow way) that leads to utterly false conclusions, and you can put yourself back right only when you step back and look at the bigger picture, by stepping outside of that narrow line of inquiry. A narrow line of inquiry is what you have here. A commenter has pointed out that Lott has an extensive body of empirical research. If he were a dishonest researcher, we should not have to rely on such a narrow line of inquiry, especially one based on an absence of evidence and somebody's declaration that here is an exception to the rule that it's not evidence of absence. The narrowness of the evidence and the oddity of its type (it is absence of evidence) should put your bullshit detector into high alert.
I have an extremely hard time remembering the names of anyone in my college years aside from close personal friends. I have an even harder time remembering the names of my professors. I have, I have noticed, a surprisingly bad memory in various respects. Nevertheless, it is genuinely bad.
I remember more about a poll I took of classmates for the 1968 election (when I was in something like sixth grade) than Lott does about this so-called survey he "supervised".
How nice for you. I have a friend who remembers every phone number that anyone he ever knew ever had. He has a really good head for that sort of thing. I have trouble remembering my own phone number. Took me half a year to memorize my own cell phone number.
We're not all you.
Perhaps that was the height of your career? In Lott’s case this survey would probably not rank in the top 50 of his biggest projects from then to now. Face it, his response in the link he provided is more than adequate.
I am glad you mentioned Bellesiles, because that is all these attacks on Lott are really about. I followed the Bellesiles situation from the start. He fabricated his entire book, the core of his research. His anti-gun defenders were humiliated, desperate for payback. That is when these attacks on Lott picked up steam. It is pathetic how weak the case against Lott is. He has more than adequately responded to the charges in his link in these comments.
Lott if anything raised the bar in gun research dramatically. He took it from what was an area full of junk political fake research (Kellerman “42 times” with secret data, measuring benefit of guns only by tallying dead criminals, JAMA, CDC, etc.) and took it to a new level huge of data sets, modern statistics, and data open for inspection and replication. As far the survey, he has the word of coauthors and colleagues at the time that his hard drive really did crash and data was lost. Bellesiles was defended to the bitter end until he was proven to have lied so strongly and in so many ways that it could not be denied. With Lott the standard for some seems to be: its kind of possible he might have lied about this little side issue (even thought there would be no motive for him to do so) so lets just assume he did (even though later surveys replicate the lost one and back up his original claim).
Lott would claim to have done a large national survey without discussing the sampling design with anyone, leaving any financial or other records of the study, or remembering anyone who had worked on it. I had never heard of a professor doing anything of that size with no funding, paid support, paid staff, phone reimbursements, or records.
If you click through, you see that the story is even more damning: Lott confabulated this study to explain away misunderstanding one of his sources. That is, the claim he originally attributed to other studies, but which was not backed up by them, migrated into a study of his own that left no trace, literally none, except the spoor-ious results he had already used.
I wouldn't think of using Bellesisles, or Ward Churchill, as a source. I won't be paying attention to Herman Rosenblat (apple hoax) on the Holocaust. That doesn't mean I'm pro-German, but that I have standards. I'd suggest the pro-gun side adopt a similar quality assurance program (and in some cases, they surely have).
It's this thing some of us call "freedom". Maybe you've heard of it.
Didn't several states argue that during the 1860s?
And contrary to his claim, the coding errors made a huge difference to his results. When Ayres and Donohue corrected the errors and reran the regressions, Lott's results went away -- there was no significant link between carry laws and crime.
This isn't true and is not the correct comparison with what Lott claimed to found in his survey. Lott has added "warning shots" to brandishing to make Kleck's numbers closer but it doesn't make them "over 90 percent" (It's 84% if you check.) But the correct comparison is brandishing (Kleck got 76%) with brandishing (Lott claims that 98% of the 25 gun users in his simple just brandished the gun).
Nor can this difference be explained by the different time frame for Kleck, since if you just look at the people whose defensive gun use was in the previous year, you still only get 76% brandishing from Kleck's survey.
1. A burglar always has to risk walking in on an occupied residence. Yes, maybe he rang the doorbell first, but some people may work nights, sleep days, and ignore a doorbell. They don't ignore the door frame cracking.
2. There a lot of people who would not buy or own a handgun except because they plan to carry it away from their home. If permits are effectively impossible to get for ordinary people, then why buy a handgun?
3. You can now get a permit to carry concealed. A handgun starts to make more sense.
4. At least some small fraction of burglars are smart enough to figure out that if they break into a home where someone is home--and has a handgun--the day could end very badly.
I agree that the net benefit of non-discretionary concealed carry laws is probably relatively small. But there's no shortage of people who, because of these laws, have been able to defend themselves from murder, robbery, and rape. See here for the concealed carry permit holders whose gun defenses have made it onto the Civilian Gun Self-Defense Blog. It's not all that rare.
You must subscribe to view the Chronicle of Higher Education article from which I quoted, but you can view the relevant section at here along with additional commentary by Glenn Reynolds and links to other discussions of the topics.
If you have a comment about spelling, typos, or format errors, please e-mail the poster directly rather than posting a comment.
Comment Policy: We reserve the right to edit or delete comments, and in extreme cases to ban commenters, at our discretion. Comments must be relevant and civil (and, especially, free of name-calling). We think of comment threads like dinner parties at our homes. If you make the party unpleasant for us or for others, we'd rather you went elsewhere. We're happy to see a wide range of viewpoints, but we want all of them to be expressed as politely as possible.
We realize that such a comment policy can never be evenly enforced, because we can't possibly monitor every comment equally well. Hundreds of comments are posted every day here, and we don't read them all. Those we read, we read with different degrees of attention, and in different moods. We try to be fair, but we make no promises.
And remember, it's a big Internet. If you think we were mistaken in removing your post (or, in extreme cases, in removing you) -- or if you prefer a more free-for-all approach -- there are surely plenty of ways you can still get your views out.