My colleague Russ Roberts writes: Over the years, I have become increasingly skeptical of the power of statistical techniques to measure causation in complex systems. … I happen to believe that concealed handguns do deter crime and allowing concealed handguns is a good thing. And you can claim that the evidence that shows I’m right is "good" statistical analysis. The other side disagrees. They claim it’s "bad" statistical analysis. Who’s right? I have no idea. But what’s clear to me is that my belief in the virtues of allowing concealed hand guns has little to do with the empirical evidence. And I would argue that the opponents are really in the same boat. They just don’t like guns and they’ve dressed up their prejudices in fancy statistical analysis. …
1) If you look across Western Europe or Europe as a whole, there is not the relationship that you imply. Switzerland has a much higher gun ownership rate than the US, but it has one of the lowest murder rates in Europe. Norway and Finland also have gun ownership rates that are just a little less than that in the US. If you read either MGLC or The Bias Against Guns, I have extensive discussions on the problem with purely cross sectional data. Let me note that the countries that currently have low murder rates in Europe tended to have even lower murder rates before gun control. England is a good example. It had no gun control prior to 1920 and yet London, a city of millions of people, had a total of two gun murders in 1900. It had five armed robberies. After stricter gun regulations in the 1950s and 1997, the UK say increases in murder and violent crime. You really need panel data to answer this question properly.
2) If you look around the US, murders are extremely geographically concentrated. 50 percent of US counties have zero murders in any given year. 25 percent have one murder. just over 3 percent of counties account for over 70 percent of all murders. Those 3 percent also happen to be the lowest gun ownership rate counties in the US. Again, though, I wouldn't put a lot of weight on purely cross sectional data.
3) If you look across all countries for which the data is available and control for income, there actually appears to be more crime where there are stricter gun regulations. The top ten countries in terms of murder have either complete or virtually complete gun bans. Even in the past during the 1970s and 1980s in the former USSR, despite the totalitarian state, they had a murder rate much higher than ours in the US. Again, however, panel data is necessary. I have just tried to frame this discussion in terms of the cross sectional data that you reference.
If you are looking for a detailed discussion of these points, please see The Bias Against Guns.
Well, Robin Hanson has pointed out that this is not the time or place to be arguing about (hand)gun control. However, since you have popped up, how about answering my questions? Why does the US have such higher homicide and suicide rates than other high income countries that pretty much all have much stricter gun controls of all sorts? And, is it because we already have so many guns around given our past lack of gun control that your studies on handgun controls in states within the US might be correct?
You apparently didn't look at many of the issues raise in just those two references.
1) The coding errors were in a paper by Plassmann and Whitley, not More Guns, Less Crime. The small number of errors that they argue did not effect their basic results were in data that was added from 1997 to 2000, not in the data that MGLC used from 1977 to 1996. Plassmann has a discussion here. You might also want to see their discussion here.2) As to the survey, see here.
John, the links you provide don't say anything about the alleged errors in your book; at most they indicate that on a couple of occasions Tim Lambert has been overzealous in identifying sockpuppets. Given your history (which is a matter of record) of self-aggrandizing sockpuppetry via "Mary Rosh" over a three-year period, I think that's pardonable; it certainly doesn't seem to me to destroy Tim Lambert's credibility.
And, btw, the "Mary Rosh" comment was mine; I am not Tim Lambert, have never met Tim Lambert, and so far as I can recall have had no interactions with Tim Lambert other than reading his blog from time to time.
I agree with everything John says. And he's so handsome too!
And whatever people say about the errors in John's book, just look at all the good reviews it got on Amazon. How could his book be bad, when he shows his faith in it by reviewing it so many times?
Regarding correlation and causation that is precisely why some research try to provide many qualitatively different empirical tests. For example, with right-to-carry laws: 1) violent crime falls, 2) the size of the drop increases over time the longer the law is in effect because more permits are issued, 3) there are differences between violent crimes where a criminal comes in contact with a victim who might be able to defend herself and a property crime where there is no contact (violent crimes fall relative to property crimes), 4) there are differences between different types of violent crimes (e.g., between murder generally and multiple victim public shootings because the probability that someone will be able to defend themselves with multiple victim public shootings is much higher than the case where there is only one or two victims present), 5) a comparison between adjacent counties on opposite sides of a state border, and 6) differential benefits across different types of victims.
Peter, that doesn't sound at all to me like what Russ said, and I don't yet see the relevant of the Duhem-Quine thesis for the project of overcoming bias. But if you'll blog an argument somewhere, I'll blog a response here.
Isn't Russ just saying that "facts" never speak to us in an unambiguous manner and instead always demand "interpretation" through various theoretical frameworks and respect for the network of statements that are made (both focal and subsidiary) in developing a hypothesis?
What about the relevance of the Duhem-Quine thesis even in the natural sciences, let alone the social sciences?
We are all interested in tracking truth and overcoming unwarranted bias, but stating that as a goal is slightly different from achieving it unambiguously in practice. But rather than go down that path, perhaps it might be better to address the Duhem-Quine thesis and its implications in not only making sense of the problem that Russ identifies, but the project of overcoming bias in general.
I expected Russ' post and Robin's response to set off a flurry of back and forth amongst econ blogs, but I haven't really seen that. I've heard Russ Roberts say that economics is really an art and not a science, which seems like an important criticism. Robin, such a disagreement seems like an elephant in the room of economics, which I would think you would want to boil down to a specific disagreement of facts. Why isn't an active, engaged debate happening? Does nobody take his argument seriously, or does everyone take it as given?
"But it is pretty clear that on a global basis, the countries that have done a better job of controlling the availability and use of guns more generally very strongly tend to have lower homicide, and especially suicide, rates."
It could be that the places people tend not to have guns and also have little cultural tradition of solving personal problems with firearms, it's easier to do gun control. The causation might run from low homicide to gun control, rather than the other way around.
Similarly, there are cultural differences between US cities and states that tend toward gun control versus none. In some places, a rise in crime rates leads to gun control -- whether that results in improvements or not. In other places, a rise in crime rates leads to capital punishment -- whether that results in capital punishment or not.
Since places where they solve problems with gun control differ in important but subtle ways from places where they solve problems with capital punishment, it's hard to make crude statistical comparisons between them.
Robin, physics is a simple theory which is discovered in lab experiments where complexity is managed, and remains valid outside and is applicable outside the lab as well, amidst the vast comlexity. For example, we can be confident that perpetual motion machines will not work. The vast complexity of the outside does not render the simple theory useless. The line that there are so many ways to apply the simple theory to the complex alleged perpetual motion machine is an argument that the crackpot inventor might try to use in order to discourage us from using our intelligence against his claims.
I get the impression that RR was looking referring specifically to cases where intelligent, well informed people have very different opinions. If all the smart people are on one side and only stupid people are on the other, then it's a pretty safe bet that we're right and they're wrong ;-) But if there are intelligent people on both sides, it may be a bit presumptuous to assume that our side formed its beliefs via pure unbiased reason but the other side is largely a product of wishful thinking.
Peter Boettke has a reply:http://austrianeconomists.t...
Dear Rosser:
1) If you look across Western Europe or Europe as a whole, there is not the relationship that you imply. Switzerland has a much higher gun ownership rate than the US, but it has one of the lowest murder rates in Europe. Norway and Finland also have gun ownership rates that are just a little less than that in the US. If you read either MGLC or The Bias Against Guns, I have extensive discussions on the problem with purely cross sectional data. Let me note that the countries that currently have low murder rates in Europe tended to have even lower murder rates before gun control. England is a good example. It had no gun control prior to 1920 and yet London, a city of millions of people, had a total of two gun murders in 1900. It had five armed robberies. After stricter gun regulations in the 1950s and 1997, the UK say increases in murder and violent crime. You really need panel data to answer this question properly.
2) If you look around the US, murders are extremely geographically concentrated. 50 percent of US counties have zero murders in any given year. 25 percent have one murder. just over 3 percent of counties account for over 70 percent of all murders. Those 3 percent also happen to be the lowest gun ownership rate counties in the US. Again, though, I wouldn't put a lot of weight on purely cross sectional data.
3) If you look across all countries for which the data is available and control for income, there actually appears to be more crime where there are stricter gun regulations. The top ten countries in terms of murder have either complete or virtually complete gun bans. Even in the past during the 1970s and 1980s in the former USSR, despite the totalitarian state, they had a murder rate much higher than ours in the US. Again, however, panel data is necessary. I have just tried to frame this discussion in terms of the cross sectional data that you reference.
If you are looking for a detailed discussion of these points, please see The Bias Against Guns.
John,
Well, Robin Hanson has pointed out that this is not the time or place to be arguing about (hand)gun control. However, since you have popped up, how about answering my questions? Why does the US have such higher homicide and suicide rates than other high income countries that pretty much all have much stricter gun controls of all sorts? And, is it because we already have so many guns around given our past lack of gun control that your studies on handgun controls in states within the US might be correct?
Dear "g":
You apparently didn't look at many of the issues raise in just those two references.
1) The coding errors were in a paper by Plassmann and Whitley, not More Guns, Less Crime. The small number of errors that they argue did not effect their basic results were in data that was added from 1997 to 2000, not in the data that MGLC used from 1977 to 1996. Plassmann has a discussion here. You might also want to see their discussion here.2) As to the survey, see here.
John, the links you provide don't say anything about the alleged errors in your book; at most they indicate that on a couple of occasions Tim Lambert has been overzealous in identifying sockpuppets. Given your history (which is a matter of record) of self-aggrandizing sockpuppetry via "Mary Rosh" over a three-year period, I think that's pardonable; it certainly doesn't seem to me to destroy Tim Lambert's credibility.
And, btw, the "Mary Rosh" comment was mine; I am not Tim Lambert, have never met Tim Lambert, and so far as I can recall have had no interactions with Tim Lambert other than reading his blog from time to time.
Is that Tim? In any case, the inaccuracy of the source that you reference is discussed here and here.
I agree with everything John says. And he's so handsome too!
And whatever people say about the errors in John's book, just look at all the good reviews it got on Amazon. How could his book be bad, when he shows his faith in it by reviewing it so many times?
Dear J Thomas:
Regarding correlation and causation that is precisely why some research try to provide many qualitatively different empirical tests. For example, with right-to-carry laws: 1) violent crime falls, 2) the size of the drop increases over time the longer the law is in effect because more permits are issued, 3) there are differences between violent crimes where a criminal comes in contact with a victim who might be able to defend herself and a property crime where there is no contact (violent crimes fall relative to property crimes), 4) there are differences between different types of violent crimes (e.g., between murder generally and multiple victim public shootings because the probability that someone will be able to defend themselves with multiple victim public shootings is much higher than the case where there is only one or two victims present), 5) a comparison between adjacent counties on opposite sides of a state border, and 6) differential benefits across different types of victims.
Peter, that doesn't sound at all to me like what Russ said, and I don't yet see the relevant of the Duhem-Quine thesis for the project of overcoming bias. But if you'll blog an argument somewhere, I'll blog a response here.
Robin and others,
Isn't Russ just saying that "facts" never speak to us in an unambiguous manner and instead always demand "interpretation" through various theoretical frameworks and respect for the network of statements that are made (both focal and subsidiary) in developing a hypothesis?
What about the relevance of the Duhem-Quine thesis even in the natural sciences, let alone the social sciences?
We are all interested in tracking truth and overcoming unwarranted bias, but stating that as a goal is slightly different from achieving it unambiguously in practice. But rather than go down that path, perhaps it might be better to address the Duhem-Quine thesis and its implications in not only making sense of the problem that Russ identifies, but the project of overcoming bias in general.
Pete
I expected Russ' post and Robin's response to set off a flurry of back and forth amongst econ blogs, but I haven't really seen that. I've heard Russ Roberts say that economics is really an art and not a science, which seems like an important criticism. Robin, such a disagreement seems like an elephant in the room of economics, which I would think you would want to boil down to a specific disagreement of facts. Why isn't an active, engaged debate happening? Does nobody take his argument seriously, or does everyone take it as given?
Robin, I'm arguing about correlation versus causation. About the difficulty of doing statistics across populations that have systematic variation.
J and others, this isn't the place to argue about gun control.
"But it is pretty clear that on a global basis, the countries that have done a better job of controlling the availability and use of guns more generally very strongly tend to have lower homicide, and especially suicide, rates."
It could be that the places people tend not to have guns and also have little cultural tradition of solving personal problems with firearms, it's easier to do gun control. The causation might run from low homicide to gun control, rather than the other way around.
Similarly, there are cultural differences between US cities and states that tend toward gun control versus none. In some places, a rise in crime rates leads to gun control -- whether that results in improvements or not. In other places, a rise in crime rates leads to capital punishment -- whether that results in capital punishment or not.
Since places where they solve problems with gun control differ in important but subtle ways from places where they solve problems with capital punishment, it's hard to make crude statistical comparisons between them.
Robin, physics is a simple theory which is discovered in lab experiments where complexity is managed, and remains valid outside and is applicable outside the lab as well, amidst the vast comlexity. For example, we can be confident that perpetual motion machines will not work. The vast complexity of the outside does not render the simple theory useless. The line that there are so many ways to apply the simple theory to the complex alleged perpetual motion machine is an argument that the crackpot inventor might try to use in order to discourage us from using our intelligence against his claims.
I get the impression that RR was looking referring specifically to cases where intelligent, well informed people have very different opinions. If all the smart people are on one side and only stupid people are on the other, then it's a pretty safe bet that we're right and they're wrong ;-) But if there are intelligent people on both sides, it may be a bit presumptuous to assume that our side formed its beliefs via pure unbiased reason but the other side is largely a product of wishful thinking.