I was reading a sourcewatch.org article--a liberal think tank--which argued John Lott (founder of the more guns less crime thesis) has coding errors in his work [1]. Interestingly, this merely shows how much liberals need to twist evidence on the gun control issue. First, lets examine the claim.
Source Watch got their data from Tim Lambert. Tim Lambert got is Data from Aryes and Donahue 2003. First, the study is attacking Plassman and Whitley 2003, not Lott's work. Lott did co-author the study, but removed his name because the law review offered an ultimatum which he did not agree with. Lambert, along with Aryes and Donahue argue once these coding errors are fixed, the results disappear. This claim is far fetched and extremely weak. For a few reasons.
1) Donahue and Aryes argue coding errors show the study is flawed. The reason they are so hostile is because that very study argues their first review is weak and does not give evidence for their assertion. So instead of defending their flawed study, they argue their data has coding errors (something which would get media attention and sway the public on the issue). Plassman responded to their study, however. He shows their Hybrid model--the one which Aryes and Donahue put emphasis on as it shows it increases crime--is flawed. In the data, there is a curved line. Before the law, crime rises. After the law, it flat lines and falls. Look at this:
See source #2
As we can see, the crime rate shows conceal carry lowers crime. But Aryes and Donahue insert these straight lines. Using the top line (as they did) it would appear in the data that crime actually rose, then fell. This, they argue, shows CCW raises crime and then other factors make it have no effect. However, as we can see, the facts do not support this position and the results are artificially created by this line (which is not the actual data).
Aryes and Donahue do not even respond to the criticism. As Plassman notes, "Ayres and Donohue do not even acknowledge this problem that plagues all of their analyses. They also decided not to cite Nic Tideman’s and my paper in their footnote 3 (p.1197) among the papers that are supportive of the “More Guns Less Crime” thesis, even though our paper appeared in the same issue of the Journal of Law and Economics as several other papers that they cite. Any econometrician will agree that this ought to be one of the most relevant issues in the whole debate. However, the public debate (as well as Ayres and Donohue in their response to our paper) decided to completely ignore this issue. In short, the fact that the extended dataset had errors does not provide card blanche to dismiss those comments that are not based on the extended data set. Even if one decides to completely ignore the results of the extended data set that we report on pp.1336-1357, it is still necessary to acknowledge (and address) our criticism of Ayres and Donohue’s own analysis."[3]
In other words, Aryes and Donahue essentially concede their study is flawed and use the coding errors as an excuse to try and cover up their weak study. So, assume this:
a) Aryes and Donahue are right, coding errors invalidate Plassman and Whitley's paper.
b) Even if it iinvalidates their data, they conceded their study is flawed and their analysis is wrong.
Outcome: Both studies are flawed, meaning neither side has ammo on this position.
So, in reality, this invalidates (already) the assertion that CCW increases crime. But is Plassman's and Whitley's paper really destroyed by these codinbg errors?
2) The answer is no, negative. The results do change some, however the more guns less crime hypothesis still holds as decreases in crime still exist after the errors are fixed. John Lott showed in "More Guns, Less Crime" (third edition, 2010) that these errors did not erase the results and that the few errors where not vital the the hypothesis. The notion that it erases the results is a lie.
1. http://www.sourcewatch.org/index.php?title=John_Lott
2. Florenz Plassman, John Whitley. "Confirming More Guns Less Crime" Stanford Law Review, 2003.
3. http://johnrlott.tripod.com/link3.html
Source Watch got their data from Tim Lambert. Tim Lambert got is Data from Aryes and Donahue 2003. First, the study is attacking Plassman and Whitley 2003, not Lott's work. Lott did co-author the study, but removed his name because the law review offered an ultimatum which he did not agree with. Lambert, along with Aryes and Donahue argue once these coding errors are fixed, the results disappear. This claim is far fetched and extremely weak. For a few reasons.
1) Donahue and Aryes argue coding errors show the study is flawed. The reason they are so hostile is because that very study argues their first review is weak and does not give evidence for their assertion. So instead of defending their flawed study, they argue their data has coding errors (something which would get media attention and sway the public on the issue). Plassman responded to their study, however. He shows their Hybrid model--the one which Aryes and Donahue put emphasis on as it shows it increases crime--is flawed. In the data, there is a curved line. Before the law, crime rises. After the law, it flat lines and falls. Look at this:
See source #2
As we can see, the crime rate shows conceal carry lowers crime. But Aryes and Donahue insert these straight lines. Using the top line (as they did) it would appear in the data that crime actually rose, then fell. This, they argue, shows CCW raises crime and then other factors make it have no effect. However, as we can see, the facts do not support this position and the results are artificially created by this line (which is not the actual data).
Aryes and Donahue do not even respond to the criticism. As Plassman notes, "Ayres and Donohue do not even acknowledge this problem that plagues all of their analyses. They also decided not to cite Nic Tideman’s and my paper in their footnote 3 (p.1197) among the papers that are supportive of the “More Guns Less Crime” thesis, even though our paper appeared in the same issue of the Journal of Law and Economics as several other papers that they cite. Any econometrician will agree that this ought to be one of the most relevant issues in the whole debate. However, the public debate (as well as Ayres and Donohue in their response to our paper) decided to completely ignore this issue. In short, the fact that the extended dataset had errors does not provide card blanche to dismiss those comments that are not based on the extended data set. Even if one decides to completely ignore the results of the extended data set that we report on pp.1336-1357, it is still necessary to acknowledge (and address) our criticism of Ayres and Donohue’s own analysis."[3]
In other words, Aryes and Donahue essentially concede their study is flawed and use the coding errors as an excuse to try and cover up their weak study. So, assume this:
a) Aryes and Donahue are right, coding errors invalidate Plassman and Whitley's paper.
b) Even if it iinvalidates their data, they conceded their study is flawed and their analysis is wrong.
Outcome: Both studies are flawed, meaning neither side has ammo on this position.
So, in reality, this invalidates (already) the assertion that CCW increases crime. But is Plassman's and Whitley's paper really destroyed by these codinbg errors?
2) The answer is no, negative. The results do change some, however the more guns less crime hypothesis still holds as decreases in crime still exist after the errors are fixed. John Lott showed in "More Guns, Less Crime" (third edition, 2010) that these errors did not erase the results and that the few errors where not vital the the hypothesis. The notion that it erases the results is a lie.
1. http://www.sourcewatch.org/index.php?title=John_Lott
2. Florenz Plassman, John Whitley. "Confirming More Guns Less Crime" Stanford Law Review, 2003.
3. http://johnrlott.tripod.com/link3.html
No comments:
Post a Comment