I want to make it clear right off the bat that I like the NY Times, grew up reading it, and my day isn't complete without it (especially the crossword). It is true that its editorial page leans somewhat left but most of what's printed is more factual than not. I do like Krugman's columns because his clarity on the morass of national and global economics is good and sometimes brilliant - there's a reason he got a Nobel Prize for economics. But every now and then, they blow it on the editorial page. Feb. 1, 2013, was one of those occasions.

The editorial was titled "Myths About Gun Regulation." (http://www.nytimes.com/2013/02/01/opinion/myths-about-gun-regulation.html?nl=todaysheadlines&emc=edit_th_20130201). And here's the statement where they screwed up: A long–range, independent study issued as Congress allowed the ban to expire in 2004 found criminal use of assault weapons had fallen by one-third or more as a share of gun crimes in major jurisdictions.

To their credit, they cited the study where this came from. But wait! I went and read the study. You can find it at: http://www.sas.upenn.edu/jerrylee/research/aw_final2004.pdf

It's a really interesting piece of work and it is both highly nuanced and not at all positive that the assault weapons ban of 1994 had an appreciable effect on violent crimes with guns. Yes, using their methodology, they did find that criminal assault weapon use fell by about a third (see p.55). But the way the NY Times phrased this result both obscured that this drop may not have been statistically significant - which is a form of lying about statistics - and that the way they developed their data might not have real correlation between pre- and post-ban gun use in crime.

The problem with their analysis is that there are no good nation-wide studies on specifically assault weapon used in crimes. There is not now nor has there ever been any uniform data gathering on crime rates involving specifically assault weapons. Congress has been an obstruction for providing support for such an effort despite calls for assault weapon stats for almost 20 years now. Because of the lack of such data, the study under discussion here used surrogates to estimate assault weapon in violent crimes. These surrogate measures included annual gross production of assault weapons, assault weapon prices, and law enforcement gun tracing requests sent to the ATF.

To their credit, the authors of this study admitted that there may no correlation between the above surrogate trends they used and trends in criminal assault weapon use. In fact, despite tha ban, the authors noted that If anything, therefore, gun attacks appear to have been more lethal and injurious since the ban.

They authors used the assumptions that rising availability had an positive effect on assault weapon usage and that rising prices had a negative effect. They honestly admitted to arm waving in assuming a correlation between gun tracing statistics and assault weapon usage. They felt there was a correlation because the surrogate trends they used appeared to fulfill their assumptions, especially when their surrogate trends appeared to correlate with assault weapon data from the few jurisdictions that actually tracked assault weapon crime statistics. The number of jurisdictions that did track assault weapon crimes can be counted on the fingers of one hand; however, all of them showed decreases in pre-ban to post-ban usage. But here's what the NY Times didn't tell you: the numbers of assault weapons crimes was so small that the observed trends may not have had statistical significance. In the jursdictions the study looked at, the typical percentage of assault weapon crimes relative to all firearm crimes was less than 3% before the 1994 ban and fell to less than 2% after the ban. The study noted that these numbers are so small that they may be meaningless when you consider that approximately two-thirds (plus or minus 5%) of all murders involve some kind of firearm - a statistic that has been stable for decades.

The NY Times editorial implied that the study under discussion demonstrated a real drop in assault weapon crime during the assault weapon ban of 1994-2004. But that's not what the study itself concluded. Here's what this study actually said:

Although the ban has been successful in reducing crimes with assault weapons, any benefits from this reduction are likely to have been outweighed by steady or rising use of non-banned semiautomatics with large capacity magazines, which are used in crime much more frequently than assault weapons. Therefore, we cannot clearly credit the ban with any of the nation’s recent drop in gun violence. And, indeed, there has been no discernible reduction in the lethality and injuriousness of gun violence, based on indicators like the percentage of gun crimes resulting in death or the share of gunfire incidents resulting in injury, as we might have expected had the ban reduced crimes with both assault weapons and large capacity magazines.

I have to wonder if the NY Times editorial staff is just plain inept with statistics or if a pro-gun-control bias was involved in ignoring the study's conclusions. I expected better reporting from the premier newspaper in the US.