Cook and Ludwig keep falsely claiming that their 1997 NIJ uses data from 1993 and 1994
If you look at the data you will see that the survey was from November 1991 to November/December 1994, not 1993 and 1994 as Cook and Ludwig keep claiming. If you go through the data, you will see you only get their sample size and numbers that match theirs if you go back for three years. Note also that people are asked about purchases over the preceding year so questions asked in November 1994 cover November of each preceding year.
You might think that this is pedantic, but more than 40 percent of their data is obtained during the November 1991 to November/December 1992 period. It has disappointed me that Cook and Ludwig keep repeating this false claim.
The longer the time period not only places more of the purchases before the February 28, 1994 Brady Act but it also creates the type of biases that I discussed regarding the Washington Post survey for Maryland. One can't even just assume that purchases are uniformly made over this period. The problem is that sales were massive before the Brady Act went into effect and then plummeted right after the law went into effect. As to the biases, the buying of a gun is probably a much more memorable event than whether there was a background check. If so, people will remember the purchases, but not remember whether there was a background check that went along with it. You will see a discrete jump in how important this bias is once you ask people about more than just events over the previous year.
Cook and Ludwig redo should get a top figure of 16%, not 22% that Cook and Ludwig claim.
Labels: background checks