In Looking at the ATF's 1998 "Crime Gun Trace Analysis Report" (CGTAR) under the Youth Crime Gun Interdiction Initiative (YCGII), I was struck by the strange nature of some of the ATF calculations and terminology. So, I did a few calculations for myself and this note is to document what was found. The intent is to insure an honest report is made. The ATF generally limits traces to guns bought since 1990. They take as an indicator (for investigations) that a trace gun was trafficked (i.e., a straw purchase was made) if the time between the first retail sale and the time the gun was "recovered" by authorities is three years or less. (Their use of the term "recovered" is interesting, but comments about that here would divert us from the intended report.) Their report spends much space discussing statistics concerning the time-to-crime data for traced guns. The ATF is careful to note "Newly purchased firearms can be stolen, for example; such a crime gun, if traced soon after the theft, would show a short time-to-crime." (see page 8 of the CGTAR). So they know not all such guns are trafficked. However, the naive media seems not so aware. The ATF uses a time-to-crime of three years as a threshold for investigation as to whether the recovered gun was trafficked. The ATF's approach is probably the result of statute of limitations and other pragramatic investigation limitations. However, as a measure of trafficking, significantly biases in the statistics will result from ATF practice. To illustrate, if the ATF were to limit tracing to guns recovered since 1994 instead of 1990, virtually 100% of the guns traced in the 1998 CGTAR would show a time-to-crime of under three years. Such a limit would probably be too obviously biasing even for the ATF, but this example illustrates that the amount of biasing should be examined. That is what I have done. My assumptions are: guns are sold at a constant rate since 1/1/1990, stolen at random, trace requests are made uniformly throughout the reporting period (1 Aug. 1997 to 31 July 1998), and only stolen guns are used in crime. Now these assumptions give a simple model that can be used to calculate what proportion of guns recovered would show a time-to-crime of less than three years. I do not suggest this model is correct (obviously there is some trafficking happening). My intent is to illustrate the biases in any data analysis with this measure. I constructed an Excel spreadsheet (which I'll be happy to supply on request) to calculate the approximate percentages of guns which would show a time-to-crime of under three years. I made the calculations for the 1998 report and the 1997 report period. To illustrate how the biases would change in time, I also made the calculations for 1993, 1994, 1995, 1996 and 1999 as well. They are presented in the table below (all assuming the 1990 cut-off for tracing and the corresponding periods for the trace year). 1993 98.0% 1994 73.9 1995 59.3 1996 49.4 1997 42.3 1998 37.2 1999 33.1 The actual CGTAR for the 1997 year only covered 10 months and not the full year of my model. I did not account for this change. The ATF CGTARs do not report an overall percentage of guns showing the short time-to-crime. They report time-to-crime statistics in four categories for each jurisdiction. The categories are age related (juveniles, youth, adult, and unknown) and only the top-ten crime guns are reported. So, direct comparisons are not possible. However, we should attempt to obtain comparable information from them, using "freedom of information" law if needed. The ATF's 1998 CGTAR reports also that new funds will permit tracing back to 1985. By a similar model, the 1999 percentage will change from 33.1 to 21.3. This result will permit the naive to believe great progress is being made against gun trafficking even if no trafficking were happening. The next step needed is to obtain more of the data from the ATF than is presented in their report. Perhaps, a friendly congressman could be approached to obtain the needed data.