surveys described in Chapter 5. Nonresponse rates make it difficult to draw precise inferences about ownership rates and use as the data are uninformative about nonrespondents. With nonreponse rates of 25 percent or more, the existing surveys alone cannot reveal the rates of ownership or use. Prevalence rates can be identified only if one makes sufficiently strong assumptions about the behavior of nonrespondents. Generally, nonresponse is assumed to be random, thus implying that prevalence among nonrespondents is the same as prevalence among respondents. The committee is not aware of empirical evidence that supports the view that nonresponse is random. Indeed, studies of nonresponse in surveys of drug consumption provide limited empirical evidence to the contrary (see National Research Council, 2001). These studies find differences between respondents and nonrespondents in terms of both drug use and other observed covariates (Caspar, 1992; Gfroerer et al., 1997).

Concerns about response errors in self-reported surveys of firearms possession and use require much more systematic research before surveys can be judged to provide accurate data to address critical issues in the study of firearms and violence. The many substantial resources that have been devoted to addressing the measurement issues in the collection of other sensitive data will almost certainly be useful, yet the issues surrounding firearms may be unique. The committee thinks that new research will extend and strengthen what is currently known about response errors on sensitive topics generally. Without systematic research on these specific matters, scientists can only speculate.

Administrative and Convenience Samples

A number of administrative data sets have been used or suggested as a way to study the market for firearms possession and use. In this section, we describe the administrative data collected as part of the Bureau of Alcohol, Tobacco, and Firearms’ tracing system, the trace data, and a proposed addendum on firearms to the National Institute of Justice survey of arrestees, the Arrestee Drug Abuse Monitoring (ADAM) survey.

BATF Firearms Trace Data: One federal source of information on firearms related to violence is the firearms trace data compiled by the Bureau of Alcohol, Tobacco, and Firearms of the U.S. Department of Justice. Because trace data are quite distinct from the other federal data sources, and because they have been subject to more criticism than most of the other systems, we provide a more extensive description of the regulatory background related to firearm tracing, the nature of the tracing process, and the uses and limitations of the resulting data.

The Gun Control Act of 1968 established the legal framework for regulating firearms transactions and the associated record-keeping. The act



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement