How the Hemenway Surveys Distorted Estimates of Defensive Gun Use Frequency

Abstract

David Hemenway and his colleagues have claimed that two national surveys that they conducted indicated there were few defensive gun uses (DGUs) in the U.S., and that the number of gun crimes committed with firearms greatly is far larger than the number of DGUs. This paper explains how these authors produced extreme underestimates of DGU frequency and gross overstatements of the number of gun crimes.

Underestimating DGUs was accomplished by (1) using an eccentric and biased wording of the DGU question, (2) using a trap question that misled Rs with a DGU into thinking they had already reported the DGU, (3) employing a long recall period that increased memory loss, and the (4) selecting a biased sample that systematically underrepresented people likely to have a DGU. Overstating the number of gun crimes was accomplished by (1) mischaracterizing incidents falling into a largely meaningless “hostile display” category as gun crimes, when the authors’ own evidence indicated that most of the “victims” of these displays did not regard them as a part of gun crime, and (2) ignoring the far more sophisticated National Crime Victimization Survey estimates of gun crime.