Background checks are one of the most common proposals to address firearm-involved homicides. Universal and/or expanded background checks are always high on the list of anti-gun groups’ proposals. But what does the research say?
The Rand Corporation’s review of firearms-related research found that dealer background checks may decrease firearms homicides but ruled the effect of private-seller background checks uncertain (based on inconclusive evidence). We discussed the Rand review earlier this month, but wanted to highlight again these findings and revisit an example of the steps anti-gun researchers will take to produce their desired findings.
Rand determined that waiting periods have an uncertain effect on violent crime and intimate partner homicide. Last October, we reviewed a study conducted by a team from Harvard Business School that claimed handgun waiting periods reduce gun deaths. We noted several crucial concerns with their variable selections (or rather, exclusions).
What we did not include in our initial article was a finding the researchers left out of their discussion. According to their models, background checks are associated with an increase in homicides. These coefficients were not statistically significant but do suggest their model is misspecified. The model produced the desired finding – that waiting periods reduce gun deaths – but did so at the cost of background checks, which they indicate increase all homicides and all suicides, as well as both firearms-related and non-firearms related homicides and suicides.
That doesn’t seem right. It also doesn’t mesh with the conclusions of the Rand review. Of course, the researchers don’t discuss their strange coefficient on background checks and media coverage of their paper focused almost exclusively on the headline. This was not the first-time anti-gun researchers have hand-selected findings to present and discuss, and we doubt it will be the last.
Remember the Kellerman study we often mention? It was funded by the CDC, was rife with methodological flaws, reached questionable conclusions, and yet is still regularly cited and referenced by pundits, politicians, and anti-gun organizations. Take a look at the table showing the final model output from the Kellerman study:
No one (except us and like-minded allies) ever mentions that the real “findings” of the Kellerman study are that renting a home or living alone are both higher risk factors than keeping a firearm in the home. Doctors want to opine about gun ownership, use, and storage but we’ve never heard of one referring anyone to a roommate finder or first-time homebuyer programs. Keeping a firearm in the home was actually the 2nd lowest risk factor in the model, but Kellerman and others only want to talk about firearms.
Anti-gun researchers, their benefactors, and adoring media never let questionable model specification or odd findings undermine their predetermined outcome. There is an old saying in statistics: “all models are wrong, but some are useful.” Anti-gun researchers seem to take this as a green light to do whatever it takes to produce a “useful” model.