Page 10 in the appendix is a handy place to put a statement like:

No significant relationships existed between gun shows and firearm injuries along known trafficking routes or when California’s 10-day waiting period was excluded.

That is where one of many key clues about the lack of viability of a recent piece or research was filed. That tax money from the federal (NIH) and California (U.C. Berkeley) governments was used to generate this paper and the minor media frenzy that followed should give voters something to consider at election time.

Key takeaways

• Raw data shows lower rates of California death and injury after gun shows.
• Even after questionable adjustments to the data, rates of change in gun misuse were small, possibly within random variations.
• Controls for other influences (confounding variables) were odd and incomplete.
• Unexplained, undocumented adjustments leave the quality of the research in grave doubt.
• No verification that the misused guns were acquired at gun shows at all, much less Nevada gun shows.
• Ignored that California residents buying at Nevada gun shows must still go through the legal importation process.

A good idea gone wrong

Many a good research project have been fouled by attempting to prove something. Such is the case with “In-State and Interstate Associations Between Gun Shows and Firearm Deaths and Injuries“. The question  was interesting, namely do the strict gun laws in California, compared to the less strict laws in Nevada have any particular effect on guns deaths and injuries. For number nuts like the Gun Facts researchers, this would be of great interest.

Painfully, like any B-grade horror film, everything starts to go wrong with this research.

Adjusting data, and leaving no trace

As confirmed by the quote above, copied directly from the study, rates of gun misuse in California did not change for the worse after a Nevada gun show. Oddly, rates of gun play fell.

QUICK NOTE: For Nevada, the raw number of firearm deaths and injuries before shows was small. Over the eight-year period of the study, there were 44 within driving distance of the shows for two weeks before the shows. This is in the statistical noise realm, and thus the entire study, at least on the Nevada side, is suspect.

This is where the rather strange relationship between researchers and the media can turn one’s stomach. Using the raw data (before statistical “adjustments”), we see that all categories of California firearm misuse, in the regions under investigation, did not change except for firearm accidents, which went down.

That would make a very uninteresting headline. However, after applying their “adjustments”, the researchers claimed a statistically significant increase in California gun play, which the L.A. Times and other former dens of journalism called a “sharp rise”. The lack of inquisitiveness by L.A. Times reporters is regrettably common.

Maladjusted?

Ignoring maladjusted reporters, we have to then ponder the adjustments to the data that researchers employed. It is not always wrong to model data, to correct for known or anticipated issues with data quality. But that isn’t the case with this work.

Troublesome are some modeling constants that are undocumented. The authors adjusted for “seasonality” without explaining it, explaining why this was an important issue, or presenting the math. They also had an “offset for the number of at-risk persons” without documenting their definition of an at-risk person or the constant employed.

Clues buried in the text

We can dismiss any research that does not fully document their processes, and thus this one is rejected as both inaccurate and incomplete. But parts of the research are entertaining for what they reveal.

For Nevada gun shows, changes in firearm injuries remained statistically significant for shorter (1-week) and longer (3-week) periods but were not statistically significant for smaller geographic ranges (60-minute drive), which yielded very few cases, or larger geographic ranges (120- and 180-minute drives for California and Nevada guns shows, respectively), which covered large portions of California.

Unpacking that bit of academic jargon, rates of gun misuse did not change much if you looked at short or long periods around gun shows, or if you looked at places nearer to or further away. This is a complex way of saying time and distance were not variables. Which then brings us to the paper’s main claim, that gun misuse rises (after “adjustments”) immediately after gun shows and within a moderate driving distance of the event.

The authors make some incorrect statements, such as:

Nevada does not require background checks or documentation for private transfers and places no regulations on gun shows.

The problem with this statement is that interstate private transfers must go through a Federal Firearm Licensee (FFL). Gun shows in Reno and Las Vegas, Nevada have California FFLs on-site to facilitate such. So, the regulations are indeed in place and are being observed by people in the gun show industry.

The authors then seem to contradict themselves, promoting the alleged viability of California’s 10-day waiting period. First:

California’s regulations and 10-day waiting period motivate buyers to cross into Nevada when seeking a faster, less-regulated source of firearms.

Then:

No significant relationships existed between gun shows and firearm injuries along known trafficking routes or when California’s 10-day waiting period was excluded.

Indecision in research is not a fault, but it is amusing.

… few firearm injuries occurred in regions exposed to Nevada gun shows

Yet their final conclusion was that exposed regions in California saw increases in firearm injuries.

… we did not examine associations with firearm injuries in Nevada populations

A curious omission given that they already gathered the data.

What’s to like?

In this bit of research, there is little to like. One gets the feeling that the authors were on a mission to tag gun shows as a source of mayhem, or to promote California gun laws as a “cure”. Yet in the raw data, we see no validation of either point.

The two big problems though are not in the research. The big problems are activist in the media who avoid due diligence in reviewing research, and government agencies investing in research that is substandard. In the long run, only news consumers and voters can fix these inequities.