In a paper titled, “Changes in firearm mortality following the implementation of state laws regulating firearm access and use,” RAND Corporation staffers have published a paper examining child access prevention (CAP), right-to-carry (RTC), and stand your ground (SYG) laws. The paper shows that the people at RAND are continuing doing very biased and poorly done research on gun control. A summary of the article was published in Science.
We have never previously seen a paper that uses 80% confidence intervals. As soon as you look at their table 1, it is clear that they did that because they wouldn’t have gotten any statistically significant results if they used 95% or even 90% confidence intervals. Indeed, only one of their findings might be statistically significant at the 85% level (the child access laws for total firearm deaths would be at about 1.00 and one would need more information to know whether it was just above or below 1.00 to see if that was statistically significant at the 85% level).
The bottom line is that NONE of their results are statistically significant.
Even their first paragraph misstates what the views of gun control researchers are. They cite the New York Times survey, but they ignore our work showing that it is horribly flawed.
The differences between the New York Times and our own respondents appear to arise for several reasons. The Times’ consideration of only policies that would increase government regulation of gun ownership and the apparent dominance of public health researchers on its panel both worked to produce answers that were more sympathetic to gun control poli- cies. However, this does not explain all of the differences between our survey results and theirs, as the Times’ small selective sample was even slightly more supportive of gun control than was the average public health researcher in our survey. That is true even when we limit ourselves to just the Times’ surveyed policies.
ARTHUR Z. BERG, JOHN R. LOTT JR., AND GARY A. MAUSER, “Expert Views on Gun Laws,” Regulation, WINTER 2019–2020.
Their regressions aren’t serious, and while they cite one of my studies on this, they ignore the arguments made in it. For example, Child Access laws should have a much bigger impact on suicides for children than for adults. Many factors impact suicide rates, so what we did was to account for the adult suicide rate in explaining the juvenile suicide rate and then see if the child access laws caused the juvenile suicide rate to change relative to the adult suicide rate.
Concerning RTC laws, they assume that the impact of so-called “shall-issue” or “right-to-carry” laws causes a large, immediate increase in the number of permits (for a discussion of the problems see here). But that is not the case, for states differ widely as to how easily they issue permits. This problem is particularly problematic for studies that have looked at the period after 2000. With their data going on to 2016, this is a real problem. Once you get past 2000, they are comparing the changes in right-to-carry states to only a few other states and eventually no other places that ban entirely concealed carry.
There are also errors in their definition of what state laws are. For example, they don’t acknowledge that Rhode Island became a right-to-carry state due to a 2015 state Supreme Court decision.
You can get a pretty good idea of where these people at RAND are coming from by seeing that 97% of the money that they gave out went to gun control advocates, and the rest went to those that were either unidentified or neutral.
“
That study used an 80% credibility interval, not an 80% confidence interval.
Does this change the results?
This study utilized credibility intervals from Bayesian statistics, not confidence intervals. Does a probability of 0.80 still undermine the study’s conclusions?
(I apologize if this is a repost– the website did not inform me if my original reply was submitted this morning).