The OP. Boring discussion about NRA crowd sizes over time; at least it’s topical.
Ronald Reagan spoke at the NRA convention in 1983. (Guns and knives were prohibited in the assembly hall where he spoke, and attendees had to pass through a metal detector. None of this is surprising.)
Create a trillion fair coins. Then mint one coin with both sides the same and toss it somewhere in the pile. Grab a random coin, flip it eight times. Suppose you somehow get a streak of eight in a row of the same result. Your p-value will be less than 0.01, so you can “reject” the null hypothesis that the coin is fair at that level.
But what’s the actual chance that the coin you grabbed is fair and you just had a strange amount of luck flipping, rather than grabbing the one single false coin in that massive pile?
[spoiler]Chances are better than 99.9999% that you got a strange streak from a fair coin. I could write more 9’s, too, but I don’t want to bother with being more precise here.
The p-value of less than 0.01 does not give you the probability that the coin is fair. That is not what it means. The coin is almost certainly fair, despite the p-value.
[/spoiler]
There is an army of people whose future career chances are improved if you they can find strange relationships in data. Any given p-value means next to nothing in that context. You can run regressions on NRA conventions. You can run regressions on NRA sales drives. You can run regressions on mass shootings compared to Wayne LaPierre’s bowel movements. If you run enough regressions, you will find a strong correlation somewhere. This result, like so many before it, is almost certainly a case of survivorship bias. The correlation exists, so it was reported on.
But come on.
Obviously we can’t expect a poster like Magiver to understand what the word “correlation” actually means. But at the same time, there is no reason to give this result more than about five seconds of attention. The veritable army of job-hungry data miners will always, always be mining their data. They will never stop. (Or for a more charitable take on the same problem, google Andrew Gelman’s writings on the “Garden of Forking Paths”. There is no requirement of deliberate statistical dishonesty for this to still be a problem.)
Our media system is designed to pimp the exceptional and surprising, not to be rigorous. But the few sensible people in the world can learn to stop with the post hoc rationalizations that human minds are so adept at making. Rather than storyboarding a causal relationship after-the-fact, we can think about more relevant things.
Crushed by Bayes again. Readers wanting an article length take can go here. See also publication bias, for those wanting a technical term.
I’m not as skeptical as Hellestal is though, for the following reasons:
The article reached a p-value of .005, which is the value advocated by the statistical reformers in the Vox.com article. Yes, yes, that doesn’t repeal Bayes law.
But this isn’t a case of journalists pulling some obscure paper out of a hat. The article was from the pre-eminent New England Journal of Medicine. Now admittedly I’ve seen a number of low sample sized articles in that journal. Still.
3a. Publications bias is a fine counter-hypothesis. But this frankly doesn’t look like the sort of relationship that someone would stumble upon at random. Presumably NRA meeting occur during at least slightly different weekends of the year.
3b. All the same, if this is an instance of publication bias, then it should be easily addressed by looking at, say, 7 pre-sample years. And since we already have prior information to go on, a p threshold of .05 should be sufficient.
Full disclosure: I haven’t fooled around with a back of the envelope calculation evaluating the paper’s plausibility. I confess though that I wouldn’t be surprised if such an effort pulverized the argument, peer review notwithstanding.
Magiver: “Statistical carrot logic”. That’s not a bad summary actually.
It’s late, so I didn’t read comments past the first page, so if this has been asked I apologize.
If, as the OP seems to be suggesting, NRA members are so dangerous that in a country with 300 million firearms a statistical difference could be made by a mere 80,000 members attending a meeting, then why aren’t those annual meetings when the venue allows open or concealed carry not a bloodbath?
If the OP really wants to know if our Nation’s firearm violence is mainly the fault of us zealot, gun toting NRA members would it be more to the point to study how many illegal deaths or injuries caused by firearms were done so by NRA members? While there will certainly be some, I’ll go out on a limb and bet the vast majority were caused by a person with zero affiliation to the NRA.
Yet another good question to be addressed in a follow-up paper, since it’s outside the scope of this one. Or more likely multiple follow-ups, because that’s a complicated one.
For the record, NRA members are significantly more likely than non-NRA gun owners to carry and to keep their firearms loaded and readily accessible. I offer this strictly for informative purposes and make no judgments about what that implies, if anything.
If you are aware of a database that tracks gun injuries in conjunction with NRA affiliation, I’m sure the OP is all ears. Unfortunately, that sort of crime statistic is not tracked by the BJS (or elsewhere, as far as I can tell). For obvious reasons, medical personnel don’t ask patients whether they are NRA members.
The maximum one-day attendance at the Florida State Fair is over 80,000 (last year, the highest-attended day had 71,000 guests). I’m not sure how you concluded that the fairgrounds can’t hold 81,000 people.
For what it’s worth, I looked at the Atlanta monthly crime statistics for April (when the NRA meeting was held there in 2017) and the rest of the first half of 2017 and didn’t see any noticeable increase in gun crimes. I did see a noticeable increase in overall crime in April 2017 over April 2016, which was interesting because crimes for the period of January to April were down in 2017 versus 2016.
I’ll concede that you are 109% correct on this. My carry weapon always had a round in the chamber and if I’m wearing pants is on my person. An empty weapon is not a defensive tool its a paperweight. Having a round chambered could be the difference between successfully defending yourself and becoming a victim. One of the very first things that tends to happen in a deadly force encounter is fine motor skills become very difficult if not impossible to perform. Racking a slide is most certainly a fine motor skill. Sadly, I’ve seen video of a man desperately trying to rack a round into his defensive weapon while under attack. Let’s just say the outcome wasn’t the best.
Anyway, I’ll agree, us NRA members tend to keep our firearms in a ready to go status. I’ll also point out they the majority of us also know they as long as we keep our finger off the trigger, that firearm cannot and will not discharge on it’s own.
I can’t find the answer to this online: the study in NEJM appears in the “Correspondence” section. I know that a “Letter to Nature” is peer-reviewed, but are submissions to the Correspondence section of NEJM likewise?
You are assuming that “dangerous” means that they are dangerous to everyone. I can easily imagine a gun-carrying man who is dangerous to #1 himself, #2 his spouse and/or children, #3 anyone whom he perceives as a deadly threat. If you put that man in a convention center with 80,000 other people who are dressed like him and talk like him, I can easily imagine that he would NOT perceive those 80,000 people as deadly threats, but rather would feel at ease. And if his spouse stayed at home, all the better. His spouse is the one person in the whole world who has the highest risk of being shot by him. If he’s away at a convention, then of course her risk of dying goes down while he’s gone.
I’m not saying this straw man represents the average NRA member. I’m saying my straw man is easy to imagine. You, on the other hand, seem to be imagining a very different straw man. The one you’re imagining hates everyone equally and is likely to shoot someone for absolutely no provocation at all. I agree with you that your straw man is very hard to imagine except as a cartoon character. It’s much more likely that a dangerous person would be more dangerous toward some people and less dangerous toward others. A depressing large fraction of gun fatalities in the US happen within the immediate family of the gun owner.
You might be right. Of all the gun toting zealots (your words) that I have met, most of them told me they are not NRA members.
You are correct to notice this. What Tired and Cranky linked to in post 7 is a Letter to the Editor, which is not a study and is subject to editorial approval, not peer review. This can be contrasted with the NEJM’s other article types. An Original Research article is what is generally considered a study and is very different from a letter. Nature also has a Correspondence section which is similarly not subject to peer review.
I screwed around with Wikipedia to gather my cites.
That said, the Florida State Fair is outdoors right? It’s my impression that the Fairground’s indoor capacity is much lower. And I assume the NRA would book locations where all attendees could mingle indoors. Gun shows are typically indoors, right? (Full disclosure: I’ve never attended one, though I’ve read about them.)
There are several ways it can be illegal for you to obtain a gun. One way is if you’re a person who is not allowed to own a gun (such as a convicted felon). In that situation, yes it would be illegal for you to inherit a gun. If your father bequeathed you a gun in his will, you’d be obligated to tell the executor of your father’s will, “No, thanks.” The executor could sell it and give you the money, instead. Assuming you are legally allowed to own a gun, then it depends on what state you live in. If you live in a different state from your deceased father, then it’s only legal if the transfer of ownership is done through a licensed federal firearms dealer. If you skip the dealer and just take the gun without telling anybody, that’s illegal. If you live in the same state as your deceased father, then it depends on the state. Some states (such as CA) require you to use a licensed federal firearms dealer anyway. Other states don’t. IANAL.
This question makes me think about what happens when you inherit a car. The car was titled in the deceased’s name. You need to go to the DMV and show them copies of the death certificate and maybe the will too. Then they have you fill out paperwork, show your proof of insurance, and then the issue you a new title for the car with your name on it. If you just left the funeral in your dead relative’s car and never bothered to go to the DMV, that would be illegal too.
Congratulations to the researchers: it’s an intriguing result. I say we’re well within the zone of suggestive evidence, given the high p-values and provided storyline. Nice letter to the editor: put in on your CV and high-five.
But it hasn’t had peer review, just editorial review.
3a. Upthread, critics questioned the plausibility of the results. I’d like to see a back of the envelope calculation showing how you can get the observed changes in firearm injuries (amounting to less than 100 people, if I read the paper correctly) from 81,000 NRA attendees. How many firearm shops need to close? How many firearm outings need to be postponed or rescheduled? What sort of multipliers are we talking about? 1.3? 2? 500? Are these suppositions in any way backed by casual observation?
3b. Remember SDMB rules. The authors are making the claim. They need to show the plausibility.
I don’t think the authors are making extraordinary claims. But they are unusual enough to ask them for reasonable robustness checks. Add a couple of weeks forward and/or backwards. See whether you can get a few more years of data. Etc.
If eg #3 fails and #4 passes is a convincing way, then we have a scientific puzzle. I say that would be publishable, though not yet sufficient for policy purposes. Be your own worst critic: good science!