Hmmm, you're right. I didn't find the correct US law.

A-ha! Found it! The bill which changed it was the Investigative Assistance for Violent Crimes Act of 2012, 28 USC 530C(b)(1)(M)(i), https://www.gpo.gov/fdsys/pkg/PLAW-112publ265/pdf/PLAW-112pu... and the law text is at https://www.law.cornell.edu/uscode/text/28/530C :

> For purposes of this subparagraph — (I) the term “mass killings” means 3 or more killings in a single incident; and (II) the term “place of public use” has the meaning given that term under section 2332f(e)(6) of title 18, United States Code.

It looks like Mother Jones is using the current federal definition.

Agreed, just because there is a definition doesn't mean it's a good definition. The idea that 18 people can be shot but only one killed should still make it be a mass shooting, but the FBI wouldn't count it. That's the argument given https://www.massshootingtracker.org/about , which uses a different definition to conclude there have been "1306 mass shootings since 1/1/2013"

Does feedback like this help you adjust your confidence levels? That is, might you think better of a Mother Jones analysis in the future, or give you less reason do the legwork?

While on the other hand, this op-ed piece appears to have undercounted the numbers, and for unspecified reasons uses "the traditional FBI definition" rather than the current FBI definition. (I, and perhaps you also, thought he was using that phrase to mean something like 'the current and long-held definition', not 'the older definition, which I prefer').

Does that make you less likely to trust the general argument?

"Agreed, just because there is a definition doesn't mean it's a good definition."

Whether it is a good definition or not is important, but you're not focusing on the salient fact: the definition was changed, recently, which makes it difficult to compare current data with older data.

"for unspecified reasons uses "the traditional FBI definition" rather than the current FBI definition."

Perhaps because he wants to actually compare the rates over time without talking nonsense.

At some level the cutoff for what constitutes a "mass shooting" is fuzzy. That doesn't make it okay to do a straight comparison of data that uses one definition with data that uses another definition.

It doesn't matter that the federal definition changed a couple of years ago.

The salient question is, given the data - no matter what definition is used - does the analysis make sense?

I gave a link to a page which goes into that more carefully, and concludes that a simple average, as the editorial author did, does not make sense.

This question about the definition is a side-issue. The fact that I cannot reproduce the numbers does make me wary, but even if it were correct, the conclusion drawn from that data is simply false.

One way to improve the confidence of the data, without more sophisticated statistical analysis, is to increase the baseline. He already has a much longer baseline for US data, and he elsewhere complained about the arbitrary time definition for an FBI report which he said artificially skews the results.

Why then did he choose the first 7 years of Obama, and not 20-30 years?

I can't believe you're seriously arguing that the definition doesn't matter.

Okay, let's define "mass shooting" to include people who are shot wholesale by government action (by the way: why shouldn't it?), and increase the time frame to 75 years.

Now compare the U.S. to Europe, and your conclusions about the relative rates will be...somewhat different.

The Netherlands lost about 2% of its population in WWII, the equivalent of about 7 million deaths in a population the size of the current United States.

You don't need to believe it because that's not what I said.

I said "just because there is a definition doesn't mean it's a good definition". That doesn't mean I think it's a good definition or a bad definition. I personally prefer the definition used by massshootingtracker.org, which concludes there have been 1300+ mass shootings since 2013. However, I also think someone will argue about any reasonable definition.

I also said "It doesn't matter that the federal definition changed a couple of years ago." That doesn't mean I think any definition will work, only that for the purposes of this analysis, we can choose either the old definition or the new one - or any reasonable definition.

I'm willing to accept the old definition and focus on the real "salient", to use your term, issue, which is that the results of this op. ed piece are worthless, because of poor use of statistics.

Why wouldn't a "reasonable definition" include people who were shot in large numbers by murderous governments?

Take your time.

Because it doesn't fit the goal of the op. ed piece, which is what we are talking about.

If anything, it says that mandated arming of large numbers of citizens (conscription), and mandatory use of such weapons, leads to a higher number of mass killings.

That's a different conversation than the mandated disarming of large numbers of citizens.

"Because it doesn't fit the goal of the op. ed piece"

Translation: it wouldn't fit the political agenda of the op. ed piece.

I'm glad that you agree with me that your definition is not relevant to the op.ed piece or to this entire thread.

That is quite clearly untrue, and you are no longer debating in an honest manner (if you ever were).

Buh-bye.

If only more kibitzers were so polite.