Skip to content
 

Kaiser Fung suggests “20 paper ideas pre-approved for prestigious journals”

I got to thinking about this after reading a post from Kaiser Fung “offering up 20 paper ideas pre-approved for prestigious journals.”

What happened is that JAMA published a silly paper claiming a 12 percent increase in fatal car crashes on April 20 (“420 day,” the unofficial marijuana holiday). Following Sam Harper and Adam Palayew, Kaiser reports:

The following 21 days have more extreme ratios than 4/20:

Jul 4 Dec 23 Dec 21 Nov 21 Sep 1 Dec 20 Sep 2 Jul 3 Dec 31 Oct 31 Nov 23 Dec 18 Dec 6 Jul 14 Sep 4 Dec 22 Mar 17 May 25 Apr 1 Mar 7 Dec 19

Will JAMA editors accept one research paper for each of these days? The work is already done – the rest is story time.

That would be just awesome if JAMA would publish those 21 papers. But it’s not gonna happen. Too bad.

And that got me thinking of all the other papers that really, really should get published. There was that paper reporting that Cornell students have ESP. There are 7 other Ivy League colleges, so I think JPSP could publish at least 7 other papers in this series. And then there are those near-Ivys: Stanford, Northwestern, WashU, etc. . . really lots and lots of options here.

Etc.

14 Comments

  1. Adede says:

    Jul 4 and Oct 31 make sense.

    • Andrew says:

      Adede:

      Sure, lots of the dates make sense. End of Aug and early Sept is Labor Day weekend, 31 Dec is New Year’s, etc.

      JAMA should just publish all 21 of these papers in a special issue and be done with it!

    • jim says:

      I think Nov 21 makes the most since, since according to Wikipedia over 100 important people have died on that date since 615AD – including David Cassidy!

      My ESP tells me someone ELSE will die on that date SOON, perhaps in the coming decade!

  2. Dan F. says:

    It seems to me that, whether you realize it or not, your blog is as much about ethics as it is about statistics.

      • Dan F. says:

        Maybe my comment was really a bit of self-realization. Your blog is constantly making the point that statistical malpractice or ignorance underlies a lot of unethical behavior. More and more I think the contrary, positive, affirmation needs more attention – by thinking statistically and carefully about statistics we can behave more ethically, and recognize potential ethical problems in behaviors that have been normalized or are considered acceptable. I suppose you and some of your readers are already making such observations/connections, at least implicitly, a lot. My comment cam from realizing that somehow this is a central point of your blog that somehow hadn’t been explicit in my mind.

        Summary: I value the discussion of ethics on this blog as much as the discussion of statistics.

        Another way of putting it: maybe one could consider “applied ethics” as part of statistics.

        • Martha (Smith) says:

          I think a big part of ethics in general is being aware of the assumptions and reasoning behind whatever you are doing (or advocating), and checking the validity of those assumptions (in the case at hand) and the quality of that reasoning.

          • Andrew says:

            Martha:

            I agree 100% . . . but, then again, what you’re expressing is a very mathematical view of ethics, perhaps not surprising given you and I, as mathematicians (ok, I’m a statistician, but that’s close) would think. It’s all about deductive reasoning: drawing conclusions from assumptions and then going back and evaluating those assumptions based on their implications. In concept, this is not so different from proving a theorem and then going back and figuring out what are the minimal conditions under which that theorem could be true.

            Others with less mathematical taste might put the foundations of ethics in other places, for example, in psychology (for most of us, doing something unethical feels wrong, it pricks our conscience) or in social relations (you can judge ethics based on other’s reactions) or in economics, law, power relations, etc.

            • Martha (Smith) says:

              Andrew:

              I think your response to my comment is a bit of an overreach. You said,

              “It’s all about deductive reasoning: drawing conclusions from assumptions and then going back and evaluating those assumptions based on their implications. “

              I didn’t say “It’s all about …” — that’s coming coming from you, something that you presumably read into (or assumed) about what I said.

              Also, when you say “Others with less mathematical taste might put the foundations of ethics in other places…”, you seem to be assuming that I put the foundations of ethics in deductive reasoning, whereas I said nothing about “the foundations of ethics”

              What I said was, “a big part of ethics in general is being aware of the assumptions and reasoning behind whatever you are doing (or advocating), and checking the validity of those assumptions (in the case at hand) and the quality of that reasoning.” As I see it, “a big part of” is quite different from “It’s all about” or taking about “foundations”.

              Indeed, something “feeling wrong” or “pricking my conscience” are to me also important parts of ethics — in particular, I do feel I’ve done something wrong (and it does prick my conscience) if I base a conclusion on faulty assumptions or poor reasoning.

              • Andrew says:

                Martha:

                I agree. Amend my comment above to say that for mathematicians such as you and me, deductive reasoning is a big part of ethics. For others, deductive reasoning may not be such an important part of how they see ethical issues.

    • Martha (Smith) says:

      Statistical inference is a means of trying to make inferences in situations where uncertainty is involved — and situations involving uncertainty are situations where ethics are inherently involved.

  3. jim says:

    I’m curious Andrew if you noticed or thought about the 3 billion missing birds.

    I don’t have access to the paper, but I’m told the figs have error bars. I don’t especially doubt the numbers but, given the methods in use and land use changes, I presume the error would be substantial ± 20%?

    I think in a research paper the error should be attached to the “news number” when that number is first mentioned – in the title or the abstract or whatever – so that it’s always visible with the number.

    Also, the conservation community seems to have grabbed the numbers to run with them, with a spate of editorials chasing the news reports in the following days. The general perception of scientists seems to be that the numbers are solid and the methodology is sound – which I presume is true at the wide scope of this work. But given land-use changes and many other factors, it would be most useful if a deeper analysis were undertaken on subsets of the data. I doubt that will happen, but it should because it would be bad to build policy around this data and later find out there was a methodological problem lurking in our friends, The Assumptions.

Leave a Reply to Martha (Smith)