Progress in post-publication review: error is found, author will correct it without getting mad at the people who pointed out the error

Valentin Amrhein points to this quote from a recent paper in the journal, Physical and Engineering Sciences in Medicine:

Some might argue that there is confusion over the interpretation and usage of p-values. More likely, its’ value is misplaced and subsequently misused. The p-value is the probability of an effect or association (hereafter collectively referred to as ‘effect’), and is most often computed through testing a (null) hypothesis like “the mean values from two samples have been obtained by random sampling from the same normal populations”. The p-value is the probability of this hypothesis. Full-stop.

Hey, they mispunctuated “its”! But that’s the least of their problems. The p-value mistake is a classic example of people not knowing what they don’t know, expressing 100% confidence in something that’s completely garbled. Or, as they say in Nudgeland, they’ve discovered a new continent.

But everybody makes mistakes. And it’s the nature of ignorance that it can be hard to notice your ignorance. It can happen to any of us.

The real test is what happens next: when people point out your error. The twitter people were kind enough to point out the mistake in the above article, and Sander Greenland informs me that the author will be sending a correction to the journal. So that’s good. It’s post-publication review working just the way it should. No complaining about Stasi or terrorists or anything, just a correction. Good to see this.

P.S. Confusion about p-values appears to have a long history. Greenland gives some references here.

4 thoughts on “Progress in post-publication review: error is found, author will correct it without getting mad at the people who pointed out the error

  1. Andrew:

    1: Re: “Hey, they mispunctuated “its”! “
    Credit where credit is due: The original article didn’t mispunctuate “its”. Here’s the original: “More likely, its value is misplaced and subsequently misused.” (https://link.springer.com/content/pdf/10.1007/s13246-021-01068-1.pdf)

    2: Here’s the text of the published correction to the original p-value indiscretion:

    Dear Editor,

    I wish to address the definition of p-values, which I misstated in my piece [1]. I made an unfortunate editing error which risks perpetuation of an incorrect definition, and I wish to correct this on the record.
    I wrote:

    The p-value is the probability of an effect or association (hereafter collectively referred to as ‘effect’) and is most often computed through testing a (null) hypothesis like “the mean values from two samples have been obtained by random sampling from the same normal populations”. The p-value is the probability of this hypothesis.
    I made an editing error, which every writer fears will happen sooner or later. Mine was to miss inclusion of the word “not.” The p-value is not the probability of this hypothesis. My goal was to try to define or identify a common misunderstanding of the value. Instead, the result was that the final sentence of my piece was false. Further, it effected the direct opposite of my general goal as a scientist: It promulgated an incorrect definition, and for a subject where heated opinions can be held. A correct interpretation, as my debater suggests, is probability of getting a result equal to or more extreme than the observed result.
    I hope that this erroneous statement does not distract from my point in this opinion piece: there is value in considering alternative statistical tools when practicing clinical medical physics. I apologize for any confusion or misunderstandings this might have caused.(https://link.springer.com/content/pdf/10.1007/s13246-021-01074-3.pdf)

    The Earth is once again spinning in its axis!!

    John

    • John:

      I went to the link and found that paragraph, and I noticed two funny things. First, he kept this phrase, “The p-value is the probability of an effect or association,” which is incorrect. Second, in that paragraph they have a different punctuation error (writing “Bacons” instead of the correct “Bacon’s”).

      • Andrew:

        I spotted this too and emailed a heads-up to Basran about his uncorrected error, with a cc to the moderator of the debate. What makes this especially cringe-worthy is that the uncorrected error was the more prominent of the 2!

        I’ll let you know if I hear back from either of them.

        John

Leave a Reply

Your email address will not be published. Required fields are marked *