Statistics professor @AdiWyner wrote a rebuttal to @TheLancet's peer reviewed mistakes on Gaza. The Lancet ignored him. ~ Elder Of Ziyon

Tuesday, March 25, 2025

Statistics professor @AdiWyner wrote a rebuttal to @TheLancet's peer reviewed mistakes on Gaza. The Lancet ignored him.

In January, The Lancet published a peer-reviewed study that used a "capture-recapture" method to estimate that the actual number of people killed in Gaza was 60% higher than the Ministry of Health statistics - 64,000 compared to the 40,000 at the time of their analysis.

While I am no statistician, I showed that the input data they used was not at all appropriate for this estimation method. The capture-recapture method requires that sources of data are independent, but two of the sources they used were complementary, making all of their fancy math worthless. (The third source was also worthless.)

Since then, my conjecture was proven correct. If their numbers were correct, presumably the missing dead people were buried under the rubble, but the number of bodies recovered during the ceasefire was nowhere close to what one would expect if tens of thousands were missing.

I noted that this is one of three peer reviewed Lancet studies on Gaza casualties, all of which have major errors that never should have passed any review. I wrote, "This is not a coincidence. It is a pattern. The Lancet would never publish papers that disprove these, they would never correct these papers in light of new information that proves them wrong, and from what I can tell, they have not published any letters pointing out these severe flaws."

It looks like I am correct about that as well.

Professor Abraham Wyner, a professor of statistics at the Wharton School of University of Pennsylvania, submitted a rebuttal paper soon after the Lancet's publication. It agrees with my main point why the original study is flawed. He then recalculated the numbers based on more reasonable assumptions and showed that the paper's estimates were way off.

I wish it was easier to read; here is part of the abstract:

A January 9th Lancet (Jamaluddine, 2025) study on traumatic injury mortality in the Gaza Strip employs a three-list capture-recapture model, widely used for estimating partially observed populations using multiple random samples. But the methodological framework and assumptions underlying this study raise significant concerns about the reliability and accuracy of its conclusions. The authors estimate that the number of recorded decedents substantially undercounts the true population of traumatic deaths by approximately 35,000, an amount larger, by far, than the number of recorded deaths. In this response, we show that this surprisingly large estimate has two causes: 1) a methodological problem and 2) the inclusion of a relatively small but highly influential subset of bad data.

Here's the kicker: Wyner submitted this paper to The Lancet two months ago, and they have essentially ignored it.

I have long complained that The Lancet was only publishing papers that fit with its political position, the exact opposite of science. Now we have proof.