Thread by @nickburbank710

Gross statistical errors committed by @Wharton Prof. Abraham Wyner in his @tabletmag article led a London School of Economics professor to call it “One of the worst abuses of statistics I've ever seen”.

Our analysis below: 🧵

His flimsy argument is based on a statistical “analysis” where Wyner cherrypicks a limited dataset to prove his point. He examines only 15 days of casualty data (between the red lines in the graph below) from more than 165 days of conflict.

Wyner's observations are skewed to the 15-day subset of data (red dots). Placing those days into the context of the full dataset (blue dots) quickly reveals substantial issues. For example, his conclusion that deaths of women and children are uncorrelated is obviously incorrect.

Wyner's conclusion that in order to maintain a quote more women casualties were reported on days few men were killed can be similarly disproven. Placing his 15 data point subset (red dots) in the context of the complete dataset (blue dots) shows the opposite.

Cherrypicking data is not the only substantial stat. error Wyner makes. He minimizes daily casualties (100s) by comparing them to cumulative values (1000s), and fails to account for the fact that 50% of Gaza's population are children in his assessment

https://x.com/Ted55643099/status/1770094598813053007?s=20

Overlooking Gaza's demographics, Wyner mistakenly argues indiscriminate bombing couldn't lead to 70% of casualties being women & children. Simple math contradicts this: with 50% children & an equal split among adults, women and children would make up about 75% of the populace.

Most disturbing of all is Wyner’s inability to understand what these numbers ARE - real people, killed by the hundreds as they sheltered in hospitals or bakeries or schools, deaths recorded by an agency overwhelmed by Nov 21 by the scale of the disaster apnews.com/article/palest…

There’s more - We couldn't replicate his R2 values. Wyner’s associate, @aizenberg55, fosters “critical debate & accountability of #HumanRights NGOs” @hrw & @amnesty. Their data source (uncited) appears to be flash assessments by the UN, which require independent verification 👇

@Aizenberg55 @hrw @amnesty For a complete assessment of the death toll since Oct7, and to replicate our analysis, see .

To ask @Wharton how they plan to address Wyner’s understanding and use of statistics as displayed in this article, email [e-mail]

@Aizenberg55 @hrw @amnesty @Wharton For more discussions of Wyner’s statistical malfeasance see these threads 👇

@ballouxfrancois -

@joftius -

@lpachter -

https://x.com/BallouxFrancois/status/1770084876055691700?s=20

https://x.com/joftius/status/1766199967364890949?s=20

https://x.com/lpachter/status/1766184928201654480?s=20