Yes there have been other statisticians that have debunked Wyner's results. I could PM you some if you want.
I'll engage with you because I think you'll be fair and can look at things somewhat objectively. Even though I know Wyner is a biased actor with an agenda, I'll pretend he's not.
The fundamental premise of Wyner's analysis is based on two things:
1. On the days with many women casualties, that should coincide with the days with a lot of child casualties. Because he's making the assumption most of the time children are with women. But his data set doesn't supposedly coincide with that according to him.
2. The amount of increasing deaths reported is "too linear." The line is too straight on the graph basically. He says “an extremely regular increase in casualties over the period” and from which he concludes that “this regularity is almost surely not real.” There would be much more variations in real life, especially a war zone. This is the MAIN crux of his argument.
So he concludes: THEY MUST BE LYING.
He gave this graph as the example of how this isn't a natural increase in deaths.
The fundamental problem with this graph is that he MANIPULATED the data to take into account cumulative deaths
to make it look linear.
When you graph the INDIVIDUAL deaths day by day in the exact same time period, the graph looks like this.
NOT LINEAR AT ALL.
1. He manipulated the graph to make it look like the numbers were linear when they weren't at all.
2. The data Wyner used was from October 26, 2023 until November 10, 2023. That’s 16 days.
This conflict has been going on for almost 6 months.
You have to ask yourself: Why would a PhD economist even attempt to perform such a statistical analysis based on only 16 days?
3. These numbers are
documented deaths. I.e., the body was brought to a hospital/medical facility, the cause of death was established, and t
he person was identified (which includes not only the name but also the ID number).
A record is considered incomplete if identity number, full name, date of birth, or date of death is missing. Only after all this info is gathered are these confirmed numbers reported.
Well almost all healthcare data is aggregated in set time periods— once a week, every 3 days, or month etc. it’s too difficult for healthcare facilities to report out daily given the nature of their work, staffing constraints, recording time, time it takes to transfer the data, etc. Health min likely receiving data once every x amount of days.
Basically, you can only count so fast. And in the middle of a war, it's hard to hire more bean counters sometimes. So even if there is some linear increase in deaths, that just indicated when the deaths were RECORDED and not when they actually died.
4. Historically — in conflicts in 2008, 2014, and 2021 — the health ministry’s fatality numbers closely matched death tolls resulting from independent research by United Nations humanitarian agencies.
5. There's no rule that women deaths have to coincide with child deaths 1:1. Children can range from 0-17 years of age. Why does a teenager have to be with a woman? That argument doesn't even make much sense.
So basically Wyner cherrypicked a tiny sample size and manipulated it to look linear when it wasn't. He's full of shit.
partner.upday-content.com
