The line used to be that a little bit of alcohol was good for you.

"But what these studies didn’t take into account is the history of those individuals and their drinking behavior in the past. “That was quite a big miss into how these studies were designed and how data was collected,” he says.

"The teetotalers included people who were sober because they had to be — people who stopped drinking because, as Morgan puts it, “they drank so much it was killing them.” When you remove those people from the data, you get a straight linear correlation: zero alcohol consumption was associated with the lowest mortality."

https://www.cbc.ca/documentaries/the-nature-of-things/the-new-science-of-alcohol-why-recent-guidance-says-more-than-2-drinks-per-week-is-risky-9.7021701

#alcohol #research #stats #history #BadStats

@MoSchoenfeld

Not only that, but most of the alleged proof of its value is really measuring how the kids who do their homework are also the ones best supported in general. Homework is more likely to actually be done by kids with supportive adults at home. It's a signal, not a cause. If anything, it amplifies the gap between kids who do & don't have that support.

@alfiekohn has a book about it :-)

https://www.alfiekohn.org/homework-myth/

#school #homework #BadStats

Back in March, the Royal College of Physicians did a survey about "Physician Associates", non-doctors who work in medical settings.

Then they had an EGM where the survey results were presented in a biased way.

Now there's an independent report about that EGM, which I found quite interesting. (Process-geekery! Governance! Bad stats!)

PDF here:
https://assets.kingsfund.org.uk/f/256914/x/d41d2041a8/rcp_independent_learning_review_09_2024.pdf

#NHS #RCP #governance #PhysicianAssociates #BadStats

9 out 10 made at least one #ransomware payment <-- This is obviously not the case #badstats

https://www.accountingtoday.com/news/9-out-10-made-at-least-one-ransomware-payment-last-year

Another #inktober come and gone. I am happy to report that this year the 11 drawings I posted to Mastodon were boosted an average of 27 times!

day 1 - none
day 2 - none
day 4 - 300 boosts
day 6 - none
dat 7 - none
day 8 - none
day 11 - none
day 13 - none
day 15 - none
day 22 - 1 boost
day 24 - none

#badStats