Paperonium

Mental health · Psychol Med · 2021

The summary’s 4.5 lies outside its own range, 1.4 to 3.8. The table’s range reaches 13.8.

The paper’s summarylies outside 1.4 to 3.8The paper’s table4.5 lies inside 1.4 to 13.84.5×should be inside1.4 to 13.8×1×2×5×10×20×same odds as people with low distress
The paper’ssummaryliesoutside 1.4 to 3.8The paper’s table4.5 liesinside 1.4 to 13.84.5×should be inside1.4 to 13.8×1×3×10×30×same odds as people withlow distress

Times the odds of losing or changing a job, serious versus low distress

The paper says

The paper’s summary: adults who reported serious mental distress in a survey before the pandemic had 4.5 times the odds (chance it happens versus chance it doesn’t) of losing or changing a job during it, compared with low-distress adults. Plausible range: 1.4 to 3.8.

Its own data show

The paper’s table prints the same 4.5 with a range of 1.4 to 13.8. Ranges of this kind reach up by the same factor they reach down: 4.5 is about 3 times 1.4, so the top should be about 3 times 4.5, near 14. Only 13.8 fits; the main text repeats the summary’s error (3.75).

A range of plausible values spans what the data can’t rule out, so the result must sit inside it. The main finding holds: the range’s bottom, 1.4, is still above 1, which would mean no difference. But the effect is far less certain than the summary says: anywhere from 1.4 to nearly 14 times the odds.

“The summary” means the paper’s abstract: the short summary printed at its top, which is the part most readers see.

Read the paper →Full technical write-up →

A longitudinal study of predictors of serious psychological distress during COVID-19 pandemic · Psychol Med (2021) · doi:10.1017/S0033291721004293