Talk:Trial data comparison / Archive 1
Archived discussion·1 closed thread
Should the bars be there at all?Resolved
The bar column invites exactly the visual cross-trial comparison the lead spends a paragraph warning against. It is a graphical argument that the article's text contradicts. Propose removing it. — — PuffRemoverPax (talk) 11:07, 24 June 2025 (UTC)
Partly agree, but the bars are scaled within endpoint type, so a weight bar and an HbA1c bar are not on the same scale and cannot be compared. That is stated under the table. The remaining risk is comparison within an endpoint type, which is the legitimate reading as long as the population column is read too. — — DataTableDunja (talk) 15:29, 24 June 2025 (UTC)
Compromise: keep the bars, but colour them by endpoint type so that the visual difference between a weight bar and a hazard-ratio bar is obvious at a glance. A reader is much less likely to compare a purple bar with a blue one. — — ForestPlotFinn (talk) 09:52, 25 June 2025 (UTC)
That is a better answer than removal. Implemented and I withdraw the proposal. ✓ Resolved — — PuffRemoverPax (talk) 10:18, 27 June 2025 (UTC)