The headline is broadly defensible, but the qualifications matter. Effect sizes vary by population, the strongest claims rest on shorter trials, and credible voices push back on how it's typically framed.
If they looked at the success rate of the subjects, what they found was that the subjects that were told that did not work, please try again, had a 68% success rate. 68% of them went on to successfully program this cursor moving through the maze. Whereas the ones that were told you lost five points had a 52% success rate, which is a significant difference.
Every Sunday: the week’s new conflicts and verdict changes — and nothing else.
Native comments, Twitter mentions, and Reddit threads about this claim — surfaced together so the conversation isn't fragmented across platforms.
Bookmarking — the dossier-vs-overview split is the right call. Most of the time I want overview; sometimes I want receipts.
Would love a "what would change this verdict" RSS feed. Sign me up if it exists.
the subjects that were told that did not work please try again tried many many more times per unit time in other words they made more attempts at programming this thing to allow this cursor to move through the maze whereas the people that were told you lost five points gave up earlier or gave up entirely
what they found was that the subjects that were told that did not work please try again had a 68% success rate 68% of them went on to successfully program this cursor moving through the maze whereas the ones that were told you lost five points had a 52% success rate which is a significant difference