[Partial Retraction] Age Gaps and Birth Order Effects

A partial-retraction post modeling exemplary epistemic hygiene. LessWrong user Bucky re-analyzed Scott's SSC-survey birth-order/age-gap results and found he'd bungled ~1.5 analyses (a mislabel and a double-count), so the striking 'cliff' where birth-order effects vanish after a 7-year age gap is probably a coincidence of the exactly-one-sibling subsample rather than a real phenomenon (the firstborn-overrepresentation finding and the basic conclusion survive). Scott partially retracts, adds a bold note to the original post pointing readers to the failed replication, logs it on his Mistakes page to help readers calibrate trust in his future results, writes this post specifically so the correction gets as much prominence as the original, and reiterates that he makes the raw data public precisely so others can check him. Short and primarily a correction, but a model of how to publicize a failed replication of your own work.

Quality 58 · Solid. High-Solid (58): an exemplary, admirably-transparent self-correction (publicizing a failed replication of his own finding, logging it for calibration, giving it equal prominence) -- valuable as epistemic-hygiene modeling, but short and primarily a correction with modest standalone content.

Claude’s paradigm shift 30 · Slight. Slight (30): the value is the hygiene and transparency, not novelty; it's a correction notice.

Real-world impact 2 · Minor. An exemplary, transparent self-correction publicizing a failed replication of his own birth-order finding; valuable epistemic-hygiene modeling within the discourse, no material-world reach → RWI 2.