Published by BitPlan
Published by BitPlan Loyalty Inc., the company behind BitStage and Barnes Signal Advisor. We publish what we find — including about our own products.
Find a number we have overstated and we will correct it in public, with a timestamp and a note saying what we got wrong. We will not silently edit. Every correction gets an entry in The Ledger.
corrections@proofoverpromises.org
Read by Tim Barnes, who also runs BitStage. That is a conflict. It is disclosed in the masthead of every page.
Retracted the Goldin & Rouse (2000) blind-audition finding from our own materials before publishing this site.
Was citingThe widely-quoted "blind auditions increased women's success ~50%" figure, previously in our internal decks as evidence for blind evaluation.
Now citesKline, Rose & Walters (2022) on employer-level discrimination in the resume screen — a substantially stronger design for the same underlying claim.
Downgraded every validity coefficient on this site to the Sackett et al. (2022) revised figures — including the ones least flattering to our own product category.
Was citingSchmidt & Hunter (1998): work samples ρ = .54, cognitive ability ρ = .51, structured interviews ρ = .51.
Now citesSackett et al. (2022): work samples ρ = .33, cognitive ability ρ = .31, structured interviews ρ = .42 — with the live Oh et al. (2023) dispute stated alongside.
Corrected the cost-per-hire figure in evidence entry A5, which overstated the cost of hiring by roughly three and a half times in our own favour.
Was citingApproximately $4,700 average cost per hire, attributed to SHRM.
Now citesSHRM's current published figure: a $1,300 median for nonexecutive roles, from 4,657 respondents fielded November 2025 to January 2026.
A corrections log with only flattering corrections is not a corrections log. If this page ever stops growing, assume we stopped looking — not that we stopped being wrong.