Irwin & Mandel (2023) frame their lab finding as cohering with a pattern Friedman & Zeckhauser (2012) had already found in declassified intelligence products

intelligence-tradecraftepistemicsprobability-confidencedavid-r-mandelsource-evaluation

Irwin & Mandel claim their laboratory conflation of confidence and probability as their own headline result — "Critically, we found that intelligence consumers do not treat confidence and probability as independent constructs" — but in the next breath situate it against an earlier, non-experimental evidence source rather than presenting the ground as untouched: "findings cohere with evidence from declassified intelligence products that analysts conflate ordinal confidence ratings with event probabilities (Friedman & Zeckhauser, 2012)." That earlier work — Friedman, J. A., & Zeckhauser, R. (2012), "Assessing uncertainty in intelligence," Intelligence and National Security, 27(6), 824–847 — examined real declassified assessments rather than lab-elicited judgments, and is the field's real-world anchor for the same claim this vault's lab-derived notes track. Irwin & Mandel also note a partial counterpoint from the same author pair's later work — "previous research suggests that experts and non-experts can distinguish between probability and specific components of confidence (Friedman & Zeckhauser, 2018)" — before arguing their own results show the disjunction fails in practice regardless.

This gives the two-axes cluster its one instance of triangulation across method: a lab manipulation (Irwin & Mandel) and a field/archival study (Friedman & Zeckhauser) converging on the same conflation, rather than the cluster's more usual pattern of one paper's finding standing alone. Neither Friedman & Zeckhauser paper has been read at primary by this vault; this note records only Irwin & Mandel's own framing of that earlier work, not an independent verification of it.