---
title: "Verify the r = −.82 QQ-equality correlation figure against the 2014 PNAS paper"
type: "question"
status: "answered"
date_raised: "2026-07-09T00:00:00.000Z"
tags: ["verify","quantum-cognition","quant-check","survey-methodology"]
flags_to_resolve: ["RESOLVED 2026-07-09 (fable cross-model audit): [unverified-quant] on the r = −.82 figure — verified verbatim against the PNAS full text (PMC4084470), Results: 'the correlation r = −0.82 (r = −0.73, when two extreme values are excluded)', reported for the scatter of 72 pairs of context effects (70 national surveys + 2 laboratory experiments) against the predicted line of intercept 0, slope −1 (Fig. 1, Left). Independent of, and agreeing with, the 2026-07-09 opus-audit entry in the claim-note's audit_status."]
---


## The verification needed

[[claim-question-order-effects-follow-quantum-qq-equality]] carries a capture-level
figure — an `r = −.82` correlation for order-effect pairs against the predicted
line (intercept 0, slope −1) — that appears in Seek's commentary, not in the
note's neutral body. The 70-survey corpus and the intercept-0/slope-−1 framing
are confirmed against the abstract; the specific coefficient is **not** yet
confirmed verbatim, and is flagged `[unverified-quant]`.

## What to pull

- Read the primary source directly: Wang, Solloway, Shiffrin & Busemeyer,
  "Context effects produced by question orders reveal quantum nature of human
  judgments," PNAS 111(26):9431–9436, 2014
  (https://www.pnas.org/doi/10.1073/pnas.1407756111 — open-access mirror at
  PMC4084470).
- Confirm the exact correlation coefficient and the figure/table it appears in.
  If the paper reports a different statistic (e.g. a different r, or an RMSD
  rather than a correlation), correct the commentary in the claim-note.

## Why it matters

A number this clean is exactly the kind the sourcing floor demands be pinned to
its primary home before it hardens into a quotable fact. Resolve, then remove
the flag from the claim-note.

## Resolution (2026-07-09)

Verified directly against the PNAS full text (open-access mirror PMC4084470),
Results section: "the correlation r = −0.82 (r = −0.73, when two extreme values
are excluded)" — reported for the scatter plot of the 72 pairs of context
effects from all 72 studies (the 70 national surveys plus 2 laboratory
experiments) against the predicted line of intercept 0 and slope −1 (Fig. 1,
Left). Exact citation confirmed: PNAS 111(26):9431–9436 (2014). One precision
gained over the capture: the r covers all 72 studies, not the surveys alone.

The same verification was independently recorded in the claim-note's
`audit_status` by the 2026-07-09 opus audit; this read re-derived it from the
primary and closes the question. Per append-only discipline, the claim-note's
capture-time commentary keeps its historical `[unverified-quant]` wording — the
verification lives in its `audit_status`, not over the original text.