Research correction · 22 October 2026 · Version 1
Telephone Study: Correcting the Record
The completer counts reconcile. Statistical and scientific review remain open.
Wahbeh, Cannard, Radin and Delorme, Who’s calling? Evaluating the accuracy of guessing who is on the phone, EXPLORE 20(2), 239–247 (2024; online 2023). DOI 10.1016/j.explore.2023.08.008.
Current descriptive completer figures: 101/210 pre-selected responses (48.1%) and 205/630 post-selected responses (32.5%).
What the author confirmed
In research correspondence dated 22 October 2026, Dr Helané Wahbeh confirmed that three triads marked as completers in Figshare v1 had not finished all their post-selected trials. Correcting their completion status reconciles the completer counts with Table 3. Their responses remain in the including-dropout totals.
She confirmed Table 3, correction of the abstract’s pre-selected 50.0% to 48.1%, and that the first-six-trials restriction was already applied correctly in v1. This was a completion-status labelling discrepancy. The trial restriction must not be applied again.
Provenance: user-authorised paraphrase of private author clarification. The correspondence body, private links and participant identifiers are not reproduced. Author confirmation is distinct from independent scientific review.
Four summaries of the same study
| Population and version | Pre-selected | Post-selected | Status |
|---|---|---|---|
| Original abstract: completers | 50.0% / 210 responses | 31.9% / 630 responses | Historical source figures. |
| Figshare v1: original Completer labels | 115/228 · 50.4% | 216/666 · 32.4% | Correct recount of the original labels; not the corrected completer population. |
| Table 3 and author-corrected completers | 101/210 · 48.1% | 205/630 · 32.5% | Author-confirmed Table 3; conditional descriptive recount agrees. |
| All statuses, including dropouts | 147/289 · 50.9% | 243/762 · 31.9% | Unchanged; matches Table 3’s including-dropout rows. |
The 35 retained completer triads yield 35 × 6 × 1 = 210 pre-selected responses and 35 × 6 × 3 = 630 post-selected responses. The 630 response records are not 630 independent calls or participants.
The abstract’s post-selected 31.9% differs from Table 3’s 32.5%. It also equals the rounded including-dropout rate, 243/762. This arithmetic observation does not establish how the abstract value arose. Its explanation remains separately unresolved.
What AkashicNET checked
The unchanged Figshare v1 workbook matches the earlier source pin: 87,755 bytes, MD5 07158e08d430cd17345428cf76728382, SHA-256 recorded in the metadata below. Independent XML and openpyxl readers agree on all aggregates before and after the author-supplied status correction. Each retained triad has six pre-selected and eighteen post-selected response records.
Only completion-status membership was adjusted, conditional on the author’s mapping. No outcome-based exclusions or repeated trial truncation were applied. The workbook was not modified. Only aggregates are published.
This verifies descriptive arithmetic conditional on author information. It does not independently establish completion status, reconstruct chronological selection, reproduce inferential tests or validate the study conclusions. Public end-to-end reproduction awaits corrected labels, the trial-level mapping and scripts.
What remains pending
- Corrigendum: the author reports a correction is being submitted to EXPLORE. No publication link was supplied or located in the bounded checks.
- Figshare v2: corrected labels, deidentified trial-level data, scripts and documentation are being prepared, according to the author. Public article and version APIs returned only v1 at this check.
- Completer randomization: the author confirms correction is required. The published p = .054 is historical reporting. No corrected p-value, uncertainty interval or significance conclusion is available here.
- Post-selected abstract: the reason for 31.9% versus Table 3’s 32.5% remains unresolved.
- Independent scientific review: response frequencies, target assignments, dependence, protocol amendments, selection, leakage, multiplicity and any affected analyses require review.
The paper’s nominal 33.3% baseline and binomial results must not be substituted for corrected randomization. Unequal response use and the actual target-assignment structure matter. No inferential result has been calculated or inferred.
This correction does not establish or refute telepathy. BQ001 remains unresolved. No scientific evidence, model support or independent-review status is promoted.
Correction history and original editions
- 21 September: public summaries carried abstract-level figures of 50.0% and 31.9%.
- 12 October: the audit recorded the mismatch and partial Table 3 reconciliation in draft PR #399.
- 14 October: methodological inquiry and AI-transparency disclosure sent. Earlier contact-status checkpoints remain historical.
- 22 October: substantive author clarification received; conditional descriptive reconciliation completed; current public summaries corrected. Formal correction, v2, corrected statistics and independent review remain pending.
The original 14-page magazine PDF and 24-page special-edition PDF remain unchanged as historical editions. Their telephone figures and significance summaries predate this note. They are not corrected editions and must be read with this correction.
Updated illustrated feature · IONS notebook · AkashicPSI summary
Sources and access limits
- Journal DOI · Indexed abstract.
- Article-in-press paper: abstract p. 1, trial restriction p. 4, randomization methods p. 5, Table 3/result p. 6. This is not a corrected publisher edition.
- Figshare v1 · Public version history.
Publisher access was unavailable. No corrigendum or v2 link is invented. Human-directed, AI-assisted source inspection and descriptive analysis; independent scientific review pending. No institutional endorsement is implied.