Discovery Lab Answers

The Honorton–Hyman disagreement is one of the most substantive methodological exchanges in parapsychology’s history. It began as an adversarial debate, produced a landmark joint document, and left several specific questions genuinely open. Before going into detail: the studies in the ESP-Nexus library represent an unmeasured share of the published literature on this question, so treat what follows as a summary of what the library holds rather than a settled account of the field.

Charles Honorton was a parapsychology researcher profiled on ESP-Nexus at https://esp-nexus.org/scientists/charles-honorton/.

Experiments

The core of the Honorton–Hyman exchange centered on the ganzfeld database — a body of psi experiments accumulated through the 1970s and early 1980s using sensory attenuation procedures. Honorton published a meta-analysis of 28 direct-hit ganzfeld studies in the March 1985 issue of the Journal of Parapsychology, explicitly as a response to Hyman’s own critical appraisal of the same database published in the same issue. Both men analyzed the same corpus; they reached different conclusions about what it showed.

Following their joint communiqué, Honorton built the autoganzfeld system at the Psychophysical Research Laboratories (PRL) in Princeton — a computer-automated testing environment designed to satisfy the methodological standards he and Hyman had co-authored. The resulting series formed the basis of the 1994 Bem and Honorton Psychological Bulletin paper, which Hyman then critiqued in the same issue.

Methodology

The 1985 exchange foregrounded three categories of methodological dispute:

• Selective reporting. Hyman argued the published database likely omitted null-result studies, inflating the apparent effect. Honorton countered that a trim-and-fill analysis made an implausibly large file-drawer necessary to explain away the results.
• Multiple analysis. Hyman identified that many studies used several outcome indices without pre-specifying which was primary, making chance significant results more likely. Honorton’s 1985 meta-analysis addressed this by restricting to a single pre-chosen index — direct hits — across all 28 studies.
• Methodological flaws and their correlation with results. Hyman identified 12 potential flaws and argued their presence correlated with significance. Honorton disputed the coding and the magnitude of this relationship.

The joint communiqué was the practical resolution at the procedural level: Honorton and Hyman co-authored a set of specific design and reporting recommendations for future ganzfeld research — including pre-registration of protocols, automated target selection and recording, blind judging, and randomization checks. Honorton then built the autoganzfeld to implement those standards directly. The communiqué documented what both agreed the existing database showed, and separately documented where they continued to disagree — most critically, over whether a demonstrated effect, if real, constituted evidence for psi.

Data

The evidence rows in the ESP-Nexus library span different, non-comparable metrics and cannot be pooled into a single figure. What the rows show:

All five rows in the library point positive. However, the corpus coverage for this question has not been measured, so this pattern reflects what the library holds rather than the full published record.

Skeptical critiques

What critics argue.

Hyman (1985) argued — before the joint communiqué — that Honorton’s meta-analysis could not be taken as evidence for psi because the database suffered from multiple undisclosed analysis paths: studies had used several outcome measures without pre-specifying which was primary, so significant results were more likely by chance than the reported p-values indicated. Hyman (1985) also argued that methodological flaw scores correlated positively with study outcome, meaning better-controlled studies tended to show smaller or null effects.

After the autoganzfeld series, Hyman (1994) shifted the critique to the reformed database. He argued that the hit rate in the autoganzfeld was partly an artifact of target-pool composition: dynamic targets (video clips) yielded substantially higher hit rates than static targets (photographs), and the mix of target types across studies was not uniform or pre-specified. He contended that target-type confounding, rather than a genuine anomalous signal, could account for the pattern Bem and Honorton reported. Hyman (1994) also raised the issue of optional stopping — whether the series was truly pre-specified as to stopping rule — and questioned whether the autoganzfeld results were genuinely independent of the earlier database Honorton had already analyzed.

What the experimental data show.

The autoganzfeld series used automated target selection and recording, blind judging, and pre-specified primary outcomes — the specific procedural requirements Hyman had co-authored in the joint communiqué. Utts (1991) noted, reviewing the broader ganzfeld database, that while Hyman’s flaw-correlation argument was the first to attempt quantification of the relationship between procedural problems and outcomes, the relationship he identified remained disputed as to magnitude and interpretation.

Analysis.

Honorton and Hyman co-authored both the problem statement and a partial resolution: they agreed the existing database showed an overall significant effect not reasonably explained by selective reporting or multiple analysis alone, and they agreed on the design standards needed for a decisive test. Where they continued to diverge was the inferential step — whether a replicated anomalous effect, even a methodologically clean one, should be interpreted as evidence for a genuine psi process or as an unidentified artifact. That question was not resolved in the communiqué and was still active in Hyman’s 1994 commentary. Milton and Wiseman (1999) later noted that meta-analytic investigation of variables Bem and Honorton identified as important in the PRL work indicated other experimenters had not replicated those moderator effects in the few areas where this had been attempted.

For the full spoke page on the communiqué and its methodological standards, see Honorton: Hyman–Honorton Joint Communiqué and Methodological Standards.