Discovery Lab Answers

Rick E. Berger, PhD is an experimental social psychologist and anomalous cognition researcher whose work spans three distinct lines: building automated experimental infrastructure, conducting forensic audits of existing psi datasets, and analyzing geophysical correlates of psi effect sizes. He is profiled at Rick E. Berger.

A necessary caveat before the detail: the library’s share of the published literature on Berger’s work has not been measured. What follows summarizes the studies currently in the ESP-Nexus library — treat it as a sample, not a settled account of his full record. The structured evidence table below covers three result rows across different, non-comparable metrics; directions disagree across them, and at least one result is below-chance.

Experiments

The autoganzfeld system — design and PRL database (Honorton et al., 1990)

The most consequential project Berger was involved in is the autoganzfeld system. Berger designed and built the hardware-and-software platform at the Psychophysical Research Laboratories (PRL) in Princeton; Charles Honorton was principal investigator and led the research program. The system ran four experimental series with eleven ganzfeld experiment conditions. An eighth-person team including Berger, Varvoglis, Quant, Derr, Hansen, Schechter, and Ferrari contributed to the database.

Geomagnetic correlates of psi effect sizes — Berger and Persinger (1991)

Working with M. A. Persinger at Operational Technologies Corporation and Laurentian University, Berger examined whether annual geomagnetic activity correlated with psi effect sizes across all published experimental psi studies from 1882 onward — a span of roughly six decades. This is a dataset-level comparative analysis, not a new psi experiment.

Forensic audit of the Blackmore psi corpus (1989)

Berger conducted a systematic critical examination of Susan Blackmore’s accumulated ESP experiment database, going back to raw source materials, unpublished dissertation data, and original reports. A related audit examined the Spinelli dissertation data on ESP in children.

Target-set distribution forensics — Bierman, Broughton, and Berger (1998)

Berger co-authored with Dick J. Bierman and Richard S. Broughton a re-examination of the PRL autoganzfeld target and target-set distributions, distinguishing between the soundness of a randomization procedure and the properties of the empirically realized sequence it produces.

Methodology

Autoganzfeld engineering

The autoganzfeld system replaced human-handled target selection, stimulus delivery, blind judging, and data logging with computer control. The design allowed a single target-blind experimenter to run a complete session. It introduced computer-accessed videotape, enabling dynamic (film-clip) targets alongside static images. The explicit goal was to close the procedural gaps — multiple-contact between sender and experimenter, manually handled target materials, non-blind judging — that critics had identified in the prior manual ganzfeld literature.

Target-sequence forensics

The 1998 paper introduced a formal distinction between two levels of randomization quality: (1) procedure integrity — whether the algorithm or hardware generator theoretically excludes sequential dependencies — and (2) sequence integrity — whether the empirically realized sequence, even from a sound procedure, contains distributional anomalies. A post-hoc correction adjusted hit-probability estimates for observed distributional skew; the paper documented that this correction reduced effect-size estimates but left a residual excess of hits not fully explained by the artifact.

Forensic auditing standard

In the Blackmore and Spinelli audits, Berger applied what the ESP-Nexus page describes as a bidirectional standard: the same scrutiny applied to apparent positive results was applied to apparent null results, on the principle that methodological flaws cut in both directions. The Blackmore audit documented pervasive inconsistencies between unpublished dissertation data and published reports.

Geomagnetic analysis design

The Berger–Persinger (1991) analysis computed standardized effect sizes for all experiments published across the study window and correlated them with annual geomagnetic activity indices, including lag/lead analyses across a ±3-year window to test whether any year other than the year before publication drove the relationship.

Data

The three result rows in the library cover different metrics and must not be pooled. The table below carries the figures; prose follows with interpretation.

Berger and Persinger (1991) found that years with quieter annual geomagnetic activity were associated with larger psi effect sizes across the full multi-decade dataset (k = 185 experiments), with the prior year’s geomagnetic index as the dominant source of variance.

Berger (1989) — the chronological ordering analysis within the Blackmore corpus — found a below-chance, psi-missing result with a Pearson r of −0.8 across k = 6 studies, at p = .056, indicating a decline over time in ESP scoring rather than a positive effect.

Berger (1987) — representing the autoganzfeld feedback-with-silent-saved (FSS) condition — found a positive correlation of 0.126 at p = .02.

These three rows use different metrics on different research questions and cannot be combined into a single summary figure. Directions are split: one below-chance result, two positive results, and one of the positive results (the geomagnetic correlation) is itself measuring a correlate of psi scoring rather than a direct hit-rate outcome.

Skeptical critiques

What critics argue.

Hyman (1985) advanced the “dirty test tube” argument applied broadly to positive psi evidence: that methodological flaws of any kind in a study constitute grounds to set its results aside, regardless of the direction of those flaws. Berger’s 1989 audit of the Blackmore corpus directly engaged this position, noting that Blackmore’s own published reports invoked methodological flaws selectively — flaws were cited when results were positive but not subjected to the same scrutiny when results were null.

Blackmore (1984) had previously published a critique of the Spinelli ESP-in-children data; Berger’s 1989 audit of the Spinelli dissertation found independent grounds to question the data’s integrity, reaching conclusions about data anomalies through a different evidentiary route than Blackmore’s critique.

On the autoganzfeld target distributions, Bierman, Broughton, and Berger (1998) themselves raised the critique — from inside the PRL research group — that the empirically realized target sequences contained distributional skew sufficient to affect effect-size estimates, and that this had not been adequately accounted for in prior analyses. The post-hoc correction they applied reduced but did not eliminate the residual excess of hits.

What the experimental data show.

The 1989 Blackmore audit found that discrepancies between unpublished dissertation data and published reports were pervasive enough that no conclusions — for or against psi — could be drawn from that database as a whole. The Spinelli audit arrived at parallel findings about data anomalies in that corpus. The 1998 target-distribution paper documented that the PRL autoganzfeld’s realized sequences departed from ideal random distributions, that a correction reduced estimated effect sizes, and that a residual excess of hits remained after that correction.

Analysis.

The Berger (1989) Blackmore audit applied the bidirectional standard explicitly: the same methodological scrutiny that skeptics apply to positive results was applied to Blackmore’s null results, with the finding that the null conclusions were equally unsupported. Whether that bidirectional standard is consistently applied across the field, and how much of the residual excess in the autoganzfeld data survives the 1998 distributional correction, are questions the retrieved sources raise but do not close. Independent replication of the autoganzfeld protocol outside PRL and the subsequent Edinburgh laboratory is a continuing reference point in discussions of the database’s evidential weight.