Deborah L. Delanoy Sources:
Automated Ganzfeld Methodology and Security
During her two decades at the University of Edinburgh‘s Koestler Parapsychology Unit, Deborah Delanoy helped design, defend, and stress-test one of the most security-conscious ganzfeld systems ever built. This spoke examines her contributions to automated ganzfeld methodology, the security architecture developed at Edinburgh, and the hard-won lessons about deception that shaped her approach to experimental control.
Deeper dives, Delanoy:
Key findings
- Delanoy was a co-author of the Edinburgh automated ganzfeld system’s security architecture, a computer-based design providing automatic data recording, shielding against sensory cues, and resistance to both subject bias and intentional experimenter bias.1
- The first full Edinburgh study using the refined automated procedure, run with an artistic population, produced significant positive direct-hit scoring while keeping receiver and experimenter blind to the sender’s presence or absence in two of three conditions.2
- An unexpected experimenter effect emerged in that study, one experimenter produced strong results while two scored at or below chance, and the authors discussed competing interpretations rather than claiming a single explanation.2
- Delanoy and colleagues extended the security mindset beyond the laboratory, designing a free-response methodology with sealed target containers that let participants test wherever and whenever they chose while protecting against sensory access to the target.3
- Her 1983/84 investigation of a self-alleged metal-bender, caught in fraudulent activity by hidden camera and later confessing to being a practising magician, became a documented case study in subject deception and informed her insistence on stringent controls.4
Overview
Delanoy’s involvement with the ganzfeld began with her doctoral work on training extrasensory perception within the ganzfeld setting at Edinburgh,5 and matured into a sustained concern with making the procedure as resistant as possible to error, leakage, and deception. The automated ganzfeld system built by the Koestler Chair of Parapsychology was explicitly developed as a means of replicating and extending the successful ganzfeld research conducted at the Psychophysical Research Laboratories (PRL), while confronting the security criticisms that earlier autoganzfeld work had attracted.1 Throughout this programme, Delanoy treated security not as an afterthought but as a design requirement on equal footing with psi-conduciveness.
The PRL Lineage and the Pressure of Criticism
The Edinburgh team framed their system’s development against a specific historical backdrop: as parapsychological testing procedures produce successful results, they attract increasingly sophisticated levels of criticism, including criticism of their security aspects, and safeguards against fraud or deviation from protocol are often challenged with regard to researchers as well as participants.1 The team noted that this is especially true for protocols involving very few individuals already regarded as talented, such as special sender-receiver pairs, many parapsychologists deliberately avoid star subjects partly to escape suggestions of fraud that tend to follow positive results.1 The Edinburgh refinement of the classical PRL autoganzfeld procedure, first presented at the 1994 Parapsychological Association convention, was designed to overcome various possible methodological problems in the PRL procedure that had been raised in methodological papers of the early 1990s.2
The Edinburgh Automated Ganzfeld System
The security paper Delanoy co-authored with Kathy Dalton, Robert Morris, Dean Radin, Robin Taylor, and Richard Wiseman described the Edinburgh system’s approach to the artifacts that had dogged earlier ganzfeld work, and offered recommendations for future improvements to automated ganzfeld settings.1 The system addressed three distinct artifact classes with specific countermeasures. Sensory leakage was addressed through what the authors described as highly effective shielding against sensory cues, combined with computer-controlled target presentation in a free-response testing framework, a control the authors presented as effective against direct cueing. Data-recording error and subject response-bias were addressed by automatic, computer-based data recording, removing hand-scoring from the chain, eliminating transcription-stage error as a contributor. Intentional experimenter bias was addressed by building resistance into the automated protocol itself, so that the procedure did not depend on any single experimenter’s honesty at the data-handling stage, a mitigation rather than a total elimination, as the authors’ own discussion of collusion scenarios acknowledged.1
System Architecture and the Collusion Argument
The Edinburgh automated system was designed as a free-response testing platform usable under a variety of experimental designs, including automated ganzfeld research. It was a computer-based system: the controlling program ran on a 33 MHz personal computer equipped with a 210 MB fixed disk and 8 MB of RAM, with a super-VGA monitor and printer, and the target presentation arrangement involved two PC/VCR units with audio recording equipment under program control.1 The system was the brainchild of several people, with hardware and software design primarily the responsibility of a dedicated engineer, and it received both praise and criticism from experimenters and critics alike during its development.1 On the fraud question specifically, the authors advanced a structural argument: in general, the more participants involved in a study, the less likely it is that deception has occurred, because a critic would need to posit increasingly complex collusion among different individuals. Distributing procedural roles across multiple personnel thus mitigated, though could not formally eliminate, the experimenter-fraud hypothesis.1
Putting the System to the Test
The refined system was not just a design exercise; Delanoy and her Edinburgh colleagues used it to ask whether positive ganzfeld scoring would survive the tightened conditions, and to probe a long-neglected process question, whether the sender actually matters. Their automated ganzfeld study with an artistic population (97 trials total, with 32 pre-specified trials for each of three sender conditions, participants drawn from an artistic population rather than pre-screened star subjects) produced overall positive scoring that the authors presented as further evidence for positive ESP results with the automated ganzfeld procedure, while the sender manipulation itself showed no significant differences between conditions.2 The blinding design directly targeted experimenter cueing: in two of the three conditions, both receiver and experimenter were blind as to whether a sender was present at all, eliminating the possibility that the experimenter’s knowledge of sender status could shape the receiver’s mentation or the judging process in those conditions.2
Direct Hits, Sender Conditions, and the Experimenter Effect
The overall psi result was 32 direct hits out of 97 trials, just statistically significant (p < .05, one-tailed, ES(h) = .18).2 Direct-hit results were nonsignificantly above chance in all three sending conditions, and an ANOVA on Stanford z-scores comparing the three conditions was nonsignificant (p = .775), giving no evidence that an active sender conveyed any tangible advantage.2 The correlation between z-scores and extroversion, a candidate individual-differences predictor from previous autoganzfeld research, was also nonsignificant (r = −.111).2 The most striking unplanned finding was an unexpected experimenter effect: one experimenter produced strong results while the other two scored at or below chance. The authors discussed and evaluated different interpretations of this effect rather than asserting one.2 This is a case where the data and the interpretation must be kept separate: the direct-hit scoring is the experimental result; whether it reflects anomalous information transfer, an experimenter-mediated process, or an undetected artifact remained explicitly contested within the paper itself, and the experimenter-effect finding, being unplanned and based on a three-way split of 97 trials, carries the usual caveats of exploratory subgroup analysis.2
The sender/no-sender comparison also illustrates how the Edinburgh group used the automated platform for process-oriented research once the procedure appeared methodologically sound, the explicit strategy stated in the study’s introduction was that established, tightened procedures could then be turned toward testing tentative models of psi functioning.2
Security Outside the Laboratory
Delanoy’s security thinking extended beyond the ganzfeld chamber. With Caroline Watt, Robert Morris, and Richard Wiseman, she developed a free-response ESP methodology for testing outside the laboratory, one that allowed participants to choose when, where, and how each trial was conducted, while keeping the actual target enclosed in a secure container in the participant’s possession.3 The motivating problem was a textbook security lesson. An earlier Edinburgh training study by Delanoy, Morris, and Watt had found no significant ESP scoring on any formal measure, yet strong above-chance scoring in informal practice exercises.6 Crucially, the authors made no claims for that informal scoring, because the practice exercises lacked adequate security precautions against sensory access to the target materials, the unsealed envelopes and self-administered handling procedures left the sensory-leakage hypothesis unresolved as an alternative explanation for the 60% hit rate.3 The new methodology was an explicit test of that alternative: could the apparently psi-conducive conditions of the informal exercises be closely approximated under genuinely secure conditions, and would the scoring survive?
The Outwith-the-Laboratory Study Design and Results
In the study testing the methodology’s efficacy, twelve participants had contributed 432 trials at the time of reporting. The first three authors (Delanoy, Watt, and Morris) acted as both experimenters and participants, but their own data were excluded from the primary measure because they were atypical of the other participants in several respects, a self-exclusion decision that removed the most obvious experimenter-as-participant contamination from the headline result. The nine non-staff participants contributed 366 trials and obtained significant positive psi scoring (p = 0.01, one-tailed, exact binomial).3 The methodology gave each participant possession of the sealed target container, a judging pack, and a means of obtaining feedback, including immediate feedback if desired, after completing the judging procedure.3 In the predecessor informal exercises, by contrast, target pictures had been housed in unsealed opaque envelopes that participants scrambled and blind-judged themselves, a procedure the authors explicitly declined to treat as evidential.3
The design philosophy here mirrors the automated ganzfeld work: identify the specific artifact (sensory access to unsealed targets), engineer a specific countermeasure (secure target containers with a structured judging protocol), and then test whether the effect persists. The positive scoring under the secured protocol is the datum; whether it reflects ESP, residual procedural weaknesses, or selection effects among the small experienced-participant sample (N = 9 non-staff contributors) was a question the conference report’s scale could not settle.3
Fraud, Deception, and the Security Mindset
Delanoy’s insistence on security architecture was grounded in direct experience of being deceived. Early in her Edinburgh years, she investigated a self-alleged PK metal-bending subject over a seven-and-a-half-month period in 1983/84, a cooperative, helpful seventeen-year-old who claimed he could bend metal at will and who participated in twenty sessions representing roughly sixty hours of laboratory work.4 Across attempts at controlled metal-bending, micro-PK, fire-raising, and metal-bending in the ganzfeld, the subject was unable to produce any PK under thoroughly controlled conditions. He was eventually caught engaging in fraudulent activity by means of a hidden camera, and subsequently confessed to being a practising magician.4 Delanoy’s published discussion drew the general lesson explicitly: researchers must never forget the possibility that their subjects may be presenting deceptive data, and must consider the ways researchers themselves may contribute unconsciously to their own deception.4
The Metal-Bender Case as a Methodological Document
The case unfolded in stages of escalating control. The first four sessions, conducted by Julie Milton and John Beloff, were deliberately informal to let the subject adjust to the laboratory; he produced several bent objects during this phase, but never under direct observation, a pattern that itself functioned as a diagnostic signal.4 Subsequent sessions implemented progressively stricter controls recommended for testing macro-PK claims, under which no genuine effects appeared, and covert video monitoring ultimately documented the fraud directly.4 Under the doctrine that methodologically uncontrolled observations yield no evidential weight in either direction, the bent objects from the informal phase counted for nothing, and the Edinburgh investigators treated them accordingly, withholding any positive claim until controlled conditions could speak. The case became one of the relatively rare published self-reports of a parapsychology laboratory detecting and documenting subject fraud from inside.4
What distinguishes Delanoy’s methodological writing is that she never treated security and psi-conduciveness as the same problem. Her 1989 tour of six U.S. parapsychology laboratories, undertaken to inform the design of the Koestler Chair’s new research facilities, catalogued the other half of the design space: laboratory environments tailored to be comfortable and welcoming, time spent putting participants at ease, experimenters chosen for social skill and positive expectation.7 The Edinburgh automated ganzfeld represents the synthesis she and her colleagues pursued: a system warm enough on the human side to give psi a chance to appear, and hard enough on the engineering side that, if it did appear, the result could withstand scrutiny.17 Her later overview of parapsychology in a university setting reflects the same dual commitment carried into her Northampton years, where she founded and first directed the Centre for the Study of Anomalous Psychological Processes.8
Modern Context
Modern-context corpus not yet populated for this topic. Curator will add 2-3 mainstream-science citations (Trends in Cognitive Sciences, Nature, PNAS, etc.) that bear on the methodology or interpretive frame of the experiments described above. This placeholder is audit-visible so the gap is explicit rather than silent.
References
- Dalton, K., Morris, R., Delanoy, D. L., Radin, D. I., et al. (1996). Security Measures in an Automated Ganzfeld System. Journal of Parapsychology, 60(2), 120–148. https://koestlerunit.wordpress.com/wp-content/uploads/2015/06/dalton-et-al-1996.pdf R001 [Dalton 1996] ↩︎
- Morris, R. L., Dalton, K. S., Delanoy, D. L., & Watt, C. (1995). Comparison of the sender/no sender condition in the ganzfeld. Proceedings of the 38th Annual Convention of the Parapsychological Association, 244-259. https://koestlerunit.wordpress.com/wp-content/uploads/2015/06/morris-dalton-delanoy-watt-1995.pdf R002 [Morris 1995] ↩︎
- Delanoy, D. L., Watt, C., Morris, R. L., & Wiseman, R. (1993). A new methodology for free-response ESP testing outwith the laboratory: Findings from experienced participants. Proceedings of the 36th Annual Convention of the Parapsychological Association. https://koestlerunit.wordpress.com/wp-content/uploads/2015/06/delanoy-watt-morris-wiseman-1993.pdf R003 [Delanoy 1993] ↩︎
- Delanoy, D. L. (1987). Work with a fraudulent PK metal-bending subject. Journal of the Society for Psychical Research, 54, 247–256. https://koestlerunit.wordpress.com/wp-content/uploads/2015/06/delanoy-1987.pdf R004 [Delanoy 1987] ↩︎
- Delanoy, D. L. (1986). Training of extrasensory perception in the ganzfeld. ERA. R005 [Delanoy 1986] ↩︎
- Delanoy, D. L., Morris, R. L., & Watt, C. (2004). A study of free-response ESP performance and mental training techniques. Journal of the American Society for Psychical Research, 98, 28-67. https://koestlerunit.wordpress.com/wp-content/uploads/2015/06/delanoy-morris-watt-2004.pdf R006 [Delanoy 2004] ↩︎
- Delanoy, D. L. (1997). Important psi-conducive practices and issues: Impressions from six parapsychological laboratories. European Journal of Parapsychology. https://koestlerunit.wordpress.com/wp-content/uploads/2015/06/delanoy-1997.pdf R007 [Delanoy 1997] ↩︎
- Delanoy, D. L. (2009). Parapsychology in a university setting. Parapsychology Foundation. R008 [Delanoy 2009] ↩︎