Participants in the Delanoy, Morris and Watt study performed informal practice exercises roughly once a day, six days a week, over a twelve-week period, in conjunction with practice of a mental technique. Across more than 700 informal practice trials, the obtained hit rate was 60% against a mean chance expectation of 50%, strong above-chance scoring on a measure devised solely to give participants practice.² The procedure used six numbered target packs delivered weekly, each consisting of two pictures housed in unsealed, numbered opaque manila envelopes; the participant generated impressions to a target, scrambled the envelopes until unsure which was which, and then blind-judged the pictures.² The authors made no claims for this scoring rate because the exercises lacked adequate security precautions to protect against sensory access to the target materials, the specific artifact (sensory leakage from unsealed envelopes in the participant’s own possession) was unresolved, not merely mitigated.² Applied symmetrically, an insecure procedure yields no evidential weight in either direction: the 60% rate is not evidence of trained ESP, and the formal null is not evidence that training fails. What the result legitimately did was motivate a better-secured design. The paper also set out Morris’s three-part framework for training research: trainee orientation (open and positive toward psi, a healthy self-attitude, comfortable with controlled experimental participation, given preliminary exposure to the specific environment and procedures); trainer characteristics (trained in how to train others, personal experience with the techniques taught, good with people, accepting of psi); and training content (basic mental skills of relaxation, concentration, and visualization, plus psi practice).¹

At the time of reporting, twelve participants had contributed 432 trials. The three experimenter-authors (Delanoy, Watt, and Morris) also acted as participants, but their data were excluded from the study’s primary measure because they were atypical of the other participants in several respects. The nine non-staff participants, described as experienced participants, had contributed 366 trials, obtaining significant positive psi scoring (p = 0.01, one-tailed, exact binomial).² The immediate-feedback option built into the judging procedure reflects the learning-theory rationale Delanoy and colleagues had reviewed in the training literature: feedback after correct responses should increase motivation, decrease boredom, and facilitate the learning of subtle internal cues associated with correct responding.¹

Sixty-four University of Edinburgh students (ages 19 to 30, median 21) participated in 32 sessions, each session involving a pair of participants, agent and percipient, who were close friends; each took part in one session only, and the sole selection criterion was willingness to participate in a parapsychology study (an opportunity sample, unscreened for prior psi performance).⁴ Conscious responses were elicited from the percipient for only half of each session, pseudo-randomly assigned to the first or second half, on the stated reasoning that recording the physiological measure was unlikely to interact with making a conscious response, although the opposite effect could occur. Significant psi-hitting was predicted for both psi measures, along with correlations between rating-scale data and both measures; analyses examining interactions between the two psi measures were designated exploratory in advance.⁴ The study built on an unpublished undergraduate project supervised by Delanoy, in which a conscious-response measure with emotional targets yielded significant overall psi-hitting, significant scoring for the sending of positive emotions, barely significant scoring for neutral ones, and near-chance scoring for negative emotions, together with rating-scale correlations between the agent’s degree of involvement in sending and psi scoring across all emotional conditions.⁴ The experimenter for all sessions was the second author, a medical student at Edinburgh with no previous experimental experience, trained by Delanoy on the basis of training she herself had received from William Braud, who visited the Edinburgh laboratory in 1993 in connection with the DMILS collaboration.⁴

The Morris, Dalton, Delanoy and Watt study ran three conditions, sender absent with receiver and experimenter blind to the sender’s presence or absence; sender present with the same blinding; and sender present with receiver and experimenter aware, with 32 pre-specified trials per condition and a total of 97 trials due to the random assignment of the first two conditions.⁶ The overall result was 32 direct hits out of 97 trials, just statistically significant (p < .05, one-tailed, ES(h) = .18), described by the authors as further evidence for positive ESP results with the automated ganzfeld procedure. Direct hits were nonsignificantly above chance in all three sending conditions; an ANOVA on Stanford z-scores across the three conditions was nonsignificant (p = .775), and the correlation between z-scores and extroversion was nonsignificant at r = −.111.⁶ The unexpected experimenter effect, one experimenter producing strong results while the other two scored at or below chance, was not predicted; the authors presented and evaluated several interpretations without settling on one.⁶ That open finding connects directly to the experimenter-orientation category of Delanoy’s six-laboratory survey, which had flagged experimenter expectation and preparation as candidate moderators deserving experimental test.⁵

The subject, a seventeen-year-old pupil at an Edinburgh school with a verified school reputation as a metal-bender, participated in twenty sessions representing approximately sixty hours of work at the Psychology Department of the University of Edinburgh.⁷ The first four sessions, conducted by Julie Milton and John Beloff, were deliberately informal to allow him to adjust to the laboratory; he produced several bent objects during these, but never under direct observation. Subsequent work included attempts at controlled metal-bending, micro-PK, fire-raising (an ability he claimed to develop during the collaboration), and metal-bending while in the ganzfeld; throughout, he was unable to produce any PK under thoroughly controlled conditions.⁷ On Delanoy’s own framing the case yields a documented confession of fraud and nothing more: the bent objects produced outside direct observation carry no evidential weight for PK, and the controlled nulls carry no evidential weight against PK in general, the case’s value is methodological, as a worked study in how deception unfolds and how investigators can be complicit in it without knowing.⁷