| Abstract/Results: | ABSTRACT:
Introduction
With a few exceptions, meta-analyses of ganzfeld ESP studies have tended to obtain results supporting the ESP hypothesis, suggesting that this line of research merits further effort. While some have suggested that questionable research practices (QRPs) may inflate the overall hit-rate (e.g., Bierman et al., 2016; Hyman, 2010), internal patterns have been detected within groups of studies. Such patterns, notably correlations between participant characteristics and study outcome, are less likely to be attributable to questionable research practices and may reveal something about underlying mechanisms. Therefore, some reviewers recommend that researchers aiming to maximise ganzfeld task performance should only recruit selected participants (Baptista et al., 2015; Storm & Tressoldi, 2020).
Our previous pre-registered exploratory ganzfeld study with 60 selected participants tested by three experimenters obtained results supporting the precognition hypothesis (Watt et al., 2020). The present paper reports our follow-up confirmatory study seeking: to optimise psi-conduciveness with participant and experimenter selection and by limiting how many participants each experimenter could test; to optimise security including by using a precognition design; and to minimise QRPs with pre-registration and sufficient power for a confirmatory study. We also aimed to explore the previously observed link between self-reported creativity and ganzfeld task performance using validated creativity measures.
Methods
Following initial software testing (CW and AP) and experimenter selection and training (CW), a ganzfeld precognition design was employed to test 240 participants selected for self-reported creativity as well as practice of a mental discipline (such as meditation) and/or prior psi experience and/or psi belief. Fifteen experimenters, selected on the basis of their interest in parapsychology and positive attitude to psi, recruited and tested up to 20 participants each. Experimenters were encouraged to take time to set the participant at ease and establish rapport, and before ganzfeld testing participants completed the Australian Sheep-Goat Scale (ASGS), the Inventory of Creative Activities and Achievements (ICAA), and the Runco Ideational Behaviour Scale (RIBS).
The ganzfeld laboratory, furnished so as to feel comfortable and welcoming, was located in a quiet and secure basement room of a university building. We used the same target pool of 200 video clips as in our 2020 study, and a computer program guided the experimenter and participant through each session. The 25-minute impression period was preceded by a 9-minute progressive relaxation exercise. Data security measures included recording session data in triplicate to make any tampering detectable, recording of mentation period, and upload of data during each session to a remote server to which none of the experimenters or PI (CW) had access. The study design and analysis plan was pre-registered (for further details including a priori power analysis, see Watt, 2023).
H1 (Confirmatory): Participants would correctly identify the randomly selected target clip at greater than chance expectation (tested using exact binomial probability, N direct hits in 240 trials, p.hit = .25). For the following exploratory hypotheses, ganzfeld task performance was indexed by the session Z-score (calculated from target and decoy ratings).
H2: Paranormal belief (ASGS score) would correlate positively with ganzfeld task performance (one-tailed).
H3a and H3b: Creativity (ICAA and RIBS scores) would correlate positively with ganzfeld task performance (one-tailed).
The study was conducted in three waves: Wave 1 (60 trials, Jan – Mar 2023); Wave 2 (120 trials, Jun 2023 – Mar 2024); Wave 3 (60 trials, Nov 2023 – Mar 2024).
Results
Seventy-two hits were obtained out of 240 sessions, a 30% hit-rate [p = .043 (1-t), z = 1.714, ES (Z/vn) = 0.11], thus supporting the study’s confirmatory hypothesis of above-chance scoring on the precognition test. Having seemingly obtained evidence for psi, in an attempt to exclude the possibility of undetected ganzfeld program error or bias, we conducted further post hoc investigations including a month-long independent software validation exercise. For the latter, a programmer who specialised in software validation created a “puppeteer” program to remotely conduct over 5000 speeded-up and over 280 real-time “sessions” in the ganzfeld lab using the same study hardware and software, but with no person present. This exercise found no subtle bias or software/RNG errors that might have inflated the hit-rate.
We also explored the possibility of experimenter error or fraud by inspecting the profile of scoring across experimenters. Ten out of 15 experimenters obtained a hit-rate above 25%, so the study’s elevated hit-rate is not due to just one or two “outlier” experimenters
For the study’s exploratory hypotheses, no significant correlations were found between psi task performance and paranormal belief and creativity questionnaires.
Discussion
Our participant selection criteria meant that participants had relatively high levels of creativity and paranormal belief/experience, which may have undermined our exploratory process-oriented analyses. The study’s confirmatory hypothesis was significantly supported, indicating that participants could identify the randomly selected future target. This result shows that optimum security and psi-conduciveness can productively co-exist.
Our study also pioneered the use of software validation methods in ganzfeld research and we recommend this to other researchers, though it is better practice to conduct such an exercise before any formal data collection. Systematic incorporation and reporting of software validation (Kennedy, 2016) is another way in which parapsychologists can set a good example for the wider scientific community.
|
Reference:
| Watt, C. A., & Pooley, A. L. (2025). Ganzfeld ESP research: Building on lessons learned. 67th Annual Convention of the Parapsychological Association, Haus zur Lieben Hand, Albert-Ludwigs-University Freiburg, Germany.
|