Confirmatory Study of Anomalous Experiences in Enchanted Spaces

Houran, J., Laythe, B., & Little, C. (2024). Confirmatory Study of Anomalous Experiences in Enchanted Spaces. Journal of Anomalistics, 24(2), 396–426. https://doi.org/10.23793/zfa.2024.396

AI Assessment

A preregistered field experiment testing an “enchantment-psi loop”: the idea that immersive, atmospheric spaces (sacred sites, reputedly haunted locations, and augmented-reality environments) heighten both anomalous experiences and psi performance. With 22 “thin-boundary” participants toured through three enchanted conditions and a plain “disenchanted” baseline, the results were mixed and mostly did not support the psi predictions. Encounter experiences and felt enchantment were reliably higher in the enchanted spaces, but the predicted psi differences largely failed to reach significance (aggregate 1.85 vs 1.82, p = .23), and, contrary to the model, felt enchantment correlated negatively with psi. The study is preregistered with open data and unusually candid, all four deviations from the plan are disclosed, but the sample is very small, the facilitators were not fully blinded, and the central prediction did not hold. This audit reports what the study found and how; it takes no position on whether the enchantment-psi loop is real.

Provenance

DOI. 10.23793/zfa.2024.396 · Journal of Anomalistics (Zeitschrift für Anomalistik) 2024, volume 24, pages 396 to 426.

Study type. A preregistered, confirmatory field experiment with a within-participants (enchanted versus disenchanted) design.

Authors. James Houran (Integrated Knowledge Systems; Editor-in-Chief of the Journal of Scientific Exploration), Brian Laythe (Institute for the Study of Religious and Anomalous Experience), and Cindy Little (same institute), all researchers specializing in anomalous and haunt experiences.

Data availability. Design, analysis, and materials were preregistered on the Open Science Framework, with raw data subsequently deposited there (osf.io/czxfd).

Source basis. Every figure below is taken from the article’s own Abstract and Results. (A cataloguing note: the study database had stored this record’s title with the abstract text appended; the clean title, used here, is from the article.)

What the paper reports

The “enchantment-psi loop” proposes that anomalous experiences and psi feed one another in atmospherically charged settings, and that the effect can be predicted from a formula combining paranormal belief, transliminality (a tendency for material to cross between conscious and unconscious), and situational enchantment.¹ The authors specified four hypotheses and tested them by taking thin-boundary participants through sacred, haunted, and augmented-reality environments, plus a deliberately disenchanted control setting, measuring their encounter experiences, enchantment, and performance on a computerized psi test. The paper reports mixed support, with the experiential predictions faring better than the psi ones.

The results lend credence to aspects of the enchantment-psi loop yet indicate the need for future research to clarify important nuances or complexities with the model that likely affect the reliability or robustness of its predictions.

How it was run

• Participants. A vetted sample of 22 “thin-boundary” individuals, organized into small tour groups.
• Conditions. Counterbalanced immersive exposure to three “enchanted” environments (sacred, haunted, and augmented reality) plus paired testing sessions in a purposely “disenchanted” setting.
• Measures. Environmental features, real-time encounter experiences (a subjective-sensory-experience or SSE measure), enchantment levels (Enchantment-ACL), and a computerized test of putative psi (Psi-App).
• Disclosed deviations. Four departures from the preregistration are reported: an independent disenchantment baseline was used, the sample was smaller than planned, counterbalancing of site order was limited, and facilitators were not fully blinded to the hypotheses.

Results, as reported

Values are reproduced from the article’s Abstract and Results. The experiential half of the model held (enchanted spaces produced more encounter experiences and felt enchantment), but the psi half largely did not: the key aggregate psi comparison was non-significant, and the central enchantment-to-psi relationship ran in the opposite direction to the prediction.

Eleven-dimension audit

Pre-registration

A genuine strength: design, analysis, and materials were preregistered on OSF with data later deposited. Crucially, the authors disclose all four deviations from the plan rather than quietly absorbing them, which is exactly how preregistration is supposed to work. The deviations do, however, weaken the confirmatory guarantee, as noted below.

Randomization

Exposure to the three enchanted conditions was counterbalanced, but one disclosed deviation is that counterbalancing of site order was limited, so order effects cannot be fully ruled out in such a small sample.

Sensory leakage

The psi measure is a computerized test, which controls the psi task itself. The larger validity concern is expectancy and demand around the subjective experience measures, given that facilitators were not fully blinded (see Blinding), which is the relevant analogue of leakage here.

Blinding

A disclosed weakness: facilitators were not fully blinded to the hypotheses. For measures as suggestible as felt enchantment and encounter experiences in atmospheric settings, unblinded facilitators are a real avenue for demand characteristics to inflate the experiential effects, which are precisely the effects that reached significance.

Optional stopping

Not optional stopping in the peeking sense, but the sample was smaller than planned, another disclosed deviation, which leaves the study underpowered and means the null psi results are as consistent with low power as with a true absence of effect.

Outcome measure

Multiple measures (encounter experiences, enchantment, and a computerized psi score). They are reasonable operationalizations, but they mix subjective self-report (SSE, enchantment) with an objective psi task, and it is the subjective measures that carried the significant effects.

Effect size

Split. The experiential effects (enchanted greater than disenchanted for SSE and enchantment) were significant, while the psi effects were small and mostly non-significant (aggregate 1.85 vs 1.82, p = .23). The one clearly significant psi-related correlation ran against the model (r = −.24).

Multiple comparisons

Four hypotheses were tested across three enchanted conditions and several measures, plus tangential emotional, motivational, and environmental analyses. Even with preregistration, this is a broad space, and the significant findings are concentrated in the experiential measures rather than the psi predictions the study was designed to confirm.

Internal replication

The three enchanted conditions provide some internal variation, and the experiential pattern is fairly consistent across them; the psi pattern is not, with above-chance scores appearing in the sacred and haunted conditions but the overall comparison failing to reach significance.

External replication

This is a novel test of a specific model rather than a replication, and its central psi prediction did not hold. Given the small sample and the contrary correlation, independent, better-powered replication is the clear next requirement.

Transparency

High relative to the field: preregistered, open data, and, most tellingly, full disclosure of all four deviations and of the finding that ran opposite to the model. Reporting that enchantment correlated negatively with psi, when the whole premise predicts a positive loop, is a mark of honest reporting.

The adversarial record

• The central prediction failed. The model predicts that enchantment and psi reinforce each other, yet enchantment correlated negatively with putative psi (r = −.24, p < .01), and the aggregate psi comparison was non-significant. On its own primary terms, the study did not confirm the psi side of the loop.
• Very small and underpowered. With 22 participants, smaller than planned, the psi nulls are hard to interpret, and the significant experiential effects rest on a slender base. Small samples in suggestive settings are exactly where demand effects are most influential.
• Unblinded facilitators, subjective measures. The effects that did reach significance are subjective experience and enchantment ratings, collected by facilitators who were not fully blinded to the hypotheses. That combination invites the reading that the experiential effects reflect expectancy rather than something anomalous in the spaces.
• A complex, flexible model. A Paranormal Belief by Transliminality by Enchantment formula, tested across multiple conditions and measures, offers many ways to find partial support even under preregistration; the honest summary of “mixed support” is the appropriate one.
• What is genuinely strong. The study is preregistered with open data, discloses every deviation, and reports the result that contradicts its own model. That transparency is the reason a reader can see clearly that the experiential effects held while the psi predictions did not, which is the correct and honest bottom line.