Overview

Bancel’s research on forced-choice precognition and collaborative platforms addresses a fundamental problem in parapsychology: how to conduct well-controlled, replicable experiments at scale. Forced-choice protocols (in which participants select from a predefined set of randomized alternatives) have historically produced positive cumulative results, yet individual experiments often fail to replicate.¹ Bancel’s approach combines two innovations: the development of Internet-based platforms that allow geographically distributed participants to contribute data from home, and the use of pre-registered hypotheses with selected cohorts to test specific predictions about psi expression.

His work is situated within the Institut Métapsychique International (IMI) in Paris, where he has collaborated with colleagues including Mario Varvoglis, Jocelyne Boban, and Anaïs Bensahra. The evolution from earlier platforms (Sharefield, Selfield) to the current Psi@Home system reflects a sustained effort to refine experimental design and address methodological challenges that have historically limited forced-choice research.

Forced-choice methodology in parapsychology

Forced-choice experiments represent one of the oldest and most rigorous paradigms in parapsychology. Unlike free-response protocols such as remote viewing or ganzfeld studies, forced-choice tasks require participants to select from a limited, predetermined set of options, for example, guessing which of four symbols will appear next, or predicting the outcome of a coin flip. The advantage of this approach is that chance performance can be calculated precisely, and statistical deviations from chance are unambiguous.

Bancel notes that despite the accumulated positive evidence from forced-choice studies, the challenge of replicability remains acute.¹ Even well-powered, presumably well-designed experiments often fail to produce significant effects when conducted as one-off studies. This inconsistency has motivated Bancel’s shift toward collaborative platforms that can gather large datasets rapidly, enabling meta-analytic approaches and the detection of subtle patterns that might be obscured in smaller samples.

The forced-choice framework also allows for precise measurement of variance in performance across sessions, a metric that has emerged as potentially more informative than overall hit rate. Rather than asking whether participants score above chance on average, researchers can ask whether performance varies significantly across sessions in ways that suggest a mixture of psi-hitting and psi-missing, or whether certain conditions or participant types show more variable results.

The Psi@Home platform

Psi@Home represents Bancel’s most recent collaborative platform for distributed precognition research.¹ The system is Internet-based and allows participants to conduct experimental sessions at home using custom software. Each session comprises 20 forced-choice trials, and participants can contribute multiple sessions over time. The platform is designed to be accessible to outside researchers, expanding the potential for collaborative, multi-site psi research.

The development of Psi@Home reflects lessons learned from earlier platforms. The Sharefield, introduced in 2013, was designed as a novel approach for forced-choice GESP research, optimizing the presentation of stimuli and response collection.³ The Selfield, developed subsequently, further refined the methodology to optimize precognition research specifically.² Each iteration incorporated feedback from experimental results and methodological analysis, gradually improving the platform’s ability to detect and measure psi effects.

A key advantage of the Psi@Home system is its capacity for rapid data collection. In the pre-registered study Bancel conducted, the platform enabled the collection of 160 sessions from 47 participants within a single month.¹ This speed is crucial for addressing replicability concerns: large datasets can be analyzed for patterns and anomalies that might otherwise remain hidden in smaller, slower-accumulating studies.

Cohort selection and participant variables

A distinctive feature of Bancel’s recent work is the use of selected cohorts rather than unselected volunteer samples. In the Psi@Home study, he compared two groups: experienced meditators and participants selected merely for their interest in participating in the study.¹ This design allows for testing whether specific participant characteristics (such as meditation experience) correlate with psi performance or with particular patterns of variance.

The hypothesis that meditators might show different psi performance is grounded in theoretical reasoning: meditation practice may enhance attentional control, reduce psychological noise, or facilitate the mental states conducive to psi expression. The pre-registered study included three hypotheses: two confirmatory predictions (that both cohorts would show increased variance in session hit rates and increased total hit rate relative to mean chance expectation) and one exploratory prediction (that meditators would show higher variance than the general-interest cohort).¹

The results revealed a striking pattern: only the exploratory hypothesis (that meditators would show higher session variance) was confirmed (p = .03).¹ This finding suggests that meditation experience may be associated with more variable psi performance, a result consistent with the hypothesis that meditators exhibit a mixture of psi-hitting and psi-missing across sessions rather than a stable, consistent effect.

Variance anomalies and psi-hitting/missing

One of the most intriguing findings from Bancel’s recent work concerns variance in session-level performance. The pre-registered study itself did not show overall evidence for a psi effect in hit rates; however, the tryout sessions, 90 sessions collected under the same participants and protocol but specified in pre-registration as exploratory data, showed a markedly strong increase in session variance (p = .00003).¹ This dramatic difference between the formal study and the tryout data points to the importance of uncontrolled psychological variables in psi research.

The variance effect is theoretically significant because it is consistent with a mixture of psi-hitting and psi-missing across sessions. Rather than all sessions showing a consistent deviation from chance (either above or below), some sessions may show strong hitting while others show strong missing, resulting in high variance even if the overall mean is at chance. This pattern has been observed in other psi research and is sometimes interpreted as evidence that psi is present but expressed inconsistently, depending on participant motivation, attention, or other psychological factors.

Bancel’s interpretation of the discrepancy between tryout and formal study data emphasizes the role of participant attitudes and psychological state. During the tryout phase, participants may have been more relaxed, exploratory, or intrinsically motivated. Once the formal study began, with explicit hypotheses and pre-registration, participants may have experienced performance anxiety, self-consciousness, or other psychological shifts that suppressed the variance effect. This observation underscores a persistent challenge in psi research: the difficulty of maintaining consistent experimental conditions across the psychological and motivational dimensions that may be crucial to psi expression.

Replicability challenges and pre-registration

Bancel’s work is explicitly framed within the context of replicability challenges in parapsychology. He acknowledges that despite accumulated positive evidence from meta-analyses of forced-choice studies, individual experiments often fail to replicate, and even well-powered, presumably well-designed studies do not guarantee significant results.¹ This inconsistency has motivated a shift toward pre-registered, collaborative approaches that can gather larger datasets and test hypotheses prospectively rather than retrospectively.

Pre-registration, the practice of specifying hypotheses, methods, and analysis plans before data collection, is a key component of Bancel’s methodology. By distinguishing between confirmatory hypotheses (specified before data collection) and exploratory analyses (conducted on data collected under the same protocol but designated as exploratory), Bancel aims to reduce the risk of false positives and increase the credibility of findings. The Psi@Home study exemplifies this approach: the three hypotheses were pre-registered, and the distinction between the formal study (80 sessions per cohort) and the tryout data (90 sessions) was specified in advance.

However, the results also illustrate the complexity of pre-registration in psi research. The formal study, conducted under pre-registered confirmatory hypotheses, yielded only one significant result (the meditator variance effect). The tryout data, collected under the same protocol but designated as exploratory, showed a much stronger variance effect. This pattern raises questions about the relationship between participant expectations, experimental framing, and psi expression. It suggests that the psychological context of an experiment, whether it is framed as formal or exploratory, whether participants know their data will be analyzed for a specific hypothesis, may significantly influence psi performance.

Future directions and broader implications

Bancel’s development of collaborative platforms and his focus on variance anomalies point toward several promising directions for forced-choice precognition research. First, the availability of Psi@Home to outside researchers creates the potential for multi-site, collaborative studies that can accumulate large datasets and test hypotheses across different laboratories and participant populations. This collaborative approach mirrors successful meta-analytic efforts in other areas of parapsychology and may help address replicability concerns by enabling larger, more diverse samples.

Second, the emphasis on variance rather than mean hit rate opens new avenues for understanding psi expression. If psi is indeed expressed as a mixture of hitting and missing across sessions, then research should focus on identifying the conditions, participant characteristics, or psychological states that predict which sessions will show hitting versus missing. Bancel’s finding that meditators show higher variance suggests that meditation experience may be one such predictor, but other variables (motivation, attention, emotional state, or contextual factors) may also play important roles.

Third, the discrepancy between tryout and formal study data highlights the need for greater attention to the psychological and social dimensions of psi experiments. If participant attitudes and expectations significantly influence psi performance, then experimental design must account for these factors. This may require new approaches to participant recruitment, instruction, and feedback that minimize performance anxiety while maintaining experimental rigor.