Detecting Psi-Missing in Forced-Choice Data

Peter A. Bancel

Institut Métapsychique International

Paris, France

Forced-choice experiments have long been a standard in psi research. Forced-choice psi tasks involve an intervention (choice) that yields an outcome that is deemed successful or not (hit or miss). Examples are various tests of PMIR, some precognitive experiments of Bem and others, online RV studies, etc. Experiments often give trial-by-trial feedback, and this may evoke unconscious participant responses, which lead to psi-missing. Psi-missing is characterized as outcomes where psi operates opposite the intended direction of the task. Psi-missing is generally recognized as a possible experimental outcome, but it is often considered a nuisance that can obscure the presence of psi when psi-hitting is the primary test of evidence.

Here, I describe ways to detect psi-missing when studies run sessions with multiple trials, and each trial is a forced-choice psi task. Because analyses usually look for psi-hitting, a mixture of hitting and missing can cause tests to lose power, even when strong psi is present. I present models of psi-missing, along with ways to detect it statistically. It’s found that the raw variance and trial autocorrelations can detect psi-missing in a wide range of scenarios. Of course, when psi-missing is absent, traditional mean tests are far more powerful.

The tests can be combined to attain a broader reach without too great a cost in statistical power. This is helpful for framing hypothesis tests in a parsimonious way. The model-based approach can also guide post-hoc analyses that aim to determine how psi-hitting and psi-missing are distributed in data. This can provide helpful input to psychological theories of psi performance.

—

Peter Bancel holds a PhD in experimental physics from the University of Pennsylvania. Following a post-doc at IBM Research, he moved to France to work with the Centre National de Recherche Scientifique in Paris and Nantes. His research interests have included quasicrystals, phase transitions in two-dimensional systems and protein crystal growth. Peter began working in psi research shortly after the US terror attacks of Sept. 11, 2001, when he joined a small group looking at data from the Global Consciousness Project. He has published work on state-induction approaches to improving psi effect sizes, theoretical assessments of meta-analyses in psi databases and experiments with meditator cohorts on precognition. Peter works at the Institut Métapsychique International in Paris, France.

—

Program chaired by Jacob W. Glazier. Download the Abstracts at https://parapsych.org/uploaded_files/pdfs/00/00/00/01/26/2023_pa_abstracts_of_presented_papers.pdf

Find us:

Facebook – https://www.facebook.com/ParapsychologicalAssociation

X – https://x.com/parapsych

Instagram – https://www.instagram.com/parapsychologicalassociation

LinkedIn – https://www.linkedin.com/company/parapsychological-association-inc-/

Donations to support the mission of the PA can be made by reaching office@parapsych.org or using this link: https://parapsych.org/section/11/support_the_pa.aspx

Published on February 7, 2025