Discovery Lab Answers

Jessica M. Utts, PhD

Jessica M. Utts, PhD is a statistician and Professor Emerita at the University of California, Irvine, whose contributions to parapsychology are methodological rather than experimental in the traditional sense — she brought the tools of statistical inference, meta-analysis, and power analysis to bear on a field whose evidentiary debates had often been conducted without them. She served as President of the American Statistical Association in 2016, and her core argument across three decades has been consistent: parapsychology should be held to the same statistical standards as any other science, no more and no less.

A note on coverage: the studies in the ESP-Nexus library associated with Utts’s work represent what the library currently holds, not the complete published literature on this question. The library’s share of that literature has not been measured, so the account below is a summary of what the library holds rather than a settled picture of the full field.

Experiments

Utts has not primarily been an experimenter — her role has been as a statistical analyst, meta-analyst, and methodological evaluator of other researchers’ protocols.

Her most cited single contribution is the 1995 government-commissioned report evaluating two decades of remote viewing research conducted at Stanford Research Institute (SRI International) and Science Applications International Corporation (SAIC). That report, republished in the Journal of Parapsychology, assessed whether the evidence for anomalous cognition met conventional scientific standards. The same AIR review process paired her assessment with an independent skeptical evaluation by Ray Hyman.

She co-authored two meta-analyses on predictive anticipatory activity (PAA) — the phenomenon of physiological responses appearing to precede randomly selected stimuli — with Julia A. Mossbridge and Patrizio Tressoldi. The 2012 paper in Frontiers in Psychology synthesized 26 studies across seven laboratories. A 2014 follow-up in Frontiers in Human Neuroscience, adding Utts, John A. Ives, Dean Radin, and Wayne B. Jonas as co-authors, extended the critical analysis and examined practical implications.

Her 1991 paper in Statistical Science applied meta-analytic methods to the autoganzfeld database and the broader parapsychological literature, evaluating replication claims through the lens of statistical power. She also contributed a statistical evaluation of the ganzfeld and remote viewing literature in a 1999 analysis covering multiple laboratories.

A separate line of work involves Bayesian approaches to evaluating ESP studies, including a co-authored response with Bem and Johnson (2011) to the Wagenmakers et al. reanalysis of Bem’s precognition data.

Her 2022 reflection in the Zeitschrift für Anomalistik describes how she arrived at parapsychology research in 1984 through the Society for Scientific Exploration at Stanford, after looking for a domain where she could apply statistical expertise.

Methodology

Utts’s central methodological contribution is the argument that the field’s replication controversy is largely an artifact of low statistical power. She argued that defining successful replication as achieving p ≤.05 is statistically incoherent when individual studies are underpowered: a null result from such a study cannot distinguish a true null from a small real effect the study lacked power to detect. Her recommended reframing was to evaluate replication through effect size estimation and meta-analytic synthesis rather than through binary significance thresholds.

In her evaluations of the SRI/SAIC remote viewing program, she applied this framework to a large accumulated trial base, assessing whether the program’s results were consistent with a small but stable anomalous signal rather than chance fluctuation.

For the PAA meta-analyses, the protocol required physiological measures (typically skin conductance, heart rate, or blood volume pulse) recorded during anticipation windows before randomly assigned emotional or neutral stimuli. The 2012 meta-analysis excluded five laboratories whose data were related to the question but did not meet inclusion criteria, and calculated a fail-safe number to assess publication bias robustness. The 2014 paper extended the analysis and addressed alternative explanations including expectation bias.

Utts has also argued for Bayesian approaches to evidence evaluation in parapsychology, contending that posterior probabilities — which incorporate prior beliefs — explain much of the persistent disagreement between researchers who examine the same data and reach different conclusions.

Data

The table below reports the audited figures from the ESP-Nexus library for studies in Utts’s record. Because these rows use different, non-comparable metrics (standardized effect sizes, raw hit rates, and a p-value only), they cannot be averaged or pooled into a single overall figure — each is described on its own terms.

On the SRI row: the figures attributed to May, Utts, Trask, Luke, Frivold, and Humphrey (1988/1989) are those of the original SRI International Technical Report, relayed and re-derived within Utts (1996/1999). They represent the SRI program’s own overall analysis for 1973–1988 and should be read as that program’s reported outcome, not as an independent finding by Utts.

All six rows in the library point in a positive direction. The library holds no null or below-chance result rows for Utts-associated work, though this reflects current library holdings rather than a complete survey of her published record — null results and non-significant outcomes may exist in the broader literature not yet held here.

Skeptical critiques

What critics argue. No published critique of this work appears in the sources retrieved for this question.

On the PAA meta-analyses, the 2014 paper itself addresses the alternative explanation of expectation bias — the possibility that anticipatory physiological responses reflect learned expectations about stimulus sequences (analogous to the gambler’s fallacy) rather than genuine precognition. The authors engaged this alternative directly in the paper’s critical analysis section.

On the Bem (2011) precognition data, the Wagenmakers et al. reanalysis used Bayesian methods to argue that the evidence was weaker than the frequentist p-values suggested. Utts, Bem, and Johnson (2011) published a direct response in the same journal, arguing that Wagenmakers et al. had applied priors that encoded strong prior skepticism, and that the Bayesian framework, properly applied, does not automatically favor a null conclusion.

What the experimental data show. Utts (1999) reported a hit rate of 0.34 across 2,097 combined ganzfeld and remote viewing sessions, against a chance expectation of 0.25. The Mossbridge, Tressoldi, and Utts (2012) meta-analysis reported a pooled Cohen’s d of 0.21 across 26 studies, with a fail-safe number of 87 contrary unpublished reports required to reduce significance to chance. Schmidt (2004) reported a smaller pooled effect of Cohen’s d = 0.11 across 36 DMILS studies.

Analysis. The Utts–Hyman paired evaluation of 1995 is the most documented instance of two credentialed reviewers examining the same evidence base and reaching different conclusions. Hyman’s objection centered on artifact control; Utts’s position was that the controls met the standards applied in other sciences. That specific disagreement — over what level of methodological documentation is sufficient — has not been resolved by a subsequent joint publication from both parties. Independent replication of the SRI/SAIC program outside the original laboratories, evaluated under the same conditions, has not been documented in the sources retrieved for this question.

For Utts’s full profile and her spoke pages on statistical literacy, replication standards, and meta-analysis, see Jessica M. Utts on ESP-Nexus.