Discovery Lab Answers

Marilyn J. Schlitz, PhD — Effect Sizes

Before drawing any conclusions, note that the ESP-Nexus library’s share of the published literature on Schlitz’s work has not been measured. What follows summarizes the studies currently held in the library — treat it as a sample, not a complete account of the field.

The evidence base covers several distinct research lines — ESP performance, distant mental influence on living systems (DMILS), distant healing intention, and precognition — and the metrics across these lines are not comparable. The structured evidence contains a mix of Cohen’s h, Cohen’s d, Pearson r, and z-only results. These cannot be averaged into a single bottom-line figure.

Experiments

ESP with arts students (1992). Schlitz tested 20 Juilliard-affiliated receivers in a forced-choice ESP task. The result for this group was statistically significant.

DMILS/remote-staring meta-analysis (1997). Schlitz, with Braud, pooled 30 electrodermal activity (EDA) distant-intentionality experiments across multiple independent laboratories.

Schlitz–Wiseman joint replication series. Three studies conducted jointly with Richard Wiseman using identical protocols, equipment, and participant pools were designed explicitly to test whether the experimenter’s orientation — proponent vs. skeptic — affected outcomes. Two rounds were published (1997, 1999) and a crossover design was added in 2006.

Distant healing intention — wound healing (2012). Schlitz and colleagues examined DHI effects on objective and psychosocial measures in surgical patients, including collagen deposition as a primary outcome.

Time-reversed priming replication (2021). Schlitz co-led a large multi-site replication of Bem’s precognitive priming paradigm, enrolling 493 participants.

Distant healing intention overview (2015). Radin, Schlitz, and Baur synthesized the broader DMILS and DHI literature across multiple meta-analyses.

Methodology

DMILS/remote-staring. The standard protocol pairs a “sender” in one room who randomly alternates between staring at a video feed of the receiver and not staring; the receiver’s EDA is recorded continuously and blind to staring epochs. Randomization of staring/non-staring epochs and blinding of the receiver are the core controls.

Schlitz–Wiseman crossover design. The 2006 study added a 2×2 crossover in which each experimenter served as both greeter and sender in separate conditions, allowing within-study separation of the greeter effect from the sender effect. This is a methodological refinement directly responsive to the earlier critiques about which stage of the procedure the experimenter influenced.

ESP with arts students. Conducted at the Psychophysical Research Laboratories, using the standard PRL forced-choice protocol with music students as a theoretically motivated subgroup.

Time-reversed priming. A multi-site, multi-experimenter replication of Bem (2011), pre-registering confirmatory analyses separately from exploratory ones. The confirmatory analysis used log-transformed reaction times at a 2,500 ms window.

Wound-healing DHI. A three-group blinded RCT design with collagen deposition as the primary biomarker outcome.

Data

Key patterns across the evidence:

The 1992 Juilliard ESP study returned the largest single-study effect in the library (Cohen’s h = 0.52), though on only 20 trials — a very small sample for effect-size estimation.

The 1997 DMILS meta-analysis (Pearson r = 0.25, pooling 30 studies) is the widest-scope positive result, but the metric is not comparable with the per-experiment Cohen’s h figures from the staring studies — do not treat them as the same quantity.

The experimenter-effect pattern across the Wiseman–Schlitz series is the most discussed feature of this dataset. In 1997, Schlitz-run receivers produced z = −2.02 (p =.04, positive direction); Wiseman-run receivers produced z = −0.44 (p =.64, null). In 1999, Schlitz-run participants reached Cohen’s h = −0.33 (p =.05, borderline positive); Wiseman-run participants were at Cohen’s h = −0.07 (p =.69, null). Yet in both years, the direct comparison between the two experimenters’ participants did not itself reach significance (p =.17 in 1997; p =.44 in 1999).

The 2006 crossover — designed to isolate whether the effect arose during greeting or during sending — produced null results in all four conditions tested, with effect sizes near zero (Cohen’s h ranging from −0.03 to −0.07).

The 2021 precognitive priming replication and the 2012 wound-healing study both returned null results on their confirmatory primary outcomes.

Eight of the 13 result rows are null or below-chance in direction. The positive results cluster in the earlier work (1992, 1997) and in the Schlitz-run conditions of the joint experiments; the later dedicated replications and the wound-healing trial are uniformly null.

Skeptical critiques

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

What the experimental data show. The 2006 crossover study was designed precisely to address this: by having each experimenter serve as both greeter and sender in separate conditions, it could in principle separate greeting artifacts from sender-phase effects. All four conditions in that design returned null results (Cohen’s h between −0.03 and −0.07, p values of.72 and.87), and the ANOVA main effect of greeter was non-significant (p =.50). The 1997 meta-analysis (Pearson r = 0.25, z = 6.17, p = 4.58 × 10⁻¹⁰) pooled results across multiple laboratories, offering a broader base; however, Wagenmakers et al. (2011), writing in the context of Bem’s related precognitive priming work, argued that analyses of this type can be partly exploratory and low-powered in ways that inflate apparent effect sizes and produce false positives.

Analysis. The within-series trajectory is notable: the 1992 and 1997 results are positive; the 1999 borderline result did not survive formal comparison; the 2006 crossover — the most methodologically refined study in the series — produced nothing above chance on any condition; and the 2021 multi-site precognitive priming replication with 493 participants returned a confirmatory effect of Cohen’s d = 0.03 (p =.22). Whether the earlier positive results reflect a real phenomenon, experimenter artifacts, or sampling variability in small studies remains unresolved in the published record. Independent replication of the staring effect outside the Schlitz/Wiseman collaborative framework, using comparable methodology, has not resolved the question.

For further background on Schlitz’s full research record, including her work at IONS and the broader DMILS literature, see her profile at Marilyn J. Schlitz, PhD.