Discovery Lab Answers

The Schlitz–Wiseman Collaboration

Coverage note: The statistics below reflect the studies currently in the ESP-Nexus library. The library’s share of the published literature on this question has not been measured, so treat this as a summary of what the library holds rather than a settled account of the field. Given that most results in the library for this question are null, it is worth holding open the possibility that the picture across the full literature differs.

Marilyn Schlitz and Richard Wiseman conducted a sustained, formally structured joint research program — one of the few cases in parapsychology where a proponent and a skeptic ran identical protocols side by side and published the results together. The central phenomenon they studied was remote staring: whether a person’s electrodermal activity (EDA) rises when an unseen sender is looking at them. Their collaboration spanned three published studies (1997, 1999, 2006) and became the most cited empirical case of the experimenter effect in the field.

The experimental design

The core logic was deliberately adversarial-cooperative. Both researchers used the same equipment, the same lab space (shared between the University of Hertfordshire and the Institute of Noetic Sciences), the same randomization protocol, and the same EDA outcome measure — but each ran their own participants independently. If results differed by experimenter, the protocol itself could not explain the divergence.

What the three studies found

\* The 1999 Schlitz result reached p = .05, but the effect ran in the opposite direction to 1997 — EDA was lower during staring epochs rather than higher, a pattern the paper notes explicitly.

The pattern across the three studies

The 1997 study showed the clearest experimenter contrast: Schlitz’s participants produced a significant result; Wiseman’s did not. The 1999 replication was more complicated — Schlitz’s arm again reached the conventional significance threshold, but with a reversed effect direction, while Wiseman’s arm was again null; and the between-experimenter comparison was not significant in either study. The 2006 study, the most elaborate (a 2×2 crossover design adding Dean Radin as a second proponent experimenter and varying who greeted participants versus who sent), found no significant effect for any condition or experimenter.

Why the design attracted so much attention

The protocol controlled for the most obvious confounds: equipment, procedure, and randomization were identical across experimenters. What varied was the person running the session — their expectations, their pre-session demeanor, and, as explored in a published interview, their behavior toward participants before the session began. Schlitz described routinely spending preparatory time with participants and maintaining an intention to influence them during staring epochs; Wiseman described making no such preparation. Whether this behavioral difference explains the 1997 and 1999 divergence, or whether something else does, was not resolved by the data alone.

The broader meta-analytic context

A separate meta-analysis of thirty DMILS and remote-staring experiments using EDA as the outcome measure, reported in Schlitz (1997), found a positive overall effect across multiple independent laboratories [see the library’s evidence table]. That pooled result sits alongside the null and mixed outcomes from the joint studies, and those are different metrics from different study sets — they cannot be directly combined.

Skeptical critiques

What critics argue. Richard Wiseman’s own 2010 paper (discussed in Baptista & Derakhshani, 2014) raises the possibility that experimenter effects in parapsychology, including those found in the joint staring studies, may reflect subtle uncontrolled differences in procedure rather than any psi-mediated influence. The implicit argument is that if the protocol were truly identical, results should not diverge by experimenter.

What the experimental data show. The between-experimenter comparisons in 1997 and 1999 did not reach significance, meaning the data did not formally establish that Schlitz’s and Wiseman’s participants differed from each other — only that one arm reached significance and the other did not. The 2006 study, designed partly to address procedural confounds through the crossover design, found no significant effect for either experimenter. Baptista and Derakhshani (2014) note that Wiseman was himself party to the 1997 study and that the results were consistent with experimenter-effect predictions made in advance.

Analysis. The published interview transcripts document concrete behavioral differences between the two experimenters in participant preparation. Whether those behavioral differences constitute a methodological confound (a procedural artifact) or a genuine experimenter effect (an anomalous influence) is precisely what the three studies were designed to probe — and what they left open. No independent laboratory has published a preregistered replication of the three-condition Schlitz–Wiseman design using a different experimenter pair.

For Schlitz’s broader research profile, see her page at Marilyn J. Schlitz.