Discovery Lab Answers

The Schlitz–Wiseman remote-staring collaboration

Marilyn J. Schlitz, PhD conducted a series of joint remote-staring experiments with the skeptic Richard Wiseman — one of the few sustained proponent–skeptic collaborations in parapsychology’s history. In these studies both investigators ran participants under the same protocol, equipment, and participant pool, testing whether an unconscious autonomic response to being covertly observed (measured by electrodermal activity, EDA) would appear for one experimenter but not the other.

Coverage: this summary reflects the studies currently in the ESP-Nexus library, not the complete published literature on remote staring or experimenter effects. How much of that literature the library holds has not been measured, so treat this as a summary of what we hold rather than a settled account of the field.

Experiments

The three joint experiments the sources document are:

• Wiseman & Schlitz (1997) — the first joint study. Each author acted as both experimenter and sender, running their own receivers on an EDA stare/non-stare paradigm.
• Wiseman & Schlitz (1999) — an attempted replication of the 1997 study, run at the University of Hertfordshire, again with each investigator running participants under the shared protocol.
• Schlitz, Wiseman, Watt & Radin (2006) — a third collaboration (“Of two minds”), which added Caroline Watt and Dean Radin and attempted to replicate the earlier experimenter-effect pattern with additional controls, including a heavily acoustically shielded chamber and a test of whether the effect depended on Schlitz acting as greeter versus sender.

A separate, related line of work by Caroline Watt & Peter Ramakers (2003) extended the paradigm to multiple believer and disbeliever experimenters — a “four-experimenter” style follow-up examining whether an experimenter-linked pattern generalizes beyond the two original investigators.

Methodology

The core design across all three joint studies was consistent: a receiver in one room had EDA recorded while, in a separate location, the experimenter/sender either stared at a live video image of the receiver or looked away, with stare and non-stare epochs randomized. The rooms were separated to block sensory leakage — in 1999 a video camera relayed the receiver’s image via a long cable to the observation room, and in 2006 the receiver sat in a steel-walled shielded chamber with roughly 100 dB of acoustic attenuation and no line of sight to the experimenter.

The distinctive methodological feature is that the same protocol, equipment, and participant pool were used by both a proponent (Schlitz) and a skeptic (Wiseman) simultaneously, so that a systematic difference between their results could not be attributed to differing procedures. The 2006 study went further by decomposing Schlitz’s role — testing whether outcomes depended on how she interacted with participants at the start of a session (greeter versus sender) and correlating session outcomes with greeter–participant rapport.

Data

The library’s structured evidence base carries the following results from the joint studies. Note the metric mix — the 1997 rows report a z score only, while the 1999 and 2006 rows report a standardized effect size (Cohen’s h) with a z. These are not directly poolable, and no single “overall effect size” should be read across them. The pattern is the important thing: in each of the first two studies, participants run by Schlitz reached (or approached) significance in the predicted direction, while participants run by Wiseman did not; in the third study both of Schlitz’s own conditions came out null.

Reading the arc: the first two studies show the candidate experimenter-effect pattern — Schlitz’s participants significant, Wiseman’s not. The third study did not reproduce it; both of Schlitz’s conditions were null, and the study found no influence of whether Schlitz acted as greeter or sender and non-significant correlations between session outcome and rapport. Across the six rows in the evidence set, most results are null — the positive results are confined to Schlitz’s runs in 1997 and 1999. One important caution the site’s own synthesis records: in each of the first two studies, the difference between the two experimenters was not itself statistically significant, even where Schlitz’s own condition was.

The z-versus-year scatterplot rendered below the answer plots these six results; note that z is a test statistic that scales with sample size, so it tracks significance over time rather than effect magnitude — read the caption for the fitted trend’s specification.

Skeptical critiques

What critics argue. Wiseman and Schlitz (1997) themselves record the central interpretive problem: because the difference between the two experimenters did not reach significance in either of the first two studies, one cannot cleanly conclude an experimenter effect drove the divergence rather than chance. Their 2006 paper (Schlitz, Wiseman, Watt & Radin) reports the direct test of that concern — a pre-registered replication that failed to reproduce the earlier pattern, with both of Schlitz’s conditions null. A broader methodological framing comes from Watt and Ramakers (2003), whose multi-experimenter design was built precisely because a two-person proponent–skeptic contrast cannot isolate which experimenter characteristics (belief, rapport, behavior) might matter.

What the experimental data show. The numbers bear out the ambiguity. Schlitz’s participants produced significant stare/non-stare EDA differences in 1997 (z = −2.02, p = .04) and marginally in 1999 (z = −1.93, p = .05, h = −0.33), while Wiseman’s were null in both (p = .64; p = .69). But the 2006 replication, with tighter shielding and a decomposition of Schlitz’s role, yielded null results for Schlitz herself (p = .87; p = .72) and no rapport or greeter/sender dependence. The Watt and Ramakers (2003) multi-experimenter extension examined believer versus disbeliever experimenters as a way to test whether such a divide generalizes.

Analysis. Where the same protocol was run by a proponent and a skeptic, Schlitz obtained significant results and Wiseman did not in 1997 and 1999; when the collaboration attempted a third replication in 2006 with added controls, Schlitz’s own participants were null and the earlier pattern did not reappear. Wiseman and Schlitz have described the exchange as much a study of the sociology of science as of psi itself. Independent characterization of the effect as an experimenter-linked moderator would require the kind of large, multi-laboratory dataset that Watt and Ramakers (2003) began to build with additional experimenters; what the retrieved sources establish is two directional successes under Schlitz followed by a directional null, with the between-experimenter difference never itself significant in the first two studies.

For the site’s fuller treatment, see Remote staring detection and autonomic nervous system activation and Experimenter effects in parapsychology research.