Discovery Lab Answers

Marilyn J. Schlitz, PhD’s collaborative work with Richard Wiseman is one of the most-discussed series of experimenter-effect studies in parapsychology — a systematic, jointly designed program in which a proponent and a skeptic ran identical protocols and got consistently different results. Before reading the evidence, one important qualification: the library’s share of the published literature on this question has not been measured, so what follows is a summary of what the ESP-Nexus library holds rather than a settled account of the field.

Experiments

The collaboration produced three studies, all investigating whether people can psychically detect when they are being stared at by an unseen sender — the “remote detection of staring” paradigm — using electrodermal activity (EDA) as the physiological outcome measure.

1. Wiseman & Schlitz (1997) — the first joint study, published in the Journal of Parapsychology. Each author independently ran participants using the same protocol at their respective institutions. The paper notes that while preparing the journal version, the authors discovered and corrected a data-transfer error from an earlier conference presentation.

2. Wiseman & Schlitz (1999) — an attempted replication using the same paradigm, again with both experimenters running separate participant groups in parallel.

3. Schlitz, Wiseman, Watt, & Radin (2006) — a third, larger joint study explicitly designed as a collaboration between skeptic and proponent, adding Dean Radin and Caroline Watt as co-investigators, using a 2×2 cross-over design in which both Schlitz and Wiseman served as both greeter and sender across conditions. The planned sample was reduced from a larger target to 100 participants (25 per cell) in April 2004, prior to any examination of data.

The three studies together constitute what the sources describe as “a systematic programme of joint sceptic–proponent experimental work”. Caroline Watt subsequently conducted and published interviews with both researchers to explore potential sources of the divergent outcomes.

Methodology

All three studies used the same core protocol: a receiver sits alone in a room monitored by continuous EDA recording. A sender (or “starer”) is in a separate room and is cued by computer to either stare at a live video feed of the receiver or do nothing, in alternating periods. The receiver makes no conscious judgment — the EDA trace is the sole dependent variable. The critical methodological feature across all three studies is that Schlitz and Wiseman ran separate participant groups under conditions as similar as possible, including identical equipment, randomization, and analysis procedures, to test whether the experimenter’s identity affected outcomes.

The 1997 paper used an ABBA/BAAB randomization scheme to minimize progressive errors such as equipment warm-up or participant habituation. The computer displayed “stare” or “control” throughout each period; a “detect score” was computed for each participant by subtracting total EDA during stare periods from total EDA during non-stare periods.

A key methodological distinction Watt’s interviews surface is the preparation each experimenter engaged in before sessions: Wiseman reported doing none; Schlitz described a more involved mental preparation before running participants. This asymmetry is a primary candidate explanation for the divergence, though neither researcher fully endorsed any single account.

The 2006 study introduced the 2×2 cross-over design — each experimenter served as both greeter and sender, producing four distinct conditions — and added two further co-investigators to increase methodological rigor and control for any single-experimenter confound.

Data

The results across the three studies break down as follows:

Several features of this pattern require honest emphasis:

• Schlitz’s participants reached significance in both 1997 and 1999; Wiseman’s did not in either year. That is the headline experimenter-effect finding.
• The direction of Schlitz’s 1999 result is the opposite of her 1997 result. In 1997, EDA was higher during stare periods (the expected direction); in 1999, it was lower (Cohen’s h = −0.33, p = .05), which is a below-chance or “psi-missing” pattern — the structured evidence flags this explicitly as a directional disagreement across studies.
• The detect-score comparisons between Schlitz’s and Wiseman’s participants were not statistically significant in either 1997 (p = .17) or 1999 (p = .44), meaning the difference between their groups did not itself reach significance, even when the within-group results diverged.
• The 2006 study found no significant effect for either experimenter in any condition, with all four conditions producing near-zero effect sizes and the overall ANOVA main effect for greeter at p = .50.
• Eight of the thirteen result rows in the structured evidence for Schlitz’s full record are null or below-chance. The metric mix (Cohen’s h, Cohen’s d, z-only, and p-only across conditions) means these figures cannot be pooled into a single summary statistic.

For broader context on the DMILS/remote-staring literature, a separate meta-analytic result covering 30 EDA distant-intentionality experiments is reported in the Schlitz (1997) entry in the structured evidence, but that is a pooled result across the wider literature, not specific to the Wiseman–Schlitz joint studies.

Skeptical critiques

What critics argue. The primary challenge the Wiseman–Schlitz series raises internally — and which Wiseman himself has foregrounded — is that the divergence between experimenters is itself the finding requiring explanation, and the most parsimonious account may not require psi. Wiseman (1999) notes that proponents might argue the randomization sequence, initiated by the experimenters pre-session, could itself be influenced by experimenter psi — a position the paper characterizes as a possible but unfalsifiable account. The 2002 Watt interview and the 2005 published interview explore a set of more concrete candidate explanations: differences in experimenter preparation (Wiseman reported no pre-session mental preparation; Schlitz described deliberate focusing), differences in how experimenters interacted with participants during setup, and the possibility that subtle behavioral cues (“tacit information”) were transmitted to receivers in ways that affected their physiological responses. These are not fringe critiques; they appear in the researchers’ own jointly authored papers and in Watt’s structured interview format specifically designed to elicit them.

What the experimental data show. The 2006 study was explicitly designed to address some of these confounds by having both experimenters serve in all roles across a 2×2 cross-over design, adding two additional investigators, and reducing the opportunity for experimenter-specific behavioral transmission. That study found no significant effect for either experimenter in any condition, with the ANOVA main effect for greeter at p = .50. The detect-score comparisons in 1997 and 1999 also failed to reach significance (p = .17 and p = .44 respectively), meaning the experimenter difference, while present in raw within-group results, did not survive a direct between-groups test.

Analysis. The 1999 finding that Schlitz’s participants produced a statistically significant result in the opposite direction from 1997 — EDA lower during staring, not higher — complicates any straightforward psi interpretation of her positive results, and neither a psi account nor a tacit-information account has been published that fully reconciles the directional reversal with the 1997 finding. The direct between-groups tests failing to reach significance in both 1997 and 1999 means the experimenter effect is present as a descriptive pattern but not confirmed as a statistically significant interaction. Publication of the Watt interviews represents a collaborative methodological self-examination that is unusual in the literature — both researchers participated and neither claimed a definitive resolution.

For the full profile of Schlitz’s work, including her broader DMILS research and distant-healing studies, see her page at Marilyn J. Schlitz, PhD.