Discovery Lab Answers

John Palmer, PhD is a parapsychologist who has served as a long-standing editor of the Journal of Parapsychology at the Rhine Research Center. His research record spans several decades and covers experimental ESP methodology, psychological and physiological correlates of anomalous cognition, meta-analytic synthesis, and the history of the field. His profile page provides the fullest site synthesis.

Experiments

Palmer’s experimental work falls into several distinct lines:

Belief and psychological correlates of ESP. His early work examined how participants’ prior belief in ESP — the “sheep-goat effect” — predicted scoring patterns in experimental tasks. Related work investigated personality traits, emotional states, and defensive mechanisms as influences on ESP performance ^[2].

Memory and ESP. A substantial review paper ^[4] examined the experimental literature linking ESP performance to memory processes. Palmer reviewed studies testing whether ESP hits correlate with memory-task performance, synthesizing work by multiple researchers including Stanford and himself (Stanford & Palmer, 1972). This included card-guessing paradigms where participants memorized ESP symbols before guessing tasks, and free-response designs examining response-bias hypotheses.

Neuropsychiatric correlates of subjective ESP experiences. With Vernon M. Neppe, Palmer conducted exploratory analyses in a neuropsychiatric population — patients at a neurology clinic — examining predictors of subjective ESP (S-ESP) experiences and temporal lobe dysfunction ^[5].

Meta-analytic and database work. From the 1990s onward, Palmer shifted substantially toward synthesizing large bodies of experimental data, including an updated meta-analysis of post-Parapsychology Laboratory ESP experiments examining methodological standardness and effect-size variability.

Historical and editorial roles. Palmer authored an 80-year historical overview of the Journal of Parapsychology ^[2] and wrote a personal tribute to Rex Stanford that documents methodological observations from decades of shared work ^[1]. He also co-edited a major handbook of parapsychology (Parapsychology: A Handbook for the 21st Century, with Etzel Cardeña and David Marcusson-Clavertz).

Palmer received his first position in parapsychology in 1971, hired by Ian Stevenson as a Research Assistant ^[3].

Methodology

Sheep-goat and correlate designs. Palmer’s early experimental work used self-report measures of ESP belief and personality to predict ESP task performance, testing within-sample correlations between psychological variables and hit rates.

Memory–ESP paradigm. The review ^[4] describes paradigms in which participants memorized randomized lists of ESP symbols (typically under time pressure — as little as 15 seconds) immediately before a guessing task. Scoring of test sheets was blind. The designs tested whether memory scores and ESP scores correlated positively, and whether response-bias toward primary (high-association) words explained any ESP effects. Palmer noted limitations in some of these designs: effects distributed across a sample may not be uniform across individual participants, and some analyses used trials rather than participants as the unit, limiting generalization ^[4].

Neuropsychiatric study design. The Palmer and Neppe (2004) study ^[5] used logistic regression on a neuropsychiatric inpatient/outpatient population. Variables were drawn from the INSET instrument (Neppe’s structured neuropsychiatric interview). The analyses were explicitly exploratory — the paper’s own title signals this. Variable selection proceeded in stages, with final models retaining predictors that survived iterative filtering. Gender was included as a covariate throughout, and the TLD (temporal lobe dysfunction) hypothesis was tested with and without gender adjustment ^[5].

Meta-analytic work. Palmer’s database analyses examined whether methodological rigor (operationalized as “standardness”) predicted effect-size variability — a design that treats the study, not the individual participant, as the unit of analysis.

Statistical methods. In the memory–ESP review, Palmer used the Stouffer method to combine z-statistics across series ^[4], and was attentive to whether component Zs should be weighted by sample size.

Data

The retrieved excerpts contain several specific findings. Because the evidence block for this question carries no structured rows, the figures below come directly from the primary sources and are reported only where the surrounding text is intact enough that attribution is unambiguous.

Memory–ESP correlates ^[4]

Palmer notes that the Stouffer Z for the RC/RW effect would have been stronger had analysis been limited to low-association word pairs and components weighted by sample size ^[4].

Neuropsychiatric S-ESP correlates ^[5]

The TLD hypothesis — that temporal lobe dysfunction predicts S-ESP — reached significance without gender adjustment but was no longer supported once gender was included as a covariate ^[5]. Palmer and Neppe flag this explicitly. The final logistic regression model retained three predictors: gender, hallucinations, and jamais vu ^[5].

Historical and editorial observations ^[2]: Palmer’s 80-year overview documents published controversies in the Journal of Parapsychology but does not report primary experimental statistics.

Stanford tribute ^[1]: Palmer describes Stanford’s detection of methodological flaws in psi experiments — including a review of 25 ESP studies comparing hypnosis conditions — but the retrieved excerpts do not contain primary statistics from those studies.

Across Palmer’s experimental and correlate work, the pattern includes positive findings on memory–ESP relationships, significant neuropsychiatric predictors of subjective ESP, and a TLD effect that did not survive covariate adjustment. His meta-analytic work identified relationships between methodological standardness and effect-size variability, though the specific statistics from that line are not in the retrieved excerpts.

Skeptical critiques

What critics argue. Palmer himself surfaces the key methodological concern in the memory–ESP literature: where trials rather than participants serve as the unit of statistical analysis, the effect “cannot properly be generalized to a population of persons because the effect in the sample may not have been distributed uniformly across the sample participants” ^[4]. This is a participant-clustering objection — significant results may reflect concentrated hitting in a few individuals rather than a generalizable effect.

What the experimental data show. Palmer acknowledges this limitation directly for at least one series, noting that the obtained effect was strong enough that “it is unlikely an analysis with the participant as the unit would lead to a different conclusion” ^[4]. For the neuropsychiatric work, the TLD hypothesis itself failed to survive gender adjustment in an analysis of variance ^[5] — a null outcome Palmer and Neppe report without concealment.

Analysis. The critiques surfaced in these sources are primarily methodological cautions raised by Palmer and his collaborators in their own papers, not external published attacks on specific findings. The TLD null-after-covariate result ^[5] represents an internally documented failure of a pre-stated hypothesis. Published external critiques of Palmer’s specific experimental findings are not present in the sources retrieved for this question.