Back

Distant Intentionality? A Critical Empirical Assessment

Principal Investigator:
S. Schmidt, PhD

Affiliations: University Hospital Freiburg, Department of Environmental Medicine and Hospital Epidemiology

Background: In a meta-analysis on so called Direct Mental Interaction with Living Systems (DMILS), a highly significant effect of medium size from a set of 19 studies was found (ES(r) = .25, p = .0000007). This finding indicates that two physically isolated persons may interact without using conventional means of communication, i.e. one person seems to be able to mentally �influence� the physiological arousal of another person (EDA, electrodermal activity).

Objective of the study: The aim of this project was to assess this claim with different empirical strategies.

Methods and Results: Part I consisted of an assessment of the applied EDA methodology in EDA-DMILS studies conducted until 1998. The comparison with standard psychophysiological methodology showed there are major methodological shortcomings in DMILS studies. Part II consisted of a pilot study with improved methodology. The aim of this experiment was mainly to evaluate several EDA parameterization procedures and different statistical approaches. The study consisted of 26 session and yielded a medium size effect size of ES(r) = .35 � .42 depending on the EDA parameter. Part III was a replication study with the same methodology as the pilot study. In an experiment with 48 sessions, 96 subjects and 4 experimenters, no significant effect was found. Effect sizes were ES(r) = �.06 and �.07. The effects of the pilot study could not be replicated. Part IV consisted of a meta-analysis of all EDA-DMILS studies completed by the end of 2000 to assess mean effect size, level of significance and the relationship between various indices of study quality and effect size. Overall, 40 experiments with 1055 sessions were identified and coded for study characteristics on 208 different variables. Four different quality indices were constructed from which study quality turned out to be one important moderator variable. After the exclusion of four flawed studies and after weighting the studies by quality and sample size, a homogenous data set was obtained. The mean effect size of this data set was ES(d) = .11 (p = .001).