Meta-analysis of mind-matter interaction experiments: 1959 to 2000

Abstract

Laboratory experiments examining the possibility of direct mind-matter interactions have been reported for over a century. Two classes of such experiments reported most frequently include tossing dice while maintaining the intention for certain die faces to appear, and mental influence of truly random bits generated by electronic random number generators (RNG). Earlier metaanalyses of publications reporting dice and RNG experiments published up to 1987 provided strong statistical evidence for mind-matter interaction phenomena. We conducted an update of the RNG experiment literature to see if the evidence persists. The updated RNG review covered all known studies from the first published in 1959 to the most recent published in mid-2000. We found a total of 515 experiments published in 216 articles by 91 different first authors, of which 423 were published through 1987, and 92 published after 1987. The magnitude of the overall effect size per experiment is small, on average less than the equivalent of 1% for binary RNGs, but statistically the overall effect is more than 16 standard errors from chance. The average z score for studies published up to 1987 is z = .73 and for studies published after 1987 is z = .61. The difference in average z scores is not significant (p = 0.48), indicating that the meta-analytic evidence for mind-matter interaction effects persists. A conservative estimate of the effect of selective reporting practices (the “filedrawer problem”) indicates that to reduce the observed statistical outcome to chance, each of the 91 researchers would have had to conduct but not report 29 additional, nonsignificant experiments. Variations in methodological quality did not correlate with experimental results (r = 0.03, p = .26), but quality did significantly imp rove over time (r = 0.50, p = 10-34). We conclude that the RNG experiments continue to provide persuasive statistical evidence for independently repeatable mind-matter interaction effects observed under controlled conditions.