Abstract

Describes a critical evaluation of 42 ganzfeld psi studies from 34 reports written or published from 1974 through 1981. Allegedly, 55% of these studies achieved significance on the primary index of psi. Taking into account ambiguities and inconsistencies in what is counted as an independent ganzfeld study, and citing evidence suggestive of a bias in reporting the studies, it is argued that the actual rate of success was at most 30%. The author calculates that, because of multiple testing, the true significance level was much higher than the assumed .05 level, perhaps .25 or higher. The author also reports a number of procedural flaws involving inadequate randomization, potentials for sensory leakage, and statistical errors. A meta-analysis based on indices of significance and effect size as they relate to the various categories of flaws is presented. The flaws of inadequate security, possible sensory leakage, and multiple testing did not correlate with significance and effect size. Flaws involving inadequate randomization and insufficient documentation, however, did correlate with these indices. It is noted that both effect size and Z scores become approximately zero when regression equations are used to predict their values for the case in which these latter types of flaws are zero. It is concluded that this data base is too weak to support any assertions about the existence of psi. A list of the studies examined is appended.