Meta-analysis of free-response ESP studies without altered states of consciousness

Seventy-eight free-response ESP studies in which participants were not in an altered state of consciousness were meta-analyzed. There was a highly significant cumulative effect (Stouffer z = 5.72, p < 5.4 x 10(-9) one-tailed). The mean effect size (z/N-1/2) was 0.16 (SD = 0.29). Crude selective reporting appears to be an implausible counterexplanation of the overall cumulation and there were no strong positive indications that methodological artifacts played a role in producing above-chance results, However, the failure to report that the outcome measure was preplanned in 96% of the studies and the frequent use of multiple measures raises the possibility of widespread post hoc data selection problems within the studies themselves. No data are available to resolve the question of whether this threat to the database's validity is merely potential. Because of this problem and the obstacles to obtaining statistical evidence of the effects of flaws in any meta-analysis, caution is recommended in drawing strong conclusions from this meta-analysis about the existence of a genuine anomaly. However, a number of possible moderator variables were identified that suggest directions for future research.