Meta-analysis that conceals more than it reveals: comment on Storm et al. (2010).

Storm, Tressoldi, and Di Risio (2010) rely on meta-analyses to justify their claim that the evidence for psi is consistent and reliable. They manufacture apparent homogeneity and consistency by eliminating many outliers and combining databases whose combined effect sizes are not significantly different-even though these combined effect sizes consist of arbitrary and meaningless composites. At best, their study provides a recipe for conducting a replicable extrasensory perception experiment. This recipe includes following a design that employs the standard ganzfeld psi methodology and uses "selected" subjects. An experiment, having adequate power and that meets these criteria, has already been conducted and failed to produce evidence for psi. Parapsychology will achieve scientific acceptability only when it provides a positive theory with evidence based on independently replicable evidence. This is something it has yet to achieve after more than a century of trying.