Radin took me to task for not mentioning the control matrix presented in his paper. I did not mention it simply because it did not answer the key question (i.e., did the experimental matrix have a significant, non-artifactual number [N

_{s}] of correlations associated with p < .05?).

In essence, Radin conducted a 1-trial Monte Carlo simulation to assess the effect of dependence among elements in the experimental matrix. But a 1-trial simulation cannot adequately characterize the distribution of N

_{s}or determine how much the significance level was affected by the dependence artifact. Radin’s control value was lower than the experimental value, but one cannot therefore conclude that the experimental value was significantly greater than chance, let alone defend the claim of p = .0004 (or p = .000779).

Had Radin conducted a more usual simulation, i.e., with a large number of trials, perhaps he could have legitimately defended the reported p value, or at least the claim of significance. But he didn’t do that.

Radin’s response to me indicates that he still considers the 1-trial control simulation to be adequate. Readers with some familiarity with statistics are now in a better position to assess his methods.

References

Radin, Dean I. (1993). Environmental modulation and statistical equilibrium in mind-matter interaction. In Marilyn J. Schlitz (program chair), The Parapsychological Association 36th Annual Convention: Proceedings of Presented Papers (pages 157-176). The Parapsychological Association.

Radin, Dean. (August 14, 2007). Trickster, or failure of imagination? [blog post]. Available at: http://deanradin.blogspot.com/2007/08/trickster-or-failure-of-imagination.html. Accessed September 14, 2007.

## No comments:

Post a Comment