Daylia,

I have read the one Bunnell study in detail. First, a general word to studies. One can criticise single studies in retrospective, but that is a suboptimal procedure for manyx reasons: One may not find what was wrong, one critical detail may not have been mentioned, the description may be wrong and all that. The ultimate test is if the results described can be repeated by other researchers.

Two examples: The first empirical test of Einstein's theory did not come out in favour of the theory. Up to now, nobody knows why. But since this has not been replicated, the theory stands and the test is forgotten.

Backster once has published a bokk upon 'The power of prayer on plants'. Many of the experiments are perfect the way they are described. Nevertheless, nobody really believes Backster, because his results have not been replicated. With that in mind, read what I have to say about Bunnell:

No randomisation procedure is mentioned (the rough rule here is: what has not been mentioned has not been done or done wrong) in the detailed description.

The short abstract mentions double blind procedure, the detailed description only mentions one blinding.

The pilot study mentions t-test per trial (which did not make sense to me), the long study mentions just one t-test. The t-test her is not optimal a nonparametric test would be better, but (as the data look) would be significant as well.

The pilot study mentions one independent variable with three levels (healed, simply handled, warmed (untreated)), the longer study first says that the results from the first three trials of the pilot study were incorporated but then goes on only to mention two levels (healed/unhealed) for all trials. What has she done here? Lumped unhealed and untreated together? We'll never know.

I can't say for sure that the one or the other detail were responsible for the result, but her way of reporting and doing the study doesn't increase my trust in her competence doing this type of research.

Wolfgang