(2019-09-29, 08:25 AM)Chris Wrote: There seem to be conflicting claims about whether these experiments with Tanous on perception in out-of-body experiences actually produced significant results. In the interview, and in his article on Tanous in the Psi Encyclopedia, Callum Cooper says they did, though he refers to a claim by Susan Blackmore that the statistical analysis was incorrect, which he says the authors rebutted:
https://psi-encyclopedia.spr.ac.uk/artic...d_Research
Blackmore herself, writing in 1983, seems to claim afterwards that the effect was only "marginal" (though she refers to a 1975 publication, whereas Cooper refers to one in 1980 and correspondence the following year):
https://www.susanblackmore.uk/chapters/o...-survival/
In a review article from 1982, Carlos S. Alvarado (at pp. 213, 214), referring to the same publication as Blackmore, says the overall results weren't significant, but some post hoc analyses produced results that were:
https://www.academia.edu/366168/_1982_._...46_209_230
To complicate matters, the film is said to have been made c. 1983, which is after both the publications by Osis. Cooper says there must be more unpublished data in the archives of the ASPR.
It's all a bit bewildering. Unfortunately I can't find either of the publications or the following criticism and response online.
Someone kindly sent me a copy of Susan Blackmore's letter to the JASPR, criticising the 1980 paper on the Tanous experiments in the same journal, and the reply by Osis and McCormick. This makes things clearer.
It wasn't so much that Blackmore was claiming the statistical analysis was incorrect, but that she thought it didn't make sense to look at the correlation between the correctness of the guesses in the visual task and another indicator (strain gauge measurements), because the overall hit rate for the visual task wasn't statistically significant, and therefore she questioned whether there was a psi effect at all.
The authors actually agreed that the overall hit rate wasn't significant, but pointed to a long history of experiments that had produced statistically significant patterns in data that were suggestive of psi, even when overall hit rates weren't significant. (I think they might also have pointed out that statistical significance depends on power, and in principle it's quite possible for the power of an experiment to be greater for a secondary effect than for the primary hit rate. The overall hit rate in these experiments was above chance expectation, after all.)
The authors had also pointed out (before Blackmore wrote her letter) that while the hit rate defined by getting at least one of the three aspects of the visual target correct was indeed not significant, the departure from chance was larger for the hit rate based on getting two or more aspects correct. Blackmore claimed this alternative hit rate was still insignificant (p=0.17), based on the data they had provided her with. I'm not actually sure how she calculated the statistic this is based on, but based on the data she prints, they got 36 hits out of 197 trials where the probability of success was 11/80, and an exact binomial calculation gives p=0.045.