It will be interesting to see how this is received among psychologists more generally. I agree that parapsychology research would benefit from a broader input from other fields, although I hope that they do not take him up on the proposal to repeat uncontrolled experiments, but rather consider improving the design.
However, even if non-artefactual, these experiments don't seem to have anything to do with what we think of as 'psi'. I realize that psi has now been defined in such a way as to incorporate these results, but it borders on bizarre to claim that the potentially tiny effect of a fuzziness in our perceptions on the order of milliseconds can be presumed to relate to claims of presentiment. It would make more sense to relate these experiments to those of Libet, I think.
Much about the presentation of this selection of experiments is the same as other experiments, so I admit to some tedium on reading through it. But there were several things I found interesting, so far. Bem (finally
) raises the issue of the actual placement of targets instead of a theoretical randomness when it comes to analysis (as Ersby referred to above). He didn't go so far as to make this a preference in terms of analysis, apply this analysis to all the experiments, or to design experiments which would obviate this effect, but at least he recognized the issue. And one of his analyses even provided a reasonable accommodation for this effect. It did modify the effect by a small amount.
The other part I found interesting was the discussion section titled "The File Drawer". It is interesting to get a glimpse of the much larger collection of data from which this selection was culled, although it does give some hints as to the extent to which statistical assumptions are violated.
However, for much of the experiments, Bem demonstrates that taking many measures and then dividing people into many different groups on the basis of those measures allows you to sometimes find 'statistically significant' differences in between group measures. He 'tests' some alternate explanations for those differences with varying success and leaves it at that. Residual 'unexplained' differences are meant to serve as confirmation of 'anomalous cognition', rather than simply a measure of the limits of his cleverness or of any interest by critics (i.e. the extent to which switching the burden of type I and type II errors helps us arrive at true results). As Ersby mentions, one wonders at the lack of control groups.
Linda
