Nonclassical Experimenter Effects

Empirical Evidence

This effect has been known in parapsychological research for awhile. Several studies reported that parapsychological effects were found in some studies, whereas in other studies with the same experimental procedure, the effects were not shown. A [Page 908]well-known experiment that has shown such a nonclassical experimenter effect is one where a parapsychological researcher who had previously produced replicable results with a certain experimental setup invited a skeptical colleague into her laboratory to replicate the experiment with her. They ran the same experiment together; half of the subjects were introduced to the experimental procedures by the enthusiastic experimenter and half by the skeptical experimenter. The experimental task was to influence a participant's arousal remotely, measured by electrodermal activity, via intention only according to a random sequence. The two participants were separated from each other and housed in shielded chambers. Otherwise, all procedures were the same. Although the enthusiastic researcher could replicate the previous results, the skeptical researcher produced null results. This finding occurred even though there was no way of transferring the information in the experiment itself. This result was replicated in another study in the skeptical researcher's laboratory, where again the enthusiastic researcher could replicate the findings but the skeptic could not. There are also several studies reported where more than one experimenter interacted with the participants. If these studies are evaluated separately for each experimenter, it could be shown that some experimenters find consistently significant results whereas others do not. These are not only exploratory findings because some of these studies could be repeated and the experimenter effects were hypothesized.

Another experimental example are the so-called memory-of-water effects, where Jacques Benveniste, who was a French immunologist, had claimed that water mixed with an immunogenic substance and successively diluted in steps to a point where no original molecules were present would still have a measurable effect. Blinded experiments produced some results, sometimes replicable and sometimes not. Later, he claimed that such effects can also be digitized, recorded, and played back via a digital medium. A definitive investigation could show that these effects only happened when one particular experimenter was present who was known to be indebted to Benveniste and wanted the experiments to work. Although a large group of observers with specialists from different disciplines were present, there was no indication how the individual in question could have potentially biased this blinded system, although such tampering, and hence a classical experimenter effect, could not be excluded.

...