Experimental Replicability

Abstract

Editorials in the last two issues  of the Journal have discussed flaws inherent in the notion that observations or measurements must be reproducible or replicable if they are to be accepted as properly scientific. The arguments and examples are compelling, and I would like to add more emphasis to one of the points that the Editorials touch on implicitly more than explicitly. That point is that human beings differ from one another in innumerable ways, as well as to varying degrees in those capabilities that they may have in common; even more important, human beings are not exactly the same from moment to moment — they are inherently unreproducible.

      To demand reproducibility in observations or measurements presumes that what is being studied is itself so stable or regular as to be able to reproduce itself on demand, at least in principle. That is simply not the case with phenomena involving human beings; indeed, it is not the case for phenomena that concern the behaviors of any living systems.

Demanding that being scientific requires observations and measurements to be replicable would therefore bar, from being accepted as scientific, not only parapsychology but all of medical science and social science.