Page:Philosophical Review Volume 2.djvu/698

684 accuracy required for astronomical purposes is something far beyond that for psychological purposes in this case; yet the very thing which we want to measure psychologically, the reaction time, and which we do measure with an accuracy at present beyond our usual needs, is used to determine the unit of measurement, namely, the second; in the latter case we arrange our experiments so that the variations of the psychological quantity are negligible, in the former the inaccuracy of the physical apparatus is negligible.

It is readily seen that if we do not eliminate or render negligible the psychological sources of variation in physical measurements, as was the case in astronomy before the discovery of the personal equation, we are introducing errors into our physical results. Likewise, if we are measuring psychological phenomena, and yet do not know how much of our results and how much of the variations are due to mental influences, and how much to the apparatus, we really do not know what our results mean. It is from the side of psychologists who are not acquainted with the science of measurements that we often hear the remark that it is of no use to be exact in psychological work. They are careless in their methods, careless in their measurements, and careless in their statements. In the published accounts of the work there is often no information as to the elimination or presence of errors. An experimental result whose reliability is unknown to us is nearly worthless. In order to form a judgment on the accuracy of the result, all the necessary data must be given. Any description of a method and result can be criticised as materially incomplete if it does not give all the data needed for such a judgment. Failure to give such data can only be ascribed to the urgent necessity for condensation or to ignorance or neglect on the part of the observer; and either of the latter casts grave doubt on the character of the work.

Variations in Measurements.–If we make n independent measurements of the same quantity, physical or psychological,