Page:Popular Science Monthly Volume 81.djvu/457

Rh

We have then for each of the hundred a record such as is shown in the case of one of them in Table 2. These data are obviously subject to

certain errors of memory, prejudice, carelessness and the like, which will, later, be given due attention. It will be best to consider first what the meaning of the records would be, were each number a perfectly true statement of the relative strength of the interest or ability in question. Consider then this sample record as perfectly true and compute from it the differences between each subject's position for interest in the last three years of the elementary-school period (column 1) and for the high-school period (column 3).