Page:Popular Science Monthly Volume 81.djvu/460

454 and computing from it the difference between each subject's position for interest in the elementary-school period (column 1) and its position for ability in the same period (column 2), we have:

These facts are repeated in the first column of Table 4. Similar facts for this same individual, for the differences between the order for interest and the order for ability in the high-school period and in the college period are given in the second and third columns of Table 4.

If at any period an individual has greatest ability in the subject which is most interesting to him, next greatest ability in the next most interesting subject, and so on, the sum of the seven differences for that period will be zero. If the order of ability was as unlike as possible to the order of interest this sum would be 24, and if the relation between interest and ability was that of mere chance this sum would be 16. Any quantity below 16 as the sum of a column in Table 4 then means some positive relation or resemblance between the individual's degrees of interest and his degrees of ability.

For the hundred individuals studied this figure is on the average approximately 5, being practically the same for the elementary-school period, for the high-school period and for the college period. This average result may be expressed as a coefficient of correlation of .88. Nearly three fourths of the individuals show records between 2 and 8, inclusive—that is, correlations of from .70 to .98.

If, in the case of any individual, we add together the three ranks for each subject in interest at the three periods and do likewise for its ability-ranks, we have measures of the general order of the seven subjects