Educational guidance : an experimental study in the analysis and prediction of ability of high school pupils / by Truman Lee Kelley.

  • Truman Lee Kelley
Date:
1914
Catalogue details

Licence: In copyright

Credit: Educational guidance : an experimental study in the analysis and prediction of ability of high school pupils / by Truman Lee Kelley. Source: Wellcome Collection.

    105/134 (page 93)
    Similarly, ^ fi _|_2 r O'Xi+Xz+X, = ^Xi Y 3 Finally, /l + (n—l)r <^Xi+Z2+ ---Xn = Y n In order to make the standard deviations of measures which are averages of var5dng numbers of estimates comparable, it is only necessary to divide the measures by their respective sigmas, i.e., if but a single measure divide by sigma Xi~\-X2 Xi] if an average of two measures divide by sigma —2— > ®tc. The follow- ing table gives the desired divisors for the various cases’.^ No. of grades averaged I. a. Intellectual ability Cons. Conscientious- ness Emo. i. Emotional interest Exp. Expression 1 .92 .92 .92 .95 2 .74 .79 .74 .76 3 .66 .73 .66 .68 4 .63 .69 .63 .65 5 .61 .67 .61 .63 6 .59 .66 .59 .61 7 .58 .65 .58 .60 The mean number of grades averaged to obtain each individual’s measure is about two and one-half, therefore the reliability coefficients have approxi- mately the following values: ^(i. a. measures as derived) (i. a. meas. similarly derived) — *493 ^(cons. “ “ ) (cons. “ “ )=.605 ^(emo. i. “ “ ) (emo.i. “ “ )=.505 *^(exp. “ *• ) (exp. » ) = .529 In 23 cases of group 2, teachers’ estimates were not available, except teachers of English and history who later had the same pupils in test courses, which were continuation courses of the first half year’s work under the same teacher. The estimates of such teachers were not used when it could be avoided, i.e., in all cases except these 23. The following correlations justify excluding such estimates: ^G(i. a.—estimate of geometry teachers) ^ ,/ P^P* “ ^G(i.a.— “ “other “ ) = -44 \ group 1 ^E(i. a.— “ *‘ English “ ) =-54 / pop. = 33 »*E(i.a.— “ “ other “ ) = .36lgroup2 ^This method of averaging varying numbers of correlated measures was used frequently in other portions of this study, e.g., in averaging grades of pupils for some given term or year where the number of studies varied appreciably.