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.

    103/134 (page 91)
    or approximately 1.803 (Fa) = 3[1.667 (7a)+1-3 (6a)+-4 (5a)+-7 (4a)]. This combination of elementary school records is designated as (7, 6, 5, 4a), and the correlation, 6. 5, 4a) = -789. The same equation is used to obtain measures in elementary school mathe- matics and English, except that division by three is omitted, giving the follow- ing: (7, 6, 5, 4m) = 1.667 (7m)+1.3 (6m)+.4 (5m)+.7 (4m) (7, 6, 5, 4e) =1.667 (7e)+1.3 (6e)+.4 (5e)+.7 (4e) Calculation gives: rFM(7, 6, 5. 4m) = -680 and e, 5. 4je) = .710 No history was taken during the first high school year so there are no history correlations. It may be noticed by reference to the preceding table that and rp^SA are less than would be expected from the other correlation coefficients. This may be due to the teachers of these particular 7th and 5th grades being less expert in estimating the ability of pupils than the 6th and 4th grade teachers. Whatever the cause, probably a better regression equation for general pur- poses can be obtained than the one given above. The accompanying curve r Between Grade IN A Given Year AND Grade One Yr. Before r Between Grade IN A Given Year AND Grade Two Yrs. Before r Between Grade IN A Given Year AND Grade Three Yrs. Before r Between Grade IN A Given Year AND Grade Four Yrs. Before Av. of 4 coef s. = .6415 Av. of 3 coefs. = .5753 Av. of 2 coefs. = .541 1 1 coef. =.624 was drawn with this end in view. A smooth curve, not rectilinear, is drawn near the points representing the ordinates for the various abscissae. The intersections of the curve with the ordinates give the values of the correlation coeflicients in the succeeding table. The falling off in correlation from year to year is thought to be reasonable and calculation will show that the sum of the deviations of the actual coeflBcients of correlation from the points where the curve crosses the ordinates at the various abscissae very nearly equals zero, so that the curve is not entirely arbitrary.