CHAPTER 4. 



REVISION OF GROUP EXAMINATION a. 



Group examination a was notably successful from the beginning. It was interesting, 

 relatively easy in administration, and for literate men yielded ratings which, in the judgment of 

 those who had occasion to use them, were surprisingly dependable. It was this method chiefly 

 which gave the psychological service its early prestige and paved the way for later developments. 



However, during the trial period, as a result of special studies which had been made of 

 this method of examination, a number of faults had become apparent. Of the 10 tests which 

 composed the scale, some were found to give low correlations with other measures, some gave 

 too large a proportion of zero or of perfect scores, some contained ambiguous items, some were 

 improperly weighted. It was generally believed tha,t the scale as a whole was rather less 

 reliable for high-grade than for average men, and that the reliability at the lower extreme was 

 also doubtful. As it was expected that a proposed new group test for illiterates would change 

 the situation with respect to requirements to be satisfied by the group test for literates, revision 

 of examination a was undertaken in connection with the general overhauling of methods in 

 preparation for their extension to the entire Army. 



Preparations for thorough revision of all the initial methods of examining were made by 

 the staff of the Division of Psychology in December, 1917. In January, 1918, after the War 

 Department had ordered the extension of examining, a group of psychologists on civil or 

 military appointment was assembled in the Office of the Surgeon General for this special work. 



The task naturally fell into three divisions — the revision of examination a ; the provision of 

 a substitute for the Stenquist group test as a test for illiterates, and the modification and sup- 

 plementation of the methods of individual examining. 



Capt. C. S. Yoakum, because of his intimate acquaintance with the conditions and results 

 of examining in the camps, was placed in charge of the revision of examination a. He was 

 assisted over a period of weeks by C. C. Brigham, Margaret V. Cobb, E. S. Jones, L. M. Terman, 

 and G. M. Whipple. Some assistance was also rendered by the psychologists who, under the 

 direction of Lieut. W. S. Foster, were engaged in developing a new group examination for 

 illiterates. The group included C. R. Brown, A. S. Otis, K. T. Waugh, and R. H. Wheeler. 



Terman and Whipple served as advisory members of the staff of the Division of Psychology. 

 Data which proved extremely useful in connection with revision of methods were supplied by 

 Terman and Haggerty as a result of the use of the Army mental tests in high schools and col- 

 leges. This supplementary examining was done by request of the Division of Psychology in 

 order that age and grade norms might be made available and the validity of the methods 

 measured for other than Army groups. 



The data used for the revision included results of investigations along the following lines : 

 (1) The value of the separate tests of the scale in differentiating officers from enlisted men, and 

 enlisted men from the institutional feeble-minded. (2) Score distributions for various types of 

 groups, both for the entire scale and for the separate tests. (3) Correlational material, includ- 

 ing correlations of each test with officers' ratings of men; of each test with total score; of each 

 test with Binet mental age, school grade, teachers' ratings, school marks, etc., in the case of school 

 children. (4) The effect of changing the time limits. (5) Effects of different methods of scor- 

 ing and weighting. (6) Relative difficulty of forms. (7) Objectionable items in the tests. 



In order that the reader may be in position to compare the merits of examination a and 

 its modified form alpha, the data upon which the revision was based will be set forth here in 

 some detail. 



327 



