CHAPTER 5. 



METHODS OF SEGBEGATION. 



Section 1. — Literacy tests. 



The unofficial trial of examination a demonstrated that a large proportion of the soldiers 

 could not properly be given the group psychological examination because of their inability to 

 read English. In August, therefore, the committee in charge recommended that this examina- 

 tion be supplemented by the Stenquist mechanical test. In January, 1918, the beta examina- 

 tion was substituted for the Stenquist. Some sort of segregation of examinees into groups of the 

 relatively literate and relatively illiterate was consequently necessary. To insure the individual 

 an opportunity to secure the highest grade of which he was capable, as well as to make possible 

 a comparison between groups, an accurate and uniform method of segregation was required. 



Modified Thorndike test. — For this purpose a modification of Thorndike's Reading Scale, 

 Word Knowledge or Visual Vocabulary, 1 was prepared by Terman. The modification consisted 

 in the following: Changes in instructions to make them applicable to larger groups; reduction 

 of the time limit from an indefinite period to 3 minutes; elimination of the preliminary test; 

 reduction of the categories from eight to four; omission of the lines of test words suited for 

 the odd-numbered and half-year grades; and reduction of the number of test words in each 

 grade from 10 to 5. Four forms of this "Literacy test" were printed and distributed to the 

 four original camps. These forms, together with the directions and rules for scoring, are repro- 

 duced on pages 279-280. 



Various unforeseen objections to the method of segregation as outlined arose almost 

 immediately. The tentative norms and the instruction to retain only men of fourth-grade literacy 

 or better set far too high a standard. Camp Devens, for instance, reported a group where men of 

 literacy grade 2 or below had been allowed to take examination a; 37 percent made a grade of C 

 or better. Camp Dix reported that, in an average group of 307 men, 35 per cent scored zero 

 and 16 per cent more scored second grade in the test; thus 51 per cent were excluded from 

 examination a. In another group of 222 cases, selected to secure a random sample, 32 percent 

 scored zero and 20 per cent second grade, leaving 52 per cent "illiterates" for the Stenquist 

 test. From a study of other large and typical groups it appeared that 19 per cent of the men 

 who scored zero in the literacy test secured grade C or better in examination a, and of the men 

 scoring second grade in the literacy test, 78 per cent made grade C or better in examination a. 

 The median score for men of second-grade literacy was 137.5 ; the lower quartile, 106.5. The test 

 correlates with soldiers' reported school grade about 0.68. 



Further evidence that the norms originally furnished to the examiners were too high appears 

 in Table 27. The first line of the table proper is an interpretation of the norms originally fur- 

 nished to the examiners, expressed in terms of the number of words correctly marked. The 

 second line of the table shows the averages and probable errors, by grades, of correctly marked 

 words in a typical study of 313 cases at Camp Dix. Grade in this latter case is the grade as 

 reported by the soldier. 



Table 27. — Norms/or the modified Thorndike literacy test. 



i Teachers' College Record, vol. 15, 1914, pp. 207-220; vol. 17, 1916, pp. 431-433. 



347 



