476 MEMOIRS NATIONAL ACADEMY OF SCIENCES. [vol.xv. 



gence or above, but who can not make the score corresponding to their intelligence level 

 because of the speed requirement. Examination b giving longer time to certain of the tests 

 was designed to retest this group. It will be discussed later. 



The percentage of recruits by company units unable to make satisfactory ratings on exami- 

 nation a ranged from per cent to 50 per cent, depending on the character of the group under- 

 going examination. The total average percentages given examinations other than examina- 

 tion a because of detected low literacy are: Devens, 20.4 per cent; Dix, 8.6 per cent; Lee, 17.7 

 per cent; Taylor, 7.4 per cent. (See p. 489 of this report.) As stated above, the men 

 failing in examination a were retested with examination b, with the group skill test, or by 

 methods devised for individual examination. All who were eliminated by segregation methods 

 above described took either the group skill test or individual examination or both. 



The definition of failure to pass the examination for literates remained open. Camp pro- 

 cedure varied according to the opinions of examiners, the tentative standards of line officers, 

 or even in obedience to limitations of space and time. At one time in the course of the work 

 one camp sent to group skill examination men who made below 100 points in examination a; 

 another sent those who made 60 or less; a third, those who made 30 or less; and the fourth 

 camp only sent to group skill those making less than 20 in examination a. This particular 

 critical point concerns those recruits who had been previously judged literate. The elimina- 

 tion of examination b made it necessary to re-examine low-score men from examination a with 

 group skill or individual examination. 



It was necessary to make reports on unexpectedly large numbers of men thus eliminated 

 from examination a. Company commanders needed more than the simple statement that 

 these men were too illiterate to take the regular examination or unable to read orders and in- 

 structions issued in the military service. The individual examination methods given in the 

 Examiner's Guide would not enable an examining staff to report results expeditiously on such 

 large percentages of the recruits. The Stenquist construction test was suggested as a means 

 of testing these men in groups or singly. (See Examiner's Guide, first revision, pp. 146 f ; 

 and pi. 12, p. 91.) It constituted, therefore, part of the equipment of the four camps and is 

 referred to in Examiner's Guide as group skill test. 



The size of groups that could be examined by this performance test varied according to 

 physical equipment, space, and number of orderlies. As high as 60 or 75 men were often exam- 

 ined at once by this group method for illiterates. Relatively few men failed to show interest 

 in the test and to make an attempt to do some portion of the test. Some score was therefore 

 obtained for all recruits by the combined use of examination a and the construction test. 



Special difficulties arose in connection with the handling of this test. The breakage of parts 

 was considerable. Efforts to force pieces together twisted and deformed parts of a unit so that the 

 subject who followed lost time and score in attempting to assemble the object. Previous ex- 

 perience with the objects could not be entirely controlled, and so far as investigation went seemed 

 to be an important factor in final score. The taking apart of objects and the scoring could not 

 be made sufficiently mechanical and both must be completed before a second groujj could be 

 examined. The tendency to watch someone else for cues was so strong that special partitions 

 were necessary to prevent copying (pi. 12, p. 91). 



In the four cantonments 14,610 men were examined by this method. The statistical results 

 are given elsewhere. Technical criticisms of the tests have also been discussed. The method 

 gave a rough estimate of intelligence and enabled examiners to report some special information 

 concerning each man examined. The actual significance of the test was not determined satis- 

 factorily. (See ch. 2, pp. 321 ff, and ch. 6, pp. 363 ff.) 



The failure of the skill test to correlate highly with any of the usual tests of general intel- 

 ligence and to correlate satisfactorily with officers' estimates and performance of recruits in mili- 

 tary duties, led the four staffs to seek other group methods of testing "illiterates." The dif- 

 ficulties of administration mentioned above directed the attention of the Devens and Lee staffs 

 toward nonlanguage tests that could be given and scored like examination a. The Taylor and 



