no. 2.] PSYCHOLOGICAL EXAMINING IN THE UNITED STATES AKMY. 



319 



Concerning the interpretation of total score, the following is quoted from Thorndike's 

 report : 



Adult inmates of feeble-minded asylums, of Binet ages 9 to 13, in general make in the test scores from —188 to 



— 60. Their median score is —140. 



Men of the Regular Army reporting themselves as "laborers" or "farmers," who have English names and report 

 at least some schooling, in general make scores from —100 to 0. The median score of 100 such taken at random is —55. 



Men of the Regular Army (including a few from officers' training camps) who have English names and some 

 schooling and who report themselves as skilled laborers, such as electricians, engineers, draughtsmen, carpenters, 

 masons, photographers, and the like, make in general scores from —70 to +50. The median score of 53 such is —15. 



Fifty men in training in an officers' training camp made scores from to +180, with a median a trifle over +100. 



Eighty-seven students in a course in psychology primarily for seniors and graduate students made scores from 

 +50 to +175, with a median at about +125. 



Four hundred and fifty men of the Regular Army (camp in Indiana) showed only 16 men, or 35 per thousand, with 

 scores of —140 or worse; and there is evidence that at least half of those men (below —140) could read English very 

 little or not at all. 



There is evidence that if a man can read English as well as, say, the average fifth-grade child, and still scores below 



— 140 in the test, he either is not admitted, or, if admitted, is not found worth retaining in the Regular Army. 



It seems, then, that any literate man who scores below —140 should be considered for exclusion from the Army, or 

 for retention only for work in which even gross stupidity could not endanger his fellow soldiers. A man scoring —100 

 should be subjected to further study, especially by the aid of performance tests (clock, maze, form board, and Sten- 

 quist) in a group test, and by tests of his ability to manage himself and other men so as to gain fair repute amongst his 

 associates. 



At the other end of the scale it may be said that no man scoring below zero is likely to succeed in the work with 

 books, maps, instruments, and the like of a training camp for commissioned officers. Such a man probably should not 

 be sent there, as he will almost certainly fail. Roughly, a minimum standard of + 50 might be set for admission to such 

 a camp, with character, leadership, and military zeal and knowledge the deciding factors for men scoring + 50 or better. 



More systematically we may interpret the scores as shown below (assuming that the individual speaks English 

 well and has had good opportunity to learn to read English): 



— 150 feeble-minded adult, of Binet age, 9 to 10. 



—100 upper level of the "borderline" or the lower level of the dull, slow- thinking day laborer. 



— 50 unskilled laborer. 



0. 

 + 50 lower level of men fit for appointment as officers. 



+ 100 level of a man who gains success in a profession or as an officer in the Army or the Navy by intellect. 

 + 150 level of a very rapid and exact thinker with words, numbers, and abstract relations. 

 For the purposes of the Army the abstract ability measured by the test above + 100 may well be irrelevant, but 

 from + 100 down the scores certainly have value. This is proved by the correlation with the officers' ratings of the 

 ability and promise of the men. This correlation averages about 0.5. It seems likely that with an omniscient judgment 

 of the intellect of a man the score in the test would correlate with that judgment to the extent of 0.7 or 0.8 (assuming, 

 as previously, that the test is used with men who have had good opportunities to learn to speak and read English). 



At the four examining stations individual examinations were given to between 200 and 300 

 men. It appears that only 50 of these tests were scored. This obviously is entirely too small a 

 number either for purposes of standardization or for important statistical treatment. The 

 only results reported by the statistical unit are the following correlations (Pearson coefficients 

 as estimated by method of unlike signs) with ratings of men by their officers. 



The data treated were scanty and the method of correlation rough. Nothing was reported 

 about the nature of the distributions of the ratings and scores from which the correlations were 

 computed. Without such data the correlation coefficients themselves have little significance. 



