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



331 



Table 12.- 



-Per cent of 920 unselected men (the Three hundred and third Engineers, Dix) who attempted one-fourth, one-half, 

 three-fourths, or all of the items of each test, and per cent who succeeded. 



Score distributions on the separate tests of examination a for various typical groups are 

 shown on pages 534-536. Other material in Part II, chapter 11, also is of interest in a com- 

 parison of alpha and examination a. 



Section 4. — Correlations with, officers' ratings of men. 



The only ratings available at the time were: (a) 313 National Guard men (same as on 

 p. 315f); (b) 438 Lee draftees (averages for three companies taken separately); (c) 267 Lee 

 subjects with total score above 100; (d) 362 Lee subjects with total score below 200; (e) 213 

 Fort Snelling officers' training camp students. 



Table 13 gives the correlation of each test with ratings for the different groups. The ratings 

 of Fort Snelling students were the intelligence ratings based on the officers' rating scale devised 

 by Walter Dill Scott and later adopted for general use in the Army. The men of the other 

 groups were rated by their officers as 1, 2, 3, 4, or 5 in general intelligence. 



Table 13. — Correlations between the different tests and officers' ratings of their men. 



Some of the conclusions from table 13 are as follows. Test 2 on the whole is a poor test 

 except in the lower range of intelligence. Test 3 is equivocal, giving a high correlation with 

 National Guard cases and a low correlation with the Lee data. Tests 4, 5, and 6 are among the 

 best tests. Test 10 is not far from average. The form of test 9, used with the National Guard 

 cases, differed somewhat from that used later, and may explain its relatively low correlation 

 in that group. 



Section 5. — Correlations with Stanford-Binet mental age. 



Correlations of the separate tests with Stanford-Binet mental age were available as indi- 

 cated in table 14. 



