68 TWENTY-FIRST REPORT. 



training principally in the respect of their lower synaptical resistence- 

 thresholds. They are the quicker, better learners, so gifted in inheritance. 

 This achievement-capacity is regarded as only slightly, if at all, influenceable 

 by training." 



The sixty subjects were members of a non-commissioned oflScers' training 

 school, from which number ten were to be selected for officer training. The 

 problem was to select, by means of tests, the men of greatest ability ; to keep 

 secret these names and after the official selections were made to compare test 

 findings with Officer's judgments ; and to determine the nature and extent of 

 variations in abilities in subjects whose ages, health, training and social 

 status was approximately the same. For, if the same wide ranges of differ- 

 ences exist after twelve years of scholastic experience, one is forced to con- 

 clude that this training has had little or no influence in equalization of 

 chances for success. 



Assisted by his colleague. Professor T. S. Henry, the writer gave six group 

 tests to the sixty men. The tests employed were selected because of their 

 known high correlation with general intelligence and because they are said to 

 measure traits essential in the officer's make-up. These were : 



1. fhe Franz Rectangular Maze (No. 75,224). 



2. Woodworth and Wells' Hard Directions. 



3. Opposites. Whipple's adaptation of Simpson's list. 



4. Analogies. Whipple's List C. 



5. .Substitution. The Digit-Symbol Test. 



6. Completion. Trabue Language Completion, Scale J. 



Securing a control group of illiterates of the same ages and opportuni- 

 ties was not only impossible but was also regarded as unnecessary. 



The tests were given and scored in the ways of accepted usage. Test con- 

 ditions were of the ^ery best, as reported by both subjects and examiners. 

 Two minutes on the first five and five minutes on the last test were allowed. 

 The subjects were then ranked according to scores in each test and these ranks 

 were then amalgamated as the final rank. By the Spearman "foot-rule" 

 method the following correlations of the tests with final ranks were obtained : 



Analogies 81 



Opposites 71 



Substitution 71 



Directions 66 



Conipletiiin 64 



Maze 40 



The tests were found to be more reliable in the selection of men of ability 



than the combined judgments of the Officers. The sixty men's scores, arrayed 



in a histogram, show practically the true Galton ogive. This means that the 



selections from the total group made by the officers was a pure one-to-one 



chance picking, and as to the final selections of the best ten the tests gave a 



more reliable sifting in the I'l minutes actual working time than the officers 



got from their 72 days of observational study of the behavior of the men. 



