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



8. Unfavorableness to cheating. — Little difference among the tests, except that copying 

 by one man from another is somewhat easier in oral directions, arithmetical reasoning, practical 

 judgment, and memory for digits than in the other tests. 



9. Nondependence upon schooling. — Members of the committee did not agree as to the in- 

 fluence of schooling, except in believing that the oral directions test is rather exceptionally 

 free from such influence. Some regarded the analogies test and the synonym-antonym test as 

 more or less objectionable on this ground. Results have given no indication that the tests 

 differ greatly in the extent to which schooling influences the result. 



10. Response without writing. — No writing is required, except the writing of figures in the 

 arithmetical reasoning test, the memory for digits test, and number series completion. 



11. Interest ami appeal. — It was expected that the tests chosen would satisfy this criterion 

 very well, although it was not expected that the rank order of the tests on this point would be 

 the same for all types of subject. 



12. Economy of time. — The synonym-antonym test and the practical judgment test stood 

 first in this respect. Counting the time necessary for reading the instructions, each required 

 about 2 minutes. Oral directions, arithmetical reasoning and memory for digits were least 

 economical of time, requiring from 5^ to 7 minutes apiece. 



By June the materials for the group test were ready for press. It was decided to print 

 an edition of 1,000 copies for preliminary trial of the group method before undertaking to pre- 

 pare methods for individual examining. Bingham accepted the responsibility of seeing the 

 material through press. Two days later printed copies were distributed and the committee 

 adjourned to gather data which would serve as a basis for revision and standardization. 



The recess lasted from June 10 to 23. During this period trial was made of the group 

 method by various members of the committee as follows. By Yerkes, 50 inmates of the Massa- 

 chusetts School for the Feebleminded, Waverly, Mass., and 25 subjects at the Boston Psycho- 

 pathic Hospital; by Wells, about 50 subjects at the Reformatory for Men, Concord, Mass. ; by 

 Haines, about 50 aviation recruits in Ohio; by Terman, about 60 high-school pupils and 43 

 prisoners in California; by Bingham and Goddard, assisted by N. J. Melville, 114 marines at 

 the Philadelphia Navy Yard and 27 men at the Carnegie Institute of Technology; by Whipple, 

 50 men in the officers' training camp, Fort Benjamin Harrison, Ind. 



The committee reassembled at Vineland on June 25 to work over carefully the results of 

 the above tests giving particular attention to correlations with other measures of intelligence, 

 and to methods of scoring. 



The 103 high-school pupils and prisoners of California had been tested by the proposed 

 Army test, by the Stanford-Binet scale, the Yerkes-Bridges point scale, and the Trabue lan- 

 guage tests. The results of the Army test, both for the scale as a whole and for the 10 different 

 tests composing it, were correlated with these other measures. A large number of the marines 

 were given an abbreviated Binet test, and these results also were correlated with the army 

 test. In addition, for certain groups of cases, each of the 10 tests in the Army scale was cor- 

 related with the total score of the scale. For certain groups each test was correlated with the 

 sum of 5 tests. 



Unfortunately some of the data of this preliminary trial are not at present available. The 

 following facts, however, can be presented: 



1. For Terman's 103 subjects (60 high-school pupils and 43 prisoners) the correlation of 

 each test with Stanford-Binet mental age was as follows : 



Test 12345 6 789 10 Total score. 



Correlation 0.66 0.56 0.71 0.73 0.73 0.75 0.77 0.25 0.82 0.29 0.87 



2. For the marines the 10 tests correlated with mental age scores by Doll's brief Binet 

 (two tests per year) as follows: 



Test 12345 6789 10 Total score. 



Correlation 0.40 0.37 0.68 0.57 0.60 0.68 0.72 0.47 0.50 0.31 0.81 



