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



405 



It will be noted that the first four and the last two tests are the same in both orders. The 

 form board, by its low position, is indicated as probably the poorest test in the scale, and one 

 that could therefore be dropped without loss. Inspection of the distributions showed that the 

 picture completion test, while not one of the lowest from point of view of the above correlations, 

 presented an unfortunately large proportion of zero scores (38 out of 134). Dropping these 

 two tests (form board and picture completion) the total score on the remaining eight tests 

 was determined for each of the 134 cases. These scores show a correlation with Stanford-Binet 

 mental-ages of 0.834, while the complete scale scores correlate only 0.841 with the same ratings. 



The five tests of the short scale seem to have been very well chosen. The picture-arrange- 

 ment test, which has a high place in both orders above, was not included because of the time 

 required to give it, and also because of its difficulty in the lower end of the scale. There 

 seemed to be no reason for changing the other tests, except that for the absolutely illiterate 

 (whether foreign or American born) the ship or the cube imitation tests should be substituted 

 for the digit symbol, which gives a rather high percentage of zero scores. 



The short performance scale gave the following correlations: Short scale with long scale 

 (134 subjects), 0.97; short scale with Stanford-Binet (260 American born), 0.84; short scale with 

 Stanford-Binet (61 foreign born), 0.70. 



Tables 61, 62, and 63 give typical correlation- arrays for performance scale and Stanford- 

 Binet scores. 



Table 61. — Short performance scale and Stanford-Binet mental age. (r=0.842). 



Table 62. — Designs and Stanford-Binet mental age (r=0.7S5). 



