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



407 



An interesting contribution from Greenleaf was a Yiddish translation of the performance 

 scale by Corpl. William Cohen. 



Criticism. — Important criticisms of the performance scale were made by camp examinera 

 regarding the method of scoring the various tests, the relative difficulty of the tests, the relia- 

 bility of results with non-English speaking subjects, etc. As to scoring, the chief points made 

 were that the scoring rule for test 1 was ambiguous; that the counting and evaluation of "moves" 

 in tests 4 and 5 were unreliable; that the possibility of chance scores in tests 1 and 9 were too 

 great and not adequately corrected in the weighting, etc. It was claimed that tests 10 and 7 

 (occasionally also 9 and 6) were too difficult, while test 1 was too easy; and that these differences 

 were not sufficiently corrected in the weighting. The manikin and feature profile (test 2) was 

 everywhere criticized as being inartistic and unrecognizable; nevertheless, results proved that 

 this test is one of the best in the scale. The "pencil and paper" tests (6, 7, and 8) received 

 an undue amount of adverse criticism, for analysis of the results shows that designs (6) is 

 probably the best single test in the scale, that maze (8) is fair, and that digit-symbol (7) is 

 very satisfactory for all subjects except the absolutely illiterate. The kinds of pictures used 

 in tests 9 and 10, especially the former, were disapproved; but it was recognized that such 

 tests would probably be valuable if a selection of pictures more suitable for adults were made. 



On the whole the present performance scale, particularly the abbreviated form, gave satis- 

 factory results, and with improvement along the lines indicated above it could be made a fairly 

 satisfactory intelligence scale, both for literate and illiterate subjects. Tests 9 and 10 were 

 not found practicable in the Army. 



Section 4. — Abbreviation of the Stanford-Binet scale. 



During the spring and summer of 1918 when 300,000 or more men were being drafted each 

 month it was found impossible to give a complete individual examination to all who failed in 

 the group examinations. Accordingly, abbreviations of the point scale, the Stanford-Binet 

 scale, and the performance scale were prepared which could be applied in from 12 to 15 minutes. 

 After September 1 these abbreviated scales were used for the majority of those individual 

 examinations which did not present exceptional or puzzling features. 



In the case of each scale the attempt was made to select for the abbreviated series those 

 tests of greatest diagnostic value as indicated by agreement with the scale as a whole. The short 

 series was intended to replace, for a certain proportion of subjects, the complete scale. It was 

 therefore necessary that it should yield results as nearly as possible the same as would have 

 been secured from the entire scale. 



The data used included 486 examinations in which four tests per year had been used, 

 usually the four-starred tests (see examination blank, Part I, p. 271). Of the 486 records, 324 

 were of unselected men who entered into the alpha-beta-Stanford experiment described in 

 chapter 7. The remairing 162 records were sent in from various camps for use in the mental 

 age standardization of the performance scale. 



The distribution of these 486 cases according to mental age, as measured by the regular 

 scale of four tests per year, is shown in the accompanying table. It will be noted that the 

 limited number of cases at the lower mental ages renders the data unreliable for the tests below 

 six years. 



Table 64. — Distribution according to mental age of subjects used in the abbreviation of Stanford-Binet scale. 



