No. 1.] 



PSYCHOLOGICAL EXAMINING IN THE UNITED STATES ARMY. 



149 



Test 9. 



If in any item the proper word in the parenthesis is clearly indicated in any manner, the item may count as right. 



However, if two or more words in a parenthesis are indicated, the items shall count as wrong. Underlining words 



outside the parenthesis may be overlooked! 



Test 10. 



If in any item the proper numbers are clearly indicated in any manner, the item may count as right, 

 two numbers are indicated in any column, however, the item shall count as wrong. 

 In any test, responses which have been corrected stand as corrected. 



If more than 



Section 3. — Instructions for scoring and combining tests of individual examination series. 



The method of scoring and combining the tests of the individual examination series was based on tests of 340 

 unselected men in Camps Devens, Lee, and Taylor. 



The method involves the following procedure: (1) A raw score is obtained for each test, A, B, C, etc. (2) Each of 

 these raw scores is transmuted into an •'equalized 3 ' or "absolute-" score by means of a table. (3) These "absolute 8 ' 

 scores are then averaged and the average is multiplied by 10 to clear it of decimals. The score thus obtained is the 

 measure of intelligence. 



The method of combining the scores of the several tests was arrived at as follows: For the tests in which time was 

 not counted, frequency distributions were plotted for number of right responses. These tests are listed in table 9. 

 For the remaining tests (listed in table 10) distributions were plotted both for time and amount accomplished. On the 

 basis of these plots, rules were drawn up for the assignment of points bath for time and accomplishment. These rules 

 are embodied in table 10. The intention of the plan is to assign equal increments of points for theoretically equal incre- 

 ments of merit both in speed and accomplishment. For each of the tests a raw score was secured according to the 

 directions in table 9, table 10, and the "Instructions for Scoring." 



The assumption was made that the degrees of performance in each test attained by the same percentage of the 

 group were equal in absolute value. The further assumption was made that the abilities of the individuals of the 

 group in each test were distributed normally. On the basis of these assumptions the achievement of the median indi- 

 vidual was assigned the absolute value of 15 points. The lower quartile achievement was assigned 10 points, the upper 

 quartile of achievement was assigned 20 points, and bther percentages were assigned values in accordance with the 

 assumption made regarding distribution of intelligence. 



This operation was accomplished as follows: Distributions of the total raw scores in each of the tests A to V, exclud- 

 ing tests F and H, were plotted in the ogive form and smoothed. A scale was then applied to the ogive in each case 

 and values were assigned to the scores attained by each percentage of individuals corresponding to the values of y 

 which would correspond to the same percentages if the ogive had been normal. The absolute values of scores ranged 

 in most cases from approximately zero to approximately 30. In several instances, however, where a large number of 

 individuals got a raw score of zero, it was necessary to give an absolute value somewhat above zero to the raw score of 

 zero. 



The intention is that a single absolute measure of the intelligence of an individual may be obtained by averaging 

 the absolute scores in any number of tests, and multiplying by 10 to clear of decimals. Assuming the data on which 

 the method was based to be representative, the absolute (that is, the weighted or equalized) total score of an individual 

 of median ability should be approximately 150. The upper and lower quartile scores should be approximately 200 

 and 100, respectively. 



Table 9. — Scoring of tests in which time is not counted. 



Table 10. — Scoring of tests in which time is counted. 

 TEST A. 



