No. 3] PSYCHOLOGICAL EXAMINING IN THE UNITED STATES ARMY. 823 



The chief interest of this study for the present report is the possibility of comparing it 

 with the Wadsworth study in certain particulars. Unfortunately there are only 29 occupa- 

 tions common to both studies. These 29 occupations were ranked according to the median 

 scores shown in the Wadsworth study and also as shown in the Devens study. These two 

 rank orders were then compared. The Spearman R value equals +0.71, which corresponds 

 to a Pearson r value of + 0.90. This indicates that in general the relative rating of these 29 

 occupations according to intelligence is much the same in the Wadsworth study as in the 

 Devens study. It is interesting to note that of these 29 occupations 23 had a higher median 

 score in the Devens study than in the Wadsworth study and only 6 had a lower median score. 

 The average difference in score is 16.3 points. The median score of the literate whites in the 

 Wadsworth draft was 121. The median score of the literate whites in the Devens survey 

 was 133. This is a difference of 12 points. It is probable that the higher median scores both 

 for the occupational groups and for the general survey groups in Camp Devens as compared 

 with Wadsworth groups was due to a slightly higher literacy standard in the former camp. 



It was pointed out in discussing the Wadsworth study that in all probability the median 

 score of 99 points for printers was unreliable. That this point of view was correct is indicated 

 by the fact that the Devens printers have a median score of 162 points. This disagreement 

 of 63 points for a given occupational group is the greatest disagreement that was found between 

 the two studies. 



The similarities found in the above comparisons indicate that these two occupational 

 investigations are roughly accurate and justify themselves on practical grounds. 



The interest shown by personnel officers in the Wadsworth study indicated that such a 

 table of occupational intelligence ratings was of considerable practical value for Army personnel 

 work. The Division of Psychology of the Surgeon General's Office planned a much more com- 

 prehensive study and sent out a letter to all chief psychological examiners requesting that 

 score distributions for certain occupational groups be tabulated and reported to the Surgeon 

 General's Office at the earliest possible moment. The letter requested data on 114 occupations, 

 an urgent list of 69, and an additional list of 45. These occupations were designated by the 

 Committee on Classification of Personnel for such special report. Instructions covering the 

 method of securing and recording the data were also included in the letter. Effort was made 

 to avoid the errors of the earlier studies. The question of precise standard of occupational 

 definition and classification was solved by referring in each case to the soldier's qualification 

 card. It was planned to secure enough cases in each occupational group to give reliable 

 standards, and to insure this each examining station was requested to report on from 1,000 to 

 10,000 cases. Provision was also made to record both alpha and beta scores. 



The work of assembling and analyzing the data was done by Dr. James W. Bridges in the 

 Office of the Surgeon General, and was completed by September 1, 1918. At that time reports 

 had been received from 16 camps, viz: Sherman, Sheridan, Lee, Greenleaf, Logan, Wheeler, 

 Custer, Grant, Cody, Jackson, Dix, Travis, Pike, Gordon, Funston, and Taylor. 



The report as made by Dr. Bridges is as follows: 



After the data were tabulated and summarized, it was found that for many occupations there were too few cases 

 to give a reliable range or central tendency. Upon the basis of reliability of results, the 114 occupations were grouped 

 into five classes. Each class contains occupations for which there is a certain range of cases, and the steps from group 

 to group represent equal decreases of reliability. The limiting numbers are roughly the squares of 16, 12, 9, and 6. 

 The groups, range of cases, and number of occupations in each are as follows: 



