No. 2.] PSYCHOLOGICAL EXAMINING IN THE UNITED STATES ARMY. 381 



language difficulty and those not so handicapped and for the sake of brevity shall designate 

 them as "English speaking" and ''non-English speaking." 



The figures of Table 4S indicate the numerical relations of the beta groups to the total 

 group of which they are a part: 



Table 48. — Camp Custer group. 



Group differentiation. 



Total number reported 



Number below 100 in alpha (D and E cases) J 



Number below 50 in alpha (E cases) 1 



Number given beta only 



Number either below 50 in alpha or given beta only.. 

 Number rated E by beta 



English speaking. Non-English speaking. 



Number. Per cent 



1,904 

 585 

 231 



19 

 250 



10 



100.0 

 30.7 

 12.1 



1.0 

 13.1 



0.5 



Number. Per cent 



386 

 194 

 133 

 134 

 267 

 63 



100.0 

 50.3 

 34.4 

 34.7 

 69.2 

 16.3 



Total. 



Number. Per cent. 



2,290 

 779 

 364 

 153 

 517 

 73 



100.0 

 34.0 

 15.9 



6.7 

 22.6 



3.2 



1 Letter ratings based on standards in lorce at time of examining. 



Correlation tables were plotted for alpha score with beta total score and with score on each 

 separate test. The values of the coefficients were not computed, as it was realized that these 

 would necessarily be low, since we were dealing with only a segment of the whole distribution, 

 and since, moreover, it was uncertain whether differences in alpha scores within this range had 

 much significance. From inspection, however, it seemed evident that if the group falling low 

 in alpha represented, with even approximate fairness, the lower portion of the total group a 

 surprising number of cases were making high scores in the various tests and on the total beta 

 score. This was brought out quite strikingly when the data on the 250 English-speaking cases 

 who had either made E in alpha or been given beta only were summarized. Reference to 

 Table 48 shows that these cases made up only 13.1 per cent of the total English-speaking group. 

 Since they were not handicapped by language difficulty it would appear safe to assume that 

 they were not far from being the lowest 13 per cent of the group. Assuming even partial 

 correctness for this statement the following proportions of cases in the various grades as 

 assigned for beta at this time seemed difficult to justify. 



Percentage!! making given grades on beta — 250 English-speaking men, either E in alpha or eliminated from alpha — 



Camp Custer. 



Beta grade.. 

 Percentages . 



E 

 1.6 



D 



26.4 



C- 



19.6 



C 

 41.2 



C+ 

 10.4 



B 



0.8 



A 

 0.0 



Table 49 shows the percentage distributions of the English-speaking cases for the separate 

 tests. Without discussing this in detail it ma}- be noted that the majority of the tests showed a 

 considerable proportion of scores in the upper part of the range of the given test. While 

 this would not be a valid objection, provided the tests were to be used only for this lower range, 

 it becomes serious when beta is used for the whole range. The only tests for which an appre- 

 ciable number of high scores did not appear were tests 4 and 5 (digit symbol and number 

 checking) . 



