670 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



[Vol. XV, 



thing, some of the differences between camps may be explained by these varying percentages 

 of men who took the different examinations. 



Another source of inaccuracy may have been in the selecting of the cards for the various 

 groups. If there was any factor of selection working when the cards were chosen it would be 

 possible to get results which did not picture the camps as they actually were. A rough index 

 of the validity of the selections is furnished by a comparison of Groups I, II, and III as a whole. 

 If these large groups give comparable results, selection is more likely to have been free from 

 gross prejudicial factors. Table 191 gives the percentage distribution of the alpha scores of 



GROUP I 

 GROUP I 

 GROUP HI 



I I I l | i I i i [ i i i i | i i ' i i j 'i I i ' I | i i i i | i r i i | i '*i " t "T" f " l 

 O 25 60 IS IOO I25 I50 I75 200 



Fio. 10.— Percentage distribution of the alpha scores o( men in Groups I, II, and III who were graded on the basis of alpha. 



the men in Groups I, II, and III who took alpha only and these same data are presented 

 diagrammatically in figure 10. Both the table and the figure show that the differences between 

 the groups are not great. In table 192 there is presented a percentage distribution of beta 

 scores of men in Groups I, II, and III who took beta only or alpha and beta only. Figure 1 1 

 presents the facts shown in table 192. In this case there is a slight difference in the groups 

 between the scores of 15 and 45. Groups I and II differ rather widely through this interval, 



GROUP I 

 GROUP E 

 GROUP m 



i 1 1 1 I 1 1 1 1 1 1 1 1 1 r 



25 50 75 



Fig. 11. — Percentage distribution of beta scores of men in Groups I, II, and III who were graded on the basis of beta. 



while Group III falls practically between them. However, when the facts are considered in 

 the light of the letter grades this difference practically disappears. Probably the chief dif- 

 ference between the three groups lies in the fact that in Group II a larger number of the men 

 were graded on the basis of beta than was the case in the other two groups, where the per- 

 centages of men who were graded on the basis of alpha and the men who were graded on the 

 basis of beta were about equal. The percentage of men in each group who received final grades 

 on the basis of each examination is shown in table 193. 



