558 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



[Vol. XV, 



Group II: White draft, additional, by States. — It was intended at the time the plan for analysis 

 was made to make a comparison of States with respect to intelligence. Examination of the 

 figures of table 1 shows that the samples for separate States in Group I would be in many cases 

 inadequate as a basis for comparison. It may be assumed that New York State would be fairly 

 sampled by 4,511 cards or that California might be represented by its 1,233 cards, but it would 

 obviously be unfair to conclude to statements of the intelligence of the Maine draft on the 

 basis of only 376 cards or of the Wyoming draft on the basis of 87 cards. The intention in 

 Group II was, therefore, to select sufficient additional cards for every State so that the total 

 number of cards in Groups I and II for that State should not be less than 1,000. Table 2 gives 

 an analysis of Group II, and the last column of table 1 gives the totals for Groups I and II 

 together. From this column it will be seen that cards proved not to be available in sufficient 

 numbers for Arizona, Colorado, Delaware, Florida, Nevada, New Mexico, Utah, Virginia, and 



Wyoming. 



Table 2. — Group II: White draft, additional by States. 



Numbers of cases in Group II of the principal sample. Additional selections intended to bring the pro-rata samples 

 of less than 1,000 in the small States of Group I up to a minimum of 1,000 for each State. This total of Groups I 

 and II is shown in the last column of table 1. It was not always possible to obtain 1,000 cases, however. 



Group HI: White draft, additional, by camps. — Groups I and II provide a sampling of the 

 white draft of about 56,000 cards. The initial intention in the plan of analysis was to make 

 the sample of the white draft approximately 100,000. It is plain that it would not be possible 

 to have drawn a larger sample, pro-rating by geographical distribution, for certain States have 

 already fallen short of the desired quota. Group III is an additional group, chosen without 

 respect to State, principally for the sake of obtaining larger numbers. It brings the total for 

 the white draft not far short of 100,000 (96,354). Subsequent investigation has shown that the 

 distribution of scores in Group III is not very different from the distribution in Group I. 

 Accordingly it appears a posteriori that Groups I, II, and III do actually constitute a fairly 

 representative sample of the white draft and one that has the advantage of including a large 

 number of cases. In order to make Group III as representative as possible, the group was 

 selected by approximately equal increments from the 15 camps listed in the description of 



