318 MEMOIRS NATIONAL ACADEMY OF SCIENCES. [Vouxv, 



Table 3. — Reported amount of schooling in successive years. (E, elementary grade; H, high school; C, college.) 



The total score in the group test (for men with English or Irish names; correlates with 

 reported amount of schooling 0.67 ±.015. 



Forms A, B, and C of the Vineland group examination were used in all four camps. Form 

 D was used in all except Syracuse, and form E in all except Brooklyn. In comparing the 

 score distributions for the purpose of determining the relative difficulty of the five forms it was 

 necessary, because of selective factors affecting camps, to treat the four camps separately. 

 Since only 700 to 900 men were tested in a camp, the number for any one form was too small 

 to allow anything like an exact determination of the relative difficulty of the forms. The 

 indications were that the differences were negligible except for a slightly greater difficulty of 

 form B. The data were summarized by Thorndike as follows: 



Form B is harder than form A by 13 points±1.7; form C, equal in difficulty to A: (difference of 0.2 points±0.1); 

 form D, substantially equal to A: (difference of 3.2 points±1.6); form E, substantially equal to A: (difference of 

 2.5 points ±1.2). 



As far as the results can be accepted as representative, they confirm the validity of the 

 "random selection" method of Wells for making up the separate forms. 



It had been expected that the results of the experiment would make it possible to rear- 

 range the items of all the tests in order of difficulty, from easiest to hardest. However, the 

 data were not satisfactory for this purpose, because the time limits allotted to the tests had 

 prevented the majority of the men examined from reaching the most difficult items in the 

 case of tests 3 to 10. The items of test 2 were already in order of difficulty, and in the case 

 of test 1 the order is not as vital as it is for the majority of the tests. 



On the relation bf total score to salary earned, Thorndike reported as follows: 

 The relation of score in the test to reported salary earned at a given age by men engaged in the same trade is of 

 interest, but our data do not permit rigorous treatment. The relation when different trades are mixed is negative 

 for young ages; positive for older ages. That is, a young man scoring low in the tests reports earning more per week 

 than a man of equal age who scores high. This reverses somewhere around age 27. This is, of course, more or less 

 what one would expect. 



