372 



MEMOIRS NATIONAL ACADEMY OF SCIENCES. 



[Vol. XV, 



The revision was based upon data from three sources: 



1. At Camp Meade (February, 1918) the Three hundred and fourth Ammunition Train 

 was examined as an experimental group. This gave literacy, alpha, and beta records for an 

 unselected group of 29S men. As it later appeared the group was one of average ability. 



2. At Camp Lee (first week in March, 1918), 479 men of a literacy-alpha group (all who 

 had made below 100 weighted score, on examination alpha) were given 10 or 12 tests of the 

 beta trial series. 



3. At the training school, Vineland, N. J. (February, 1918), 77 feeble-minded males whose 

 mental ages ranged from 6 to 11 years were given the beta trial series. 



In addition 35 men were tested at Camp Meigs, but as the data from this source played 

 little part in the revision they are not set forth here. 



Practically all these examinations were given by Yoakum and Brigham or under their 

 direction. 



The procedure followed in giving and scoring the tests at Meade was essentially that 

 described in section 3, above. The following experimental variations were introduced at Lee: 

 For test 6 (cube analysis) the time was reduced from 4 to 3 minutes ; for test 8 (picture secjuence) , 

 from 4 to 3 minutes; for test 9 (digit symbol), from 3 to 2 minutes; test 10 (spot pattern) 

 was shortened to 10 items; in test 15 (designs) the models were exposed in pairs. Tests 4 (dot 

 imitation), 7 (letter line), and 11 (analogies), were omitted for all Camp Lee groups, and tests 2 

 (form recognition) and 13 (X-0 series) for most of them. With one Vineland group test 11 

 (analogies) was given by using the third picture of each analogy as a stimulus for free asso- 

 ciation with a response picture, without regard to the first two terms of the analogy. 



The method of demonstration by pantomime proved to be feasible. Most of the tasks were 

 clearly indicated. As arranged in the preliminary form, however, test 11 (analogies) was 

 responded to as though it were a free association test, with the third term as stimulus. Even 

 so the test showed 30.7 per cent of zero scores in the unselected group. Test 2 (form recog- 

 nition) gave 19.3 per cent zero scores in the unselected group. Test 7 (letter line), though ex- 

 tremely easy for one who got the idea, nevertheless gave 14.9 per cent zero scores in the un- 

 selected group. Test 8 (picture arrangement) was difficult to demonstrate, giving 11.1 per cent 

 zero scores in the unselected group. It was decided, partly also on other grounds, to drop these 

 four tests. Of those retained, test 12 (geometrical construction) was the most difficult to 

 demonstrate. The introduction in the revised form of a simpler sample effected some 

 improvement. 



During the study several different total scores were made up. Five tests were discarded 

 before the first totals were added and were never included in any beta totals. These were 

 tests 2 (form recognition), 4 (dot imitation), 7 (line comparison), 8 (picture sequence), and 11 

 (analogies). The total score first used was the sum of the raw scores in the other 10 tests, 

 that of test 9 (digit-symbol) being divided by 5 and that of test 12 (geometrical construction) 

 multiplied by 2. This gave a maximum score of 185. (It should be borne in mind, in looking 

 at the correlations of beta tests with beta total, that the five tests omitted played no part in 

 determining that total. The distributions of total scores for the three groups of subjects were 

 characteristically different.) 



The Pearson correlations of each test with alpha total and with beta total are given in 



table 44: Tablk 44. 



