no. 2.] PSYCHOLOGICAL EXAMINING IN THE UNITED STATES AKMY. 



349 



Devens literacy test. — In an attempt to develop a literacy test, the problem of which could 

 be readily grasped by men of low literacy and which would nevertheless measure a degree 

 of literacy over a wide range, another form of test was developed at Camp Devens. Four im- 

 proved forms of the test with the directions are given on pages 281 to 285. 



The significant words in the test questions were selected from the Ayres spelling scale, 

 Terman's vocabulary test and the vocabulary test in the first edition of the Examiners' Guide. 



Questions were arranged in blocks of seven. Each of the first four blocks contain words 

 appropriate for a lower grade, and the last two blocks words for high school and college, respec- 

 tively. It was believed that this arrangement might arouse interest, facilitate scoring, and 

 possibly permit segregation by simple inspection of a critical block. The original form of the 

 test was very similar to Form I. It was mimeographed in camp. Through the courtesy of 

 Prof. E. A. Shaw, of Tufts College, it was given in November, 1917, to 817 pupils in the grades 

 and high school at Somerville, Mass. Mr. L. B. Hoisington, of Cornell University, also con- 

 ducted the test with 99 college students at that place. The junior high school group was com- 

 posed of 39 eighth grade, 51 ninth grade, and 52 tenth grade pupils. The senior high school 

 group was composed of 50 sophomores, 49 juniors, and 50 seniors. The college group was com- 

 posed of 28 sophomores, 40 juniors, and 31 seniors in arts and science. The distribution of 

 scores by grades is shown in table 28. 



The differentiation between grades on the basis of scores is reasonably good. Only 17 per 

 cent of the so-called second-grade pupils scored zero, yet in reality, of course, they were first 

 graders who had just begun work in the second grade. These facts were sufficient evidence 

 of reliability to warrant further experimentation. Slight changes in the composition and order 

 of a few questions were made and the test was printed. The printed form was given to 947 

 pupils in the schools of Medford, Mass. The distribution is shown in table 29. The median 

 scores in the lower grades are now somewhat raised, due probably chiefly to greater legibility. 

 The percentage of zero scores for the second grade is thereby also reduced from 17 per cent to 

 about 4 per cent. 



Table 29. — Devens literacy, printed form — Medford School and Cornell University. 



In January, 1918, three additional forms of the test were devised. Determination of the 

 relative difficulty of the four forms was made by trials of each (in compensating chronological 

 order) with 835 pupils in the schools of Trenton, N. J. The tests were conducted under the 

 direction of Principal J. M. McCallie. Distributions for these forms are given in tables 30 to 

 33 and norms may be computed therefrom. 



