no. 2.] PSYCHOLOGICAL EXAMINING IN THE UNITED STATES ARMY. 303 



selection. Wells had also had experience in making up an information test; accordingly this 

 test also was assigned to him. The information test, in accordance with committee decision, 

 draws systematically from a number of fields of knowledge. 



The words used in the synonym-antonym test were all taken from a "vest-pocket" diction- 

 ary (Funk & Wagnalls), in order to guard against the inclusion of rare or technical terms. A 

 provisional list of about 500 pairs of words was arranged by Terman and subjected to the 

 criticism of other members of the committee. After the list had been considerably reduced by 

 the omission of unsuitable words, 10 sets of 40 each were drawn by lot for the 10 "forms." 



In making up the items of the analogies test, various types of relationship between the 

 first two words were systematically drawn upon, such as genus-species, part-whole, cause-effect, 

 object-attribute, opposites, etc. The three incorrect alternative responses for each item were 

 secured by listing and using the third term of the analogy as a stimulus word in a free association 

 test with three subjects. The three reaction words thus secured for each stimulus word were 

 used as the three alternative responses in the corresponding test item. This method, suggested 

 by Whipple, gave distinctive and definite character to the analogies test. To respond correctly 

 in the test as constituted, means to select one of several habitual associations as determined by 

 the particular type of relationship existing between the first two words. 



Mention has been made of the use of the principle of random selection in making up the 

 different "forms" of the tests. In 5 of the 10 tests it was also necessary to utilize the principle 

 of chance in arranging the items within a given test. For example, the succession of true and 

 false sentences in the disarranged sentence test was determined by tossing a penny. The dis- 

 tribution of synonyms and antonyms in the synonym-antonym test was decided in the same way. 

 In the information test, the practical judgment test, and the analogies test, the position of the 

 correct response in relation to the alternative responses was also determined by chance. The 

 necessity for these precautions is obvious. 



Effort was made to arrange the items in other teste in order of difficulty by inspection. In 

 the case of certain of the tests, the individual items were rated for difficulty by several members 

 of the committee. Later results, however, showed that the ratings are unreliable. 



The instructions for giving a test were in each case formulated first by the person or persons 

 who had been made responsible for collecting the items of that test. Every tentative formu- 

 lation was submitted to the entire group for criticism. The instructions for several of the tests 

 were recast by Whipple. The guiding principle in the wording of instructions was to make 

 them as direct and simple as possible, and to this end the directions for each test were several 

 times revised. 



The tests chosen appeared to satisfy fairly well the criteria which had been originally laid 

 down. The consensus of opinion in the committee on this point was as follows: 



1. Adaptability for group use. — Thoroughly met by all. 



2. Validity as a measure of intelligence. — A majority of the tests were known to correlate 

 highly with good measures of intelligence. Perhaps most doubt was felt with respect to memory 

 for digits, number comparison, and practical judgment. Memory for digits had usually given 

 only moderately high correlations with other measures, but it had the following advantages: 

 (1) It would lend variety to the scale; (2) it could be made up in any number of equally difficult 

 forms; (3) it would cover a fairly wide range; (4) it was perhaps the best memory span test 

 available for group use. 



3. Range covered. — Met fairly well by all. It was anticipated, however, that the analogies 

 test probably would not reach down as far as would be desirable and that the number compari- 

 son test would probably not measure well in the higher ranges. Opinion was divided on the 

 disarranged sentence and the practical judgment tests. 



4. Objectivity of scoring. — Met perfectly by all the tests. 



5. Rapidity of scoring. — Met satisfactorily by all. 



6. Unfavorableness to coaching. — Each test could readily be made up in many "forms." 



7. Unfavorableness to malingering. — Little difference among the tests. Malingering is 

 possible in all. Diagnosis of mental deficiency on group test alone would not be safe. 



