PSYCHOLOGICAL MEASUREMENTS. 635 



short time, without errors or erasures, hence his score is comparatively 

 high. Two other subjects, Fre and Tom, were faster, but they made a 

 number of errors, particularly the latter, and so reduced considerably 

 what might be termed their ''efficiency" result. Pec was slow, and 

 somewhat inaccurate, hence his combination score for points per min- 

 ute is small, being 3.6. 



Considerable improvement may be expected in successive per- 

 formances of this kind of task. A comparison of the low-diet averages 

 shown at the bottom of table 189 with the figures for September 29, 

 which have just been considered, reveals the fact that the improvement 

 was marked. The average for the 9 sessions, October 13 to February 

 2, inclusive, show total time results ranging from 384 with Pea, to 725 

 seconds with Pec, who was noted to be the slowest on September 29; 

 the average for the 10 men was 500 seconds, a reduction in time of 

 about 25 per cent from the first trial. The range for the total number 

 of points made in the test had become smaller, being 71 to 95. It 

 could hardly be expected that any subject would make absolutely per- 

 fect records throughout. Bro, who made 100 points on September 29, 

 did so only two other times out of the nine following experimental 

 sessions. The average number of points for the squad during the 

 reduction period was 83.2, a gain of about 10 per cent. The number of 

 points per minute ranged from 6.3 with Pec to 13.3 with Pea, with an 

 average of 11.1, representing a gain of approximately 60 per cent. 



Squad B (see table 190) in their first experiment with this test 

 (October 6) averaged a longer time to complete the tasks, requiring 

 860 seconds as compared to 686 seconds for Squad A. Their total 

 average score was, however, 10 points better than that of Squad A, 

 84.6 compared to 74.6 points. Their combined score in points per 

 minute, due to the longer time required to fill out the blank, was 0.9 

 lower, i. e., 6.11 as compared with 6.99 points for Squads B and A. 

 In the averages for the 5 normal sessions and for the 3 low-diet dates. 

 Squad B shows less efficiency than Squad A in the performance of 

 this test. 



The comparison in points per minute for the 10 men in each squad 

 and for the successive experiments is conveniently shown in figure 

 123. The chief characteristic of these curves is the rapid improvement ; 

 the irregularities are no larger than those commonly shown with such 

 measurements. Squad B presents a rapid and uninterrupted improve- 

 ment in the first four experiments; the depression on January 5 is 

 coincident with the return from the Christmas vacation and also con- 

 sistent with the findings in several of the other measurements for this 

 date. During the low-diet period, Squad B continued to make im- 

 provement but very slightly. The change from 7.0 of September 29 

 to 8.6 of October 13 with Squad A is found to be not quite so large as 

 the percentage change made by Squad B between their first and second 



