Haggerty: Studies in Arithmetic 



95 



months, this being more than one year older than the average 

 for the grade. The lower half shows that 74 per cent made a 

 gain, that the average gain was 30 per cent, and that the average 

 of the group was raised from 35.5 per cent to 54.8 per cent. 



The average dependability of the upper quartile in February 

 was 89.5 per cent and in June was 77.5 per cent. Eight and one- 

 half per cent made a gain, 62.5 per cent lost with an average 

 loss of 28 per cent. To the 28J per cent who made a gain we may 

 add the 9 per cent who made no change, since their scores were 

 100 per cent on both tests. This 37J per cent were benefited by 

 the drill, but what about the 62J per cent? To many of these 

 the loss was due to overspeeding. For example, such a score as 

 this: 



February: 7 attempts; 7 rights; dependability 100 per cent. 

 June: 8 attempts; 7 rights; dependability 87 per cent. 

 In some cases there was no change in number of attempts 

 but a loss of one in rights. As for example: 



February: 11 attempts; 10 rights; 90 per cent dependability. 

 June: 11 attempts; 9 rights; 82 per cent dependability. 

 These cases do not represent serious losses, but this can hardly 

 be said of such a case as this: 



February: 5 attempts; 5 rights; 100 per cent dependability. 

 June: 5 attempts; 1 right; 20 per cent dependability. 

 Here is a clear case of over-speeding: 



February: 11 attempts; 10 rights; 91 per cent dependability. 

 June: 15 attempts; 10 rights; 67 per cent dependability. 

 We admit frankly that we do not know how to account for the 

 losses, but have merely attempted to indicate some probable 

 reasons for them. 



