Changes in Human Performance with Age 153 



subject's performance: a single score of overall achievement 

 leaves out of account variations in the manner whereby this 

 achievement was attained. Thus, one subject may have been 

 deliberate and accurate, whereas another may have been 

 quicker in his actions but have wasted so much time making 

 errors that he has achieved no more than the one who was 

 slower. We need to measure various component actions of 

 the performance separately; ending up, ideally, with a double 

 analysis both of what the subject has done in terms of ac- 

 curacy, types of error, form, etc., of his actions, and also of 

 the time he has spent over each. 



Thirdly, all these scores need to be examined for variation 

 during the course of the performance: thus one subject may 

 start slowly and speed up while another maintains an even 

 tempo throughout, and yet another proceed by a series of 

 bursts of activity with intervening pauses. 



Many important studies have been done without entering 

 into this amount of detail, and indeed few have dealt with all 

 three types together, but significant differences have been 

 shown between age trends in closely similar tasks, in different 

 constituent actions of the same task and in the serial course of 

 performance, so that each type of detailed treatment should 

 be considered and only rejected after making sure that im- 

 portant data are not likely to be thrown away. 



The same considerations apply to industrial studies based 

 on production records. Figures for overall output can often 

 with advantage be supplemented by records of wasted 

 material and faults, and by figures for changes in all these with 

 variations in the precise nature of the job and over periods of 

 time. This is not to say that significant use of production 

 records cannot be made without this extra information, 

 but that there are indications that it is in such features as 

 continuity of activity rather than overall production that 

 early signs of change with age at industrial work are to be^ 

 found. S^<\ \ 





