Changes in Human Performance with Age 151 



Smith (1938), for instance, found in an experiment in which 

 subjects were required to assemble nuts and bolts, that the 

 lowering of performance between men in their thirties and 

 fifties was somewhat greater for an eight-hour spell than it 

 was over a period of half an hour. The second objection is 

 that if older people are slower at learning an unfamiliar task, 

 difficulty shown by them in performing an experiment might 

 disappear with long-continued practice, and the experiment 

 would thus give a falsely unfavourable impression of their 

 capacities. 



Both these objections tend to be exaggerated. What little 

 we know about fatigue in older people suggests that mental 

 fatigue effects may sometimes become smaller with age, at 

 least until the sixties (e.g. Botwinick and Shock, 1952). As 

 regards improvements with practice, these have been studied 

 in a few laboratory experiments and appear, in some cases at 

 least, to be in about the same proportion in all age groups, 

 suggesting that absolute differences of performance might be 

 reduced by long-continued practice but that relative differences 

 would remain about the same (see, e.g., Brown in Welford, 

 1951, p. 64). However, our present knowledge bearing upon 

 these problems is scanty and both would repay further study. 

 Meanwhile the point would seem to be in principle valid that 

 the observations made in brief experiments need to be checked 

 against long-practised performances to sort out continuing 

 differences of performance from changes of capacity to deal 

 with short-duration and unfamiliar tasks. 



These objections do not apply to studies of industrial 

 work. Observation and measurement of work in industry 

 cannot be so closely controlled as an experiment, but are 

 usually more so than other everyday activities and thus 

 provide tasks practised to an extent far beyond what is 

 possible in the most protracted experiments, yet capable of 

 quantitative study in terms of output and a number of other 

 measures. 



For the clearest results experimental and industrial studies 

 need to be closely integrated together. Experiments need 



