158 A. T. Welford 



statement in terms of the proportion of subjects attaining or 

 exceeding a given level of performance. Fortunately for ease 

 of statement the use of the mean does seem to be justified in 

 most instances. 



Other implications of the role played in human performance 

 by past experience concern problems of sampling and differ 

 according to whether the studies are (a) experimental or 

 industrial; (b) cross-sectional, examining groups of different 

 ages in the population at one time, or longitudinal following a 

 single group over a substantial period of time; and (c) aimed at 

 obtaining theoretical insight, or at the solution of an immediate 

 practical problem. Each of these classifications is independent 

 of the others giving a somewhat unwieldy array of eight 

 types. The independence of (a) and (c) may at first sight be 

 surprising since experiments tend to be used in theoretical 

 research and industrial studies in practical problems. We 

 have already pointed out, however, that industrial studies can 

 contribute significantly to theoretical issues and there is a 

 growing tendency to use experiments as a means of furthering 

 practical research. 



The obvious difference between theoretical and practical 

 studies as far as work on ageing goes is that the former are 

 directed to an attempt to understand the nature of ageing, 

 whereas the latter are concerned with problems such as em- 

 ployment or social welfare to which the ageing process may be 

 an important but nevertheless ancillary consideration. The 

 main methodological difference between the two is concerned 

 with sampling for cross-sectional studies. 



Cross-sectional researches with a practical aim need to be 

 done with a strictly representative sample of the population 

 to whom the practical issues relate. They might thus require 

 a representative sample of the population of the country as a 

 whole, or of a particular town or factory or occupational 

 group. The sample should be of the present living population 

 in each case with no consideration of whether death, disease, 

 emigration, education or any other cause has had a selective 

 effect on the older age groups. Refinements to this basic 



