152 A. T. Welford 



verification from industrial investigations; these need guid- 

 ance from experimental results and can in turn give rise to 

 further experimental enquiry. 



(2) Problems arising from the Complexity of 

 the Human "Mechanism" 



When we study the behaviour of a whole human being we 

 are attempting, even more than in other biological studies, to 

 understand the working of a mechanism of extreme com- 

 plexity which for almost all practical purposes we cannot 

 directly observe. We have normally to be content to deal 

 with relationships between behaviour and external variables 

 impinging upon the organism. As a result, inference bulks 

 large in the building of hypotheses, and interpretations are 

 easily upset by new facts. Psychological theory thus tends 

 to be unstable, and, as older people often highlight certain 

 aspects of performance not normally considered by the general 

 run of human psychological work done almost exclusively on 

 young people, the student of psychological changes with age 

 frequently has to re-think parts of his theoretical background 

 to provide a setting for his findings. 



The most important result of this complexity, however, is 

 that investigations, whether experimental or industrial, have 

 to sort out the effects of the many different mechanisms, 

 sensory, central and motor, contributing to a single final 

 result. This means that in experiments and direct studies of 

 industrial work it is often not sufficient to take a simple 

 measure of performance such as amount achieved in a given 

 time at a single task. 



We need to expand our study in three ways: first, by 

 measuring performance under conditions which are systemat- 

 ically varied to limit one of the constituent mechanisms only, 

 varying, for example, either the display of information or the 

 responding action required while keeping the remaining 

 conditions the same. 



Secondlv, we need to take several different measures of a 



