160 A. T. Welford 



a "test of concentrated attention" (Paeaud, 1955). The 

 tasks have in common that they involve fairly high-grade 

 intellectual activity, judgment or the use of verbal or other 

 symbols. With more straightforward sensory-motor tasks 

 differences of performance with educational and occupational 

 level are usually small (Paeaud, 1955). 



These findings may be linked with the result, often obtained 

 in experiments and mental tests, that the difference with age 

 between the best performers shows less decline than between 

 those whose performance is poorer. This result occurs with a 

 wide variety of tasks such as the Matrices test and word 

 meanings (Foulds and Raven, 1948); memory, both short- and 

 long-term (Gilbert, 1941); and sensory-motor tests (Paeaud, 

 1955). It is clearly another facet of the tendency noted earlier 

 for variation between individuals to increase with age. It 

 could in some cases be an artefact in the sense that the task 

 may not stretch the ablest individuals and may thus fail to 

 show differences at the highest levels of ability. This cannot 

 be the complete explanation, however, and there seems no 

 doubt that age declines are less among the best performers at 

 least in a relative if not always in an absolute sense. 



(b) Effects of occupational skills can sometimes completely 

 reverse a group age trend for a particular individual. For 

 example in an experiment by Szafran (1951) in which hand 

 actions had to be carried out blind, a subject aged 49 who 

 had considerable experience in dark rooms was able to do the 

 task far better than any other subject tested either younger or 

 older. The connection between occupational skill and experi- 

 mental performance in a case like this is clear and reasonably 

 easy to guard against when making up groups of subjects. 

 Aspects of skill and tendencies to action of occupational 

 origin do, however, appear sometimes to "work loose" from 

 their original settings and become generalized to performances 

 seemingly remote from their originals. For instance, in an 

 experiment where subjects were required to throw at a target 

 (Szafran and Welford, 1949) an apparent slowing of perform- 

 ance with age turned out to be due to the inclusion of a large 



