: ON VOCATIONAL TESTS. 279 
almost complete omission of any attempt to use objective tests for the estimate of 
temperament and character traits. 
This year inquiries have been circulated to laboratory departments of Psychology 
and to individual firms in order to obtain information concerning tests for vocational 
_ fitness and to ascertain what research is being carried out in this direction. 
Altogether twenty-five replies have been received from these sources; a list is 
appended (Appendix A). In these replies there is a conspicuous absence of tests 
which attempt to obtain a measure of temperament and character. 
One noticeable feature of the work that is being done is the fact that a systematic 
follow-up of the results of the tests is being made in a considerable number of cases. 
The deviations from the diagnosis of the selection tests and subsequent performance 
in factory or workshop are frequently attributed on good evidence to the interference 
of temperamental factors, and this fact again emphasises the need for research in this 
direction. 
As more systematic following-up is being carried out the need for reliable criteria 
of success has become very noticeable. The extent to which following-up has been 
done and the value of the methods of comparison that have been used varies greatly. 
_ In some cases the groups of subjects who have been tested have subsequently been 
: under the observation of several supervisors or foremen, and have not been in 
_ sufficiently large groups to admit of a statistical evaluation of the success of the 
selection tests. A crude comparison with other groups of workers who have not 
previously been selected by tests is the only possible means of assessing the practical 
value of the tests. In such cases the evidence is dependent on the subjective estimate 
of the supervisor as to whether on the whole the performance of the workers is better 
than those not specially selected by tests. At the other extreme it has been possible 
to keep accurate records of the work of the persons selected, as shown by their progress 
in the training school and in the workshop. Where this has been done the evidence 
is certainly in favour of the efficacy of selection tests. Notable instances of this are 
seen in the accounts of tests in the firms of Carl Zeiss, Jena ; Bergmann-Elektricitiats- 
Werke, Berlin; Philips’ Glowlampworks, Eindhoven; The Vitkovice Mines, Steel 
and Iron Works Corporation, Vitkovice, Czechoslovakia ; Linke-Hofmann-Werke, 
Breslau; A. Borsig, Berlin; Fried. Krupp, Essen; Osram, Berlin; Société des 
Transports en commun de la Région Parisienne, Paris; Ganz & Comp.-Danubius, 
Budapest ; Fried. Krupp, Kiel; Verband Berliner Metall-Industrieller, Berlin; and 
in the follow-up of results of tests for motormen by Viteles in Philadelphia. 
Various criteria are used to estimate the value of the tests. The assessment of 
teachers in the training school and foremen or supervisors in the workshop is perhaps 
the most obvious method, though it is sometimes difficult to obtain reliable and 
impartial ranking as to performance from those who are in close contact with the 
workers owing to the intrusion of personal factors which in some cases markedly and 
admittedly modify judgment. 
Ranking a group by piece-rate earnings is not entirely satisfactory, as it seldom 
happens that the distribution of work is uniform, and there is often considerable 
inequality in the allocation of piece-rates. Variation in the supply of materials and 
in the tools used affects individual output and also introduces errors that cannot be 
satisfactorily adjusted. When records are kept for a number of years, trade fluctua- 
tions make accurate comparison of work done in one period with that done in another 
extremely difficult. Attempts are also being made to use statistics of labour turnover 
as a measure of the success of selection tests. A reduction in labour turnover is 
assumed as an indication of better selection, but this measure is subject to variations 
due to changes in general economic conditions, so that comparison between different 
periods is often unreliable. Where, for instance, there is a general improvement in 
trade, the tendency for the worker to take advantage of this may show itself in an 
increased labour turnover. 
Records of accidents and the amount of spoiled work produced are in some cases 
used as a measure of the efficiency of selection, but here again the issue is often not 
clear. Thus the members of a group of motormen selected by the help of suitable 
_ tests will inevitably be operating on different routes where the accident liability varies 
considerably. In many accidents, too, it is often extremely difficult to determine 
which of the parties involved was responsible. 
Some of the attempts that have been made to prove the value of selection tests 
is by the usual statistical methods are, it is feared, open to grave objections. The 
vay 
— iatiiiad i head 
PD Ee A ++ 
