J.— PSYCHOLOGY. 183 



static and dynamic steadiness, and — to discover if the training in the 

 skilled action had affected more purely ' mental ' functions — tests in 

 mental arithmetic and tests involving the rapid and accurate cancellation 

 of specified letters in a page of print. 



This series of tests was given on three occasions : (1) before training, 

 (2) at the end of the first week, (3) at the end of the fortnight. They may 

 be called transfer tests, 1, 2, and 3. 



Identical tests were given, in the same order and at the expiration of 

 the same three periods, to twenty-eight similar subjects who meanwhile 

 received no training. These were the control group. 



Since the trained group contained thirty-two and the control group 

 twenty-eight subjects, statistical treatment is justifiable. In no instance 

 was the difference between the trained and the control group, with regard 

 to their improvement in transfer test 3 as compared with 1, of such a 

 magnitude as to exclude the possibility of its being due to chance factors. 

 In some results the brief practice afforded by the test itself was definitely 

 shown to have had more effect than the intensive training in an apparently 

 analogous performance. 



The experiment supports the view that in such conditions training in 

 a low-grade skill is specific rather than general. These manual habits 

 did not transfer. 



How may such a clear-cut result be explained ? The following con- 

 siderations may be suggested : 



Writers upon transfer of training'^'' who know the experimental 

 evidence believe that one of the chief agents of transfer is the formation 

 of a sentiment. In the present experiment there was no encouragement 

 to form a general sentiment about the acquisition of skill, which might 

 spread to other skills. 



The conditions were as unsentimental as might be. The workers 

 were never exhorted to do their best. The only encouragement was the 

 very real one of immediate personal gain. Conversely, slack work 

 automatically caused less pay. This was made known to the learner with 

 little delay. The personal influence of the experimenters was as little 

 and as unchanged as possible. The workers were paid, and highly paid, 

 to transfer. Yet demonstrable transfer did not occur. 



It may be urged that when practice in a skill has hardened it into a 

 ' habit-unit ' this latter becomes partially dissociated from the rest of the 

 personality. Examples might be given of the way in which low-grade 

 industrial skills require minimal attention. Transfer, therefore, might 

 not be expected between this almost ' insulated ' entity and the rest of 

 the personality. Hardening the skill into a series of habits may have 

 decreased the possibility of ' ordinary ' transfer. 



Since the test was given three times ; the subjects were not ' saturated 



•^0 Ballard, P. B. The Changing School, London, 1925. Fox, C, Educational 

 Psychology, Cambridge, 1927. Pear, T. H., Skill in Work and Play, Chapter V. Perrin, 

 F. A. C, and Klein, L. W., Psychology, London, 1927, pp. 280-286. Sandiford, P., 

 Educational Psychology, London, 1928, pp. 275-300. Thomson, G. H., Instinct, 

 Intelligence and Character, London, 1925. Thomdike, E. L., ' Mental Discipline in 

 High School Studies,' Jour, of Educational Psychology, XV., January and February 

 1924, pp. 1-22, 83-98. 



