608 SECTIONAL TRANSACTIONS.— J. 



By comparing the relative scores of the two groups in ' a ' to ' c ' the efiEect of the 

 training ' b ' can be estimated. 



It is essential in such experiments that the conditions in tests ' a ' and ' c ' for 

 both groups be kept constant. An indispensable aspect of this constancy is the 

 optimal control of the subjects' motives. With this in view the following experiment 

 which may serve as an example was carried out. 



3. An experiment in transfer of manual dexterity. 



The general plan was as described above. The training operation ' b ' was a 

 simple one of manual dexterity in removing and replacing links on spindles. This 

 process gave a satisfactory practice curve. The subjects were trained in it for a 

 fortnight. Tests ' a ' and ' c ' were planned to be as similar as possible to training 

 ' b.' These were given to both trained and control groups of unemployed boys, 

 aged 16-17, who were paid for all tests and training alike on their improvement. By 

 this means the motives were controlled by financial incentives. 



4. Results and Conclusions. 



On the statistical treatment of the data obtained, by the method of comparison 

 of means of the respective groups, it was found that there was no direct evidence of 

 transfer. 



5. The more important experimental work and 



6. The possible significance of the conclusions were discussed. 



Monday, September 10, 



Mr. E. R. Clarke. — The more refined Analysis of Group Mental Tests. 



The earliest age at which Mental Tests (Binet) can be used has been determined 

 by the age at which speech is reasonably well established, and the age-assignment of 

 the easiest test is 2i years. Similarly, the use of Group Tests has been limited to 

 children able to conform with a printed paper of instructions, and the use of a pencil 

 for underlining and crossing out, and in the case of the N.I.I.P; No. 34 Group Tests, 

 we find norms of performance for the age of ten years and onwards ; such a situation 

 being clearly more difficult to adjust to than an individual oral questioning by a. 

 kindly if somewhat standardised adult. These limits set by the practical testing 

 situation have been considered sufficient, but many errors have been made in 

 diagnosis by the tests when used with ages near the lower age limits of the tests. 



A method of statistical analysis has been devised to study the characteristics of 

 the ability to answer the tests as it exists among children, and one feature studied 

 was the variability of this capacity in an unselected group all of the same age. There 

 are statistical methods of evaluating variability, and the ratio of it to the mean 

 attainment of the group (Coef. of variation) was found at each age, 3-14 years for 

 Binet Tests, and 9-18 years for Group Tests. For Binet Tests the ratio was found to 

 decrease during the period 3-6 years from -5 to -225, and then remain constant at 

 this value from 6-14. For Group Tests, the nine sub-groups were individually analysed, 

 and in this case the ratio decreased during the period 9-13 years, and then remained 

 constant at values whose mean was -22 for the period 13-18. The constant ratio is a 

 widely occurring characteristic of biological development, and may be compared to 

 the proportional lengthening out of the field in a long-distance race as the race proceeds. 

 Little heed would have been paid to the initial peculiarities as discovered with Binet 

 Tests, but when the same occurred at a later stage with the Group Tests in results 

 which were so much in accord otherwise as to suggest strongly that both were 

 examining the same mental function, the method of testing was suspected. One is 

 led to believe that superimposed on the ordinary variation of this capacity is another 

 factor of variation which in the case of the Binet Tests from 3-6 years, and of Group 

 Tests from 9-13 years, gradually decreases, and either settles down to a fixed amount 

 or vanishes for later ages. 



It is advanced that this extra factor is either an extra emotional reaction called 

 up by the strangeness of the situation to a young child, or is the result of the variation 

 of environmental influences at the initial stages of learning, and which are eliminated 

 as the child later settles down to his real place in the scholastic scheme. Not until 

 after this later stage, which in one case is after six years, and the other after thirteen 

 years, will the strangeness of the situation, or the lack of settlement of the speaking 



