278 REPORTS ON THE STATE OF SCIENCE, ETC. 
groups, each of which practised a different operation. Nowhereis there any evidence of 
practice in one operation bringing about improvement in another. The combined results 
of the experiment suggest that this conclusion will hold over a wide range of operations. 
The relationship between ability, improvability, variability and intelligence have 
also been studied from the data collected in the previous year and from experiments 
carried out during this year. There is a general, and in some cases a marked inverse 
relationship between ‘initial’ ability and improvability both when the latter is 
measured as an absolute gain, and when measured as a percentage of initial ability. 
A similar, though less marked, relationship holds between total ability and these two 
measures of improvability adopted. The relation of improvability and variability is 
less uniform. Improvability, both ‘ absolute’ and ‘ percentage,’ tends to an inverse 
relation with intelligence. Comparison of each of these functions with initial ability 
suggests that this relation is due to the handicap to further progress imposed by the 
greater initial ability which accompanies superior intelligence, rather than to any 
hindering effect of the intelligence itself. 
Since the last report the mental operations involved in mechanical assembling 
and routine assembling have been analysed in great detail. In routine assembly 
work there appear to be in the main two types of activity : type I leads to a know- 
ledge of the general spatial character of the movements to be imparted to the 
material ; type II leads to knowledge concerning the way to bring about the requisite 
movements. These types have been further subdivided and analysed into their 
constituent mental processes. A detailed description of the whole research has been 
prepared for publication. It comprises 14 chapters together with an appendix of 
35 tables, 65 curves and other figures. 
In the last report suggestions were put forward (i) that the ‘ motor’ factor dis- 
covered in the data may be associated with the complexity of the operation and may 
tend to disappear when very simple movements are concerned ; and (ii) that although 
no transfer effect was observable when the subjects were left to their own devices, it 
might be possible to give general instructions or training, the effects of which would 
be transferable to other operations. An investigation into the first of these questions 
has been begun. Sixty boys constituting the two top classes of a London elementary 
school have been given two groups of tests (1) consisting of complex tests such as 
assembly and stripping containers, (2) simpler tests such as stripping screws, screwing 
up a turnbuckle, threading large beads, etc., and they have also been given a test of 
general intelligence. The correlation between the twelve tests given has been deter- 
mined. Time has not yet permitted of the statistical analysis necessary to determine 
how far the observable differences of these figures are indicative of different factors 
and how far they are due to chance fluctuations. The following indications are, 
however, given in the cumulative coefficients. There is little, if any, correlation 
between the motor tests and the general intelligence tests. The coefficients range 
from ‘09 to -25. There is usually a definite, and in some cases, a high positive inter- 
correlation between the motor tests, ranging from ‘07 to °75. Thus, further evidence 
is provided of at least one routine factor in these tests. The average intercorrelation 
of those tests classed as complex is ‘41, that of the simpler tests is 32, that of the simpler 
and the complex is ‘27. If the differences between these averages should prove of 
statistical significance, they would indicate the presence of a second factor (or factors) 
such as ‘speed,’ in the simpler tests, bringing about the observed closer agreement 
among themselves. And if, as seems more probable, the difference between *27 and 
-32 above should appear in the light of statistical criteria to be not significant, we must 
suppose that the intercorrelation of the simpler tests is brought about by the same 
factor as that which has already been observed in the more complex test, and that as 
they become simpler they become less saturated with this factor. Further work is 
contemplated in which still simpler tests are included and where further classification 
in the light of the subjective analysis which has since been completed may be 
employed. : 
\ Report For 1927-1928, 
TOGETHER WITH ADDITIONAL INFORMATION COLLECTED DURING 1930-1931. 
At the last meeting of Section J a report on intelligence tests was presented. This 
report brought out the need for more uniformity in the method of giving instructions, 
and emphasised the limited range of mental activity tested by the majority of the 
test groups in common use. One prominent feature in the replies received was the 
Cl 
a eee ol 
