90 Estimation of the General Intelligence of School Children 
of considerable intelligence may take more nearly his true place under an exami- 
nation test. Again, when the organisation of the school is such that each master 
has charge of a particular form only a small portion of the time, the most reliable 
estimate of intelligence is obtained after a boy has been several months in the 
school. It seems clear from these results that the teacher's appreciation of General 
Intelligence judged by a reasonably defined system of categories and with sufficient 
time for observation, if corrected for Class and Age gives at least a close idea 
of what a boy's Place in Form will be, and a very close measure of his examina- 
tional value. In the face of this, it is surely very unreasonable to assert, as a recent 
writer has done, that Place in Class is a much better measure of ability than the 
teacher's estimate of General Intelligence. If we use Examination Place and 
correct for Age and Form we reach a result which is in close agreement with the 
teacher's classification by General Intelligence. The present results were of course 
made in one school, by using a defined system of categories, and under the super- 
vision of the Head Master, but they do not differ substantially from those considered 
by Miss Jones, in which the categories were far more loosely defined, the teachers 
were rather more numerous, and the children were younger. In each investigation 
the teachers were left to their own judgment as to intelligence, although in the 
present case a rough average was taken of their estimates. The point is very 
much more important than it appears, because there is a growing tendency to 
question examination results as a measure of General Intelligence. No single 
psychological test or complex of tests is in the least likely to replace our present 
methods of judging general efficiency for public or other service. These are the 
examination test, and the testimonial of the teacher, be he primary schoolmaster 
or university professor. If the teacher's opinion be given honestly and thought- 
fully, as it usually is, there appears to be very little difference between the results 
that flow from it and from an examination classification. The real criticism of 
the examination test lies not so much in its giving results that diverge from an 
experienced teacher's estimate of General Intelligence, but in the fact that the 
stress of competitive examinations, upon success in which future prospects largely 
depend, may injure either the physical or mental capacity of the candidate. 
The present numbers are peculiar to a school in which there is considerable 
selection with age, only the boys of about average intelligence not prolonging their 
education ; thus Form and Age are quite sensibly correlated with Intelligence. 
The superficial observer might suggest that this condemned the system of 
intelligence categories, which should give an appreciation independent of age. 
But the partial correlation of Age and Intelligence for a constant Form in the 
first series is either '021 or — "209 ; for the second series the same partial corre- 
lation is either - 040 or — '223, either value showing that greater age has not led 
the teachers to attribute greater intelligence to the boy. It is reasonable to 
suppose that this negative correlation is due to selection within the form, i.e. 
even within the form there is a tail of elder boys with something less, and a tail 
of younger boys with something more, than the average intelligence 
