mance. Note any hesitation or inability to grasp the instructions.
Note especially any absurd procedure, such as placing all the blocks
without hefting any of them, comparing only some of them, holding them
up and shaking them, hefting two at once in the same hand, etc. The
ideal method, of course, is to try all the blocks carefully before
placing any of them, then to make a tentative arrangement, and finally,
to correct this tentative arrangement by means of individual
comparisons. A slight departure from this method does not always bring
failure, but it renders success less probable. As a rule it is only the
very intelligent children of 10 years who think to test out their first
arrangement by making a final and additional trial of each block in
turn. Contrary to what might be supposed, success is slightly favored by
hefting the blocks successively with one hand rather than by taking one
in each hand for simultaneous comparison, but as the child cannot be
expected to know this, we must regard the two methods as equally
logical.

The test of arranging weights has met universal praise. Its special
advantage is that it tests the subject's intelligence in the
manipulation of _things_ rather than his capacity for dealing with
_abstractions_. It tests his ability to do something rather than his
ability to express himself in language. It throws light upon certain
factors of motor adaptation and practical judgment which play a great
part in the everyday life of the average human being. It depends as
little upon school, perhaps, as any other test of the scale, and it is
readily usable with children of all nations without danger of being
materially altered in translation Moreover, it is always an interesting
test for the child. Bobertag goes so far as to say that any 8- or 9-year
child who passes this test cannot possibly be feeble-minded. This may be
true; but the converse is hardly the case; that is, the failure of older
children is by no means certain proof of mental retardation. The same
observation, however, applies equally well to many other of the Binet
tests, some of which correlate more closely with true mental age than
this one. A rather considerable fraction of normal 12-year-olds fail on
it, and it is in fact somewhat less dependable than certain other tests
if we wish to differentiate between 9-year and 11-year intelligence. But
it is a test we could ill afford to eliminate.[63]

[63] Compare with V, 1.

IX, 3