62 THE FACTORS OF THE MIND

equation (or its equivalent) for estimating success at the
work from the tests, without the mediation of any hypo-
thetical factors. This, indeed, is the procedure that I
myself have always advocated wherever conditions permit.
Since estimation inevitably involves an error at every stage,
the more stages we introduce into the total chain of estima-
tions, the greater will be the cumulative error in the final
prediction. Hence, as in educational selection, so in voca-
tional selection, when dealing with any specific case, the
empirically ascertained regressions will always be more
trustworthy than indirectly reconstructed inferences based
on hypothetical factors.

But once again the analytic or factorial approach is not
without its value. One great advantage, as it seems to me,
of analysing out the supposed factors and giving them names
is that the isolation and the naming force us to see that our
choice of tests covers a sufficiently wide range. An empirical
selection is not always so mechanical as it professes to be :
unconscious preferences tend often to narrow its scope,
But this is simply to echo the maxim that I have so often
emphasized in educational work—namely, that the psycho-
logical investigator with his tests and observations must be
sure that he covers all the different aspects both of the child
and of the child's task. Or, to reword it in statistical
terms, he must see that the various qualities that he pro-
poses to measure, although each of them is highly correlated
with efficiency at the job, are not themselves correlated one
with another (except, of course, so far as some small degree
of correlation can scarcely be avoided).1

But the factorial approach has a further advantage in
vocational work, which though present is not so striking in
the field of education. It makes for economy of thought,
and, what is still more important in practice, for economy
of labour. To devise tests and to undertake researches for
every conceivable vocation and for every conceivable group
is not a feasible proposal. Hence the psychologist hopes
that both the initial test-results and the ultimate vocational
performances may always be reducible to terms of the same

1 See * Some Principles of Vocational Guidance/ Brit. J. Psych., XIV,
1924, pp. 344 et seq. (Cf. also Thomson [132], pp» 114 et seq+)