426 THE FACTORS OF THE MIND

once made either by partialling out the general factor for
persons or by some equivalent device.

If a simple ranking is used both for the standard and for the
individual examinee, the following device provides a crude but
fairly appropriate way of eliminating the influence of the general
factor. Such empirical rankings, when averaged, usually show a
correlation of between -2 and -4 with the general factor for persons ;
and this in turn corresponds with a total rank-difference of about
30 out of a possible 60. Hence, for rapid assessments in the clinic,
I suggest determining the coefficient for extraversion by the formula

JV

r = i-------, instead of by the usual formula for correlation by

3°

rank-differences. This roughly allows for the general factor for
persons ; and, when gradings from inexperienced judges are alone
available, the result as a rule seems to come fairly near to the co-
efficient that would be obtained with a more accurate grading,
standardized in terms of the general population, and assigned by an
experienced judge on the basis of actually observed behaviour.

For psychological surveys and for theoretical research a composite
quantitative assessment is indispensable ; but for practical purposes
a simple synoptic chart or diagram, showing the original measure-
ments plotted in a fixed order on a conventional scale, is a speedier
and more informative device. Such graphs have been freely used
both by teachers and by school psychologists for reviewing and
card-indexing the educational attainments of individual pupils.1
Similar diagrams, too, sometimes called * mental profiles,' have been
employed in vocational guidance, occupational analysis, and, indeed,
most branches of individual psychology. And the same device
would be of equal utility in clinical work.

The essential principles are illustrated in Fig. I. The scale on
the right is for the saturations describing the types ; the scale on

1 The general method is illustrated more fully by the £ psychograms'
printed in the L.C.C. Report on the Distribution and Relations of Educational
Abilities (figs. 9, i, ii, iii, and iv: these show graphs for the commoner
educational £ types'—the c verbal' and * non-verbal' types, the ' non-
arithmeticaP type, etc.). There, as here, the traits are arranged in order
according to the secondary factors rather than the general. The unit is the
' mental year/ which for educational abilities is almost exactly equal to the
standard deviation during the middle of the school period ; here the standard
deviation has been employed.

I may add that the plotting of contour-graphs from correlations is a simple
and neglected technique that may often take the place of the slower and more
laborious computations that figure in factor-analysis. For example, besides