P-, Q-, AND R-TECHNIQUES 193

type, c low / ? with the c obstructed ' or repressed). To
illustrate his views Stephenson discusses a typical extravert
and a typical introvert in terms of these factors; and his
analysis of clinical cases is said to show " a correspondence
with clinical types when patients are examined in terms of
these factors and of their relations " (pp. 37, 4.3, cf. 51).
All these conclusions are based on correlations between
traits. Strangely enough, nothing whatever is said in any
of the papers about correlating persons, although in the
article a few pages back, describing that method, " cor-
relating persons instead of tests " was said to be the " essen-
tial tool to check the theories of type psychologists."

In a symposium on the same subject [91], all the con-
tributors seem to have assumed that R-technique was
entirely adequate. Hitherto, it is stated, the work of the
c Spearman school' has been chiefly confined to the " field
of individual differences " as distinct from " general or
* type' psychology." In the latter field " it is easy to
work to c types,' such as introvert or extravert, but difficult
to measure the degrees ; yet . . . factors of the Spearman
kind are the only means whereby these quantitative sug-
gestions can be justified and regularized" (p. 102, my
italics). Such factors, deduced by correlating tests or
traits, form the best * classificatory device' for use in
psychiatry: " the whole individual can be described in
their terms." Stephenson's first paper closes with a section
on * The Future of Factor Studies in Psychiatry' ; yet
even here there is not a word about the utility of correlating
persons.

Dr. Stephenson's conversion to the alternative procedure,
therefore, seems, not only sudden, but even startling in its
thoroughness. For in his next publication he gives an
analysis of clinical types among patients from the same
mental hospital, carried out this time by Q-technique, and
now assures us that this is the sole method for " isolating
types and measuring persons for their approximation to
such types " (the schizophrene and the manic-depressive
among abnormals, introversion and extraversion among the
normal). " It has never before been possible," we are
told, " to measure for types " ([98], pp. 357, 366). Surely,

13