P-, Q-, AND R-TECHNIQUES 183

briefly ' Q-technique.? Q-techniqiie is defined as " the
factor study of persons as variables." Like P-technique,
it thus involves " a reversal of the usual roles of persons
and tests in factorial work." " Prof. Burt," he continues,
" has suggested one way of effecting this reversal : but it
leads to no new principles or premises in mental-test theory.
The function of a door is not changed by hanging it back
to front : what I want of the concept of correlating persons
is a new door, not a badly hung back-to-front door. . . .
If Q-technique is to be of any importance on its own account,
it must seek new fields of endeavour in psychometry"
([136], pp. 33-4). " P-technique," in short, "is still
R-technique ; but Q-technique lies poles apart from either."
There is much that is attractive in these various sug-
gestions ; but, apart from the new terms, are they quite so
novel as they sound ? As Davies and others have pointed out,
it is not at first sight easy to see where precisely Stephenson's
c technique ' departs from that of previous workers who had
based their factors on correlations between persons rather
than tests. Let us note the points of agreement first of all.
To begin with, he now accepts the use of standard measure
(or some such equivalent) as a means of removing incom-
parability of units ; and in his first experiment on corre-
lations between persons ([92], p. 21), he actually adopted
the standard scale and frequencies I myself had drawn up
([35], p. 49, column 6). In his c Introduction to Inverted
Factor Analysis' [98] he begins by reproducing my matrix
formulation of the problem and describing the methods of
standardization I had proposed : but then he introduces a
novel concept—that of ' significance/ which, as he subse-
quently claims, * affords an entirely new basis of quantifica-
tion.' In his later, formal algebraic proof he adopts my
own proposal that, for theoretical work, c unitary standard
measure ? should be adopted instead of ' ordinary standard
measure ' ([96], p. 197, eqs. (2) and (3) ; cf. [93], p. 272) :
but once again this leads him to draw a sharp and instructive
distinction between the alternative ways in which such
standardization may be applied. In spite of these new
suggestions, however, we seem so far to be in pretty close
agreement.