The development of the Q-sort technique by Stephenson (1953) has been a valuable aid to personality research and has provided psychologists with a research tool for investigation of numerous problems as has been indicated by Wittenborn (1961). This technique consists of having an S sort cards containing various statements into a number of piles. In the sorting procedure Stephenson uses a forced distribution which is a flattened normal distribution for statistical convenience. For example, an S may have 80 cards containing self-descriptions ( 5 replications for each of 16 treatment conditions) and is instructed to sort the cards into 11 categories with the frequencies 2, 4, 6, 9, 12, 14, 12, 9 ,6 ,4 , and 2.2 The extreme categories represent Mosc and Least Like the S (or some similar designation) and the intermediate categories represent gradation berween the two extremes. For each card sorted into the extreme Mosc Like category a score of 10 is assigned; each card in the next category receives a 9; the next an 8; etc., to the extreme Least Like category for which each card is given a zero. The forced distribution sort has been criticized by a number of people. For example, Cronbach and Gleser (1954) maintained that the forced procedure deletes valuable information and involves conditional, rather than independent, probabilities. Thus some statistical tests of significance, such as the analysis of variance F tests, would be precluded. They argued for the use of unforced sorts. Furthermore, Cronbach and Gleser (1953) have suggested the use of an analytical technique which is more general than Stephenson's Q correlation method, uiz., D, a measure of distance berween persons. Although a number of criticisms have been made of the forced distribution, many investigators have continued to report on their use of it (e.g., Kogan, et d., 1958; Turner & Vanderlippe, 1958; Nahinsky, 1956; Engel, 1959). According to Stephenson, the Q sortings provide a basis for the use of variance analysis to deal with explanations or theories of concern. This type of analysis has been much less used than have correlation procedures. However, the former has been employed enough to warrant consideration. For example, one may be concerned with the theory of Jung, and the variance analysis might include three variables: Attitude (extrovert and introvert), Mechanism (con-
[1]
J. Gaito.
Non-Parametric Methods in Psychological Research
,
1959
.
[2]
M. Engel.
The stability of the self-concept in adolescence.
,
1959,
Journal of abnormal psychology.
[3]
S. Geisser,et al.
An Extension of Box's Results on the Use of the $F$ Distribution in Multivariate Analysis
,
1958
.
[4]
J. R. Wittenborn.
Contributions and current status of Q methodology.
,
1961,
Psychological bulletin.
[5]
L. Cronbach,et al.
Assessing similarity between profiles.
,
1953,
Psychological bulletin.
[6]
G. Box.
Some Theorems on Quadratic Forms Applied in the Study of Analysis of Variance Problems, II. Effects of Inequality of Variance and of Correlation Between Errors in the Two-Way Classification
,
1954
.
[7]
J. Gaito.
Repeated measurements designs and counterbalancing.
,
1961,
Psychological bulletin.
[8]
H. S. Ripley,et al.
Some methodological problems in the quantification of clinical assessment by Q array.
,
1957,
Journal of consulting psychology.