集団ロールシャッハ法により心理診断に用いられるスコアリング仮定の統計的分析

Abstract

In scoring a Rorschach protocol it is customary to classify each response with respect to several classes of factors, such as content, portion of ink blots interpreted, and ink blot characteristics determining the response. As a second step in scoring, the total number of responses falling in each category is ascertained.The device of categorizing the responses and then of finding the total number of responses in each category carries the following two assumptions.1. All the responses falling in a given category are similar in some behavioral respect.2. The psychological significance of responses falling in a given category is different in some respect from responses placed in other categories.In order to test these assumptions the purpose of the present study is to provide an answer to the following hypotheses. A) The tendency that the relationships between pairs of the responses differing from each other with respect to scoring category are less predominant than the relationships between pairs of the responses falling in the same scoring category, is sufficiently significant to support the implications of the two assumptions above mentioned, B) The degree to which Rorscrach responses are intercorrelated is a function of the number of scoring categories that the responses have in common. C) The response clusters revealed by the factor analysis of the intercorrelations among carefully selected sixteen responses are congruent with any current theory and practice concerning the Rorschach, particularly with those which are reflected in abstract scoring categories.The data for the present analysis were the responses of 120 Kagawa University undergraduates to the Rorschach Ranknig Test. The intercorrelations among selected sixteen responses were computed and submitted to a centroid factor.The findings obtained may be summarized as follows:1. The tendency for human movemet (M) responses to be associated with other human movement (M) responses was significantly stronger than the tendency for such responses to be associated with controlled color responses (FC) and (CF). But hypothesis A did not apply to other scoring categories.2. The evidence for hypothesis B met a criterion for statistical significance. But the trend was slight and there were numerous important exceptions.3. The data of the factor analysis suggested that response content with its implied associations and projections might play an important role in detemining the functional similarities and dissimilarities of certain responses, particularly those which involved movement.The scant support which Hypothesis A, B and C receive is considered to be the indication that the usual abstract scoring procedures have not so much value as we have expected in attempts to appraise the behavioral significance of Rorschach responses.