On Differences in Meaning between Normals and Neurotics

Abstract

In a recent study Luria (1959) attempted to investigate some differences between people in therapy and individuals who had never been in therapy. More specifically, Luria was interested in finding out if these two groups of individuals differed with respect to the meaning they assigned to certain concepts; the measuring instrument used in assessing meaning consisted of the semantic differential (Osgood, Suci, & Tannenbaum, 1957). While several groups were used for the therapy (T) and control (C) conditions, the major comlparisons were berween two matched groups of 25 Ss each (16 females and 9 males). The T group was selected from a pool of clients involved in therapy within a university setting. Luria does not indicate whether or not the T group was tested before or after therapy had begun, rhough the implication is that the testing was done early in therapy. Ss in both groups were asked to rate 15 concepts (MY SPOUSE, MY MOTHER, MY FATHER, ME, CHILD, MY DOCTOR or [MY THERAPIST or MY COUNSELOR], LOVE, MY JOB, MENTAL SICKNESS, PEACE OF MIND, FRAUD, SELF-CONTROL, HATRED, CONFUSION, and SEX) on the semantic differential. The semantic differential consisted of 10 scales, three for each of the evaluative, potency, and activity factors, as well as the additional scale relaxecdtense. In analyzing the results of these ratings, however, Luria discarded seven of the concepts and used only the findings obtained on the eight concepts of ME, MY MOTHER, MY FATHER, MENTAL SICKNESS, PEACE OF MIND, HATRED, CONFUSION, and SEX. The rationale glven for excluding the other seven concepts was based on a portion of the study which dealt with the test-retest reliability of the ratings. Luria states that the variability on these seven concepts was low, and that their inclusion have inflated the estimates of reliability. One might expect that low variability of ratings would have spuriously deflated any estimate of reliability, as such a restriction in range of scores typically results in a lower coefficient of correlation (McNemar, 1955). This, however, is not the important issue. What is important is that this prompted Luria to include o?zly rhe eight concepts listed above in the evaluation of differences between the T and C groups. The differences between the T and C conditions were analyzed by means of chi square. For each oB the three factors and for the relaxed-tense scale, 'The author would like to thank Melvin Zax for helpful comments on this paper.