PsychologxcalReports, 1990, 66, 845-846. O Psychological Reports 1990 COMPARISON OF SELF-REPORT AND PROJECTIVE MEASURES OF DEPENDENCY ' JOHN H. EHRENREICH State Univeniy of New York College at Old Wertbury Summary.-No correlation was found between the interpersonal dependency of subjects assessed by two different methods, a self-report measure and a measure of de- pendency based on TAT stories. Issues related to lack of consensual definition of "dependencyn and the differences in what projective and self-report tests measure are held accountable for the lack of agreement between the measures. Interpersonal dependency is a concept of considerable theoretical and clinical significance. Both for research purposes and for clinical assessments, a simple measure of dependency would be useful. Clinicians have frequently used the TATand the Rorschach to assess dependency (e.g., Holt, 1978; Schahr, 1948, 1954). Hirschfeld, et al. (1977) proposed a self-report "Measure of Interpersonal Dependency," which produces a total interpersonal dependency score and three separate subscale scores: Emotional Reliance on Another Person (described by Hirschfeld, et al. as reflecting a wish "for contact with and emotional support from specific other persons, as well as . . . a dread of loss of that person" and "a general wish for approval and attention from others"), Lack of Social Confidence (reflecting "wishes for help in decision-making, in social sit- uations, and in taking initiative"), and Assertion of Autonomy (reflecting a tendency "to deny either dependency or attachment" and a "preference . . . for being alone and for independent behavior"). Whether these different approaches measure the same thing is unknown, however. In the present study, subjects' scores on the Measure of Interpersonal Dependency were com- pared with an assessment of their dependency based on TAT stories. Forty-one subjects wrote stories in response to five TAT cards (Cards 1, 2, 3BM, 12M, and 13MF) and completed the Measure of Interpersonal Dependency. TAT stories were scored for pattern of dependency-whether figures in the story were described as (a) one dependent on another, (b) interdependent, (c) independent and autonomous, or (d) noninteractive. The inter- rater reliability coefficient (kappa; Cohen, 1960) between two raters scoring the stories was 0.82. For the subscales scores on the Measure of Interpersonal Dependency were: Emotional Reliance on Another Person: 43.7 (SD 8 . 3 , Lack of Social Confidence: 31.6 (SD 5.7), Assertion of Autonomy: 24.5 (SD 6.0). The Total Dependency Score (calculated, foUowing Hirschfeld, et a/., as the sum of three times the score on the first subscale plus the scores on the remaining two subscales) was 187.2 (SD 30.2). These scores were quite similar to those obtained by Hirschfeld, et al. There was no significant correlation between the dependency scores derived from the sub- iects' TAT stories and the scores the same subiects achieved on the Measure of Interoersonal Dependency. Pearson correlation coefficients between TAT scores and those on the subscales, Emotional Reliance on Another Person, Lack of Social Confidence, Assertion of Autonomy, and the Total Dependency Score were .116, ,238, -.057, and .135, respectively. The largest of these correlations, that between the TAT scores and those on the Lack of Social Confidence, was not statistically significant, as measured by a one-way analysis of variance (F,,,, = 2.34, p = .13). There are several likely explanations of the lack of relationship between the self-report and 'Address correspondence to John H. Ehrenreich, Ph.D., 19 Holiday Point Road, Sherman, CT 06784.