Simplicity, Complexity, and the Inkblot Puzzle

Abstract: Hans Zulliger’s three-card inkblot test was created in the 1940s to function as a short-form of Rorschach’s inkblot method. Despite having been continually published and used, the Zulliger has undergone less stringent research than the Rorschach. We administered both tests to 101 psychiatric outpatients using the Rorschach Performance Assessment System (R-PAS) as the basis for administration and coding, to investigate the Zulliger Test’s convergent validity and utility of assessment. In this study, we focused on Complexity and Simplicity, as these constructs represent a fundamental dimension of Rorschach data. We conducted a principal component analysis on 13 variables from both tests to compare their underlying structure. To assess linear correspondence and between-method agreement, we analyzed Complexity (as standard scores) and F%, along with two study-derived proportion variables: Fluency (R/per card) and Density (Complexity per response). We employed scatterplots with Loess smoothing, Pearson correlations, t tests, predictive performance metrics, and Bland–Altman analyses for these four variables. The underlying structure of R-PAS and Zulliger data showed convergence, confirming our hypothesis that data from the two tests are generated in a comparable manner. Further, we found that Zulliger provides unbiased measures of Complexity, F%, Fluency, and Density. However, Bland–Altman analyses revealed wide limits of agreement for Complexity and F%, indicating potential limitations in clinical utility at the individual level. Predictive performance metrics showed satisfactory specificity and negative predictive values (NPV), but sensitivity and positive predictive values (PPV) varied widely and were poor at some thresholds. Overall accuracy ranged from 71% to 85%, suggesting moderate to strong agreement between R-PAS and Zulliger scores.