The Rorschach Texture Response: A Construct Validation Study Using Attachment Theory

Abstract

Using attachment theory, in this research, we explored the construct validity of the Rorschach (Exner, 1974) Texture (T) response as a measure of interpersonal closeness and contact. A total of 40 men and 39 women completed the Rorschach and 2 attachment inventories. Their romantic partners also completed an informant version of the attachment measures. Attachment styles were measured by factor scores involving both self-report and partner report. Results indicate that attachment theory, as a broad conceptual framework, is associated with T. Specifically, T = 1 is most closely associated with a secure attachment style, T > 1 with aspects of the preoccupied style, and T = 0 with aspects of the avoidant style and an absence of secure attachment. Needs for closeness and contact associated with T can be couched within an adult attachment theory, but in this study, we did not test for problematic aspects of insecure attachment. Gender is a complicating factor and deserves more study.

Acknowledgments

We thank Joseph Obegi and Lorraine Watson for their consultations on this article. We also thank Joseph McCullaugh, who assisted with the preparation of this article. We greatly appreciate the recommendations from the Journal of Personality Assessment review process, which helped to focus this article. It is based on a doctoral dissertation completed at the California School of Professional Psychology at Alliant, San Diego, in 1999. ²

Notes

¹Future research should address the issue of using the Sum of Shading as a moderator variable. We would have investigated this aspect of our data, but only 5 participants gave no Shading responses. Accordingly, there were too few for any meaningful analysis. Obviously, a very large sample would be required to investigate Sum Shading of zero as a moderator.

^a 5 couples are missing income information, N = 74.

²In the original dissertation study, Cassella (1999) examined psychopathology as a potential classification variable for further identifying T = 0 participants. As pointed out by the reviewers of this study, this sample reflects a normal population sample and thus does not include many individuals with appreciable psychopathology. Accordingly, any analyses utilizing measures of psychopathology would suffer from a restriction of range. Because of this flaw, the measure and hypothesis for psychopathology have been excluded from this research report.