Circumplex Structure of the Impact Message Inventory (IMI–C): An Empirical Test With the Dutch Version

Abstract

The Impact Message Inventory–Circumplex (IMI–C; Kiesler & Schmidt, 2006) is an interpersonal personality measure purporting to capture the command and relationship messages experienced by the IMI–C respondent in her or his interactions with a target person. Mixed support for the circumplex structure of the American IMI–C was found by Schmidt, Wagner, and Kiesler (1999). In this study, we sought to establish the circumplex structure of the Dutch IMI–C adaptation (Hafkenscheid, 2003) in a group of clinician raters assessing a total of 350 psychiatric patient targets. The Dutch IMI–C exhibited closer fit to circumplex models in this study than in the Schmidt et al. (1999) study of the English language IMI–C. However, using the most stringent and conservative circumplex criteria, the circumplex structure of the Dutch IMI–C could not unequivocally be confirmed. Our findings do not justify the generation of 8 separate octant scores when the IMI–C is used as an interpersonal measurement tool. Clinical use of the IMI–C for interpersonal assessment and diagnosis should probably be limited to the location of patient targets on the 2 main dimensions of Control and Affiliation, sacrificing the specificity of octant level interpersonal assessment.

Acknowledgments

This work was approved and funded by the Jewish Mental Health Services, Amersfoort, Amsterdam, The Netherlands (Director Cor de Bode, PhD). We are much indebted to the two anonymous reviewers and to the associate editor for their helpful comments on earlier versions of this article. Furthermore, we are grateful to Michael B. Gurtman (University of Wisconsin, Parkside) for generously providing an EXCEL worksheet enabling the calculation of the discrepancy functions between obtained and ideal angles presented in Table 6.

Notes

^a Means and standard deviations based on raw scores of the entire data set (N = 2 × 350 = 700)

^b N = 350; Raw scores = simple sums of item ratings for that octant. Mean-centered scores = octant scores based on the deviations of raw item scores from the item means for that particular clinician rater.