Reliability of an idiographic Q-sort measure of defense mechanisms

Despite the important insights the concept of defense mechanisms may offer to our understanding of human behavior, no standardized definitions of defense mechanisms have been universally accepted. Inconsistencies in the definition and conceptualization of defense mechanisms has limited the practical utility of research involving these constructs. In addition, lack of interrater reliability, use of anecdotal evidence, and reliance on self-reports has retarded their investigation. Conducting methodologically rigorous investigations with a psychometrically sound instrument is the first step in addressing some of the issues concerning defense mechanisms and their theoretical postulates. This study was conducted to determine the reliability associated with the Defense-Q, an observer-based Q-sort measure of defense mechanisms. Thirty participants who had undergone an interpersonally stressful interview (the Type A Structured Interview; Rosenman, 1978) were rated by 11 trained coders, both for their use of the 25 defense mechanisms and for their ego strength. Reliability was assessed using Cronbach's alpha. Individual defense mechanisms demonstrated reliability ranging from .28 (undoing) to .92 (humor), with an average reliability of .73. Coder reliability ranged from .63 to .76, with an average of .69. These results indicate that defense mechanisms can be reliably assessed by the Defense-Q. Reliability of the Defense-Q is compared to existing observational measures of defense mechanisms.