Estimating the narcotic effect: Choosing techniques that fit the problem.

This paper compares the Butler and Ehrenberg analysis of the narcotic effects of impasse procedures, presented in the preceding article, with the purposes, methods, and empirical results of the authors' earlier paper on the same subject. The authors use the differences in the two papers to argue that a need exists to achieve a better blending in industrial relations research of model building and testing, of quantitative and qualitative data, and of simple and complex statistical tests. The differences in results presented in the two papers are shown to reflect, in part, differences in the definition of the problems examined. These authors stress the importance of choosing statistical techniques that fit the theoretical and policy problems of interest to industrial relations researchers and practitioners, and the power gained from mixing qualitative and quantitative methods. (Abstract courtesy JSTOR.)