随莉莉1979Interpreting the Measurement of Change anizatinal Research.pdf

PERSONNEL PSYCHOLOGY 1979, 32 INTERPRETING THE MEASUREMENT OF CHANGE ANIZATIONAL RESEARCH' ACHILLES A. ARMENAKIS Auburn Technical Assistance Center Auburn University ROBERT W. ZMUD Associate Professor of Information Systems ia State University Numerous advances have been made in the last two decades in the measurement as well as in the conceptuaUzation of change. Go- lembiewski, Billingsley, and Yeager (1976) conceptualized change as: (1) gamma change—^subjects change their understanding of the variable being measured; (2) beta change—subjects recalibrate the measurement scale; and, (3) alpha change—change detected with a consistent measurement scale (., no beta change) and for which gamma change has been ruled out. The present research examined these types of change by administering the Survey anizations questionnaire to a sample of military trainers at Timei (iV, = 222) and Timei iN2 = 242) with no intervention. This design may be re- ferred to as two premeasures of an abbreviated time series design. The results are discussed within the context of the alpha, beta, gamma change typology. After ruling out gamma change, the results revealed that alpha and beta change could be distinguished from each other and that beta change had been measured. The potential causes of the detected beta change (., the rival hypotheses affect- ing internal validity) are analyzed. The implications of these results are discussed for the present research as well as for other empirical investigations anizational change. GOLEMBIEWSKI, Billingsley and Yeager (1976) made a substantial conceptual advancement in evaluative research by defining alpha, beta and gamma change.^ The measurement of change was shown to ' The authors vrould like to express their sincere appreciation to Professors Hubert S. Feild, Jr., and Arthur G. Bedeian for ments in revising an earlier version of this manuscript. ^ The alpha, beta, and gamma labels for the types of change are preferred to other labels and wil