Isolating the Excellence Factor
We began our analysis of the quantitative data from the Excellence study by attempting to reduce as much of the data as possible into a single index of excellence in communication management. This was necessary because both the Excellence theory and the data gathered in the study are complex. The theory consists of relationships among variables from such subtheories as public relations roles, participation in strategic management, and models of communication. The causal relationships among these subtheories could begin at different points for different organizations. The subtheories were operationalized into some 1,700 questions on three questionnaires administered to one or more heads of communication, the CEO or other executive manager, and an average of 14 employees in each of the 327 organizations we studied.
We tried to reduce these variables to a single index after first combining a number of indicators of variables into indexes or by using factor analysis to produce broader variables for related variables. We then factor analyzed these indexes or factors to isolate a single factor of excellence, which we then used to calculate an overall index of excellence in public relations. Factor analysis looks for clusters of variables to which people respond with similar answers when they complete a questionnaire. In the case of communication Excellence, our theory suggested that most of the characteristics of excellence would cluster together, as would the characteristics of less excellent programs. The most excellent public relations departments should possess most of the characteristics of excellence. The least excellent organizations should possess few of the characteristics of excellence and most of the characteristics of less excellent public relations. The majority of organizations, which fall in the middle of a normal distribution of excellence, should have midrange scores for most characteristics of both excellent and less excellent public relations or should possess some but not all of both sets of characteristics.