A model of compliance‐gaining message selection

Abstract

This paper examines the dimensionality of compliance‐gaining message use ratings. Data from three studies are reanalyzed. The reanalyses suggest that the multidimensional solutions presented in these studies are spurious. Instead, a unidimensional model is shown to fit the data. This model predicts a nonlinear regression of compliance‐gaining message use ratings onto total compliance‐gaining message use score. Nonlinearity causes the regression of one compliance‐gaining message use rating onto another to be nonlinear. The degree of nonlinearity is a function of the discrepancy between the mean ratings for the two messages. Nonlinearity affects the correlations between messages, so that messages with similar mean ratings correlate more highly with one another than with messages with dissimilar mean ratings. This result introduces variance into the correlation matrix, which in turn produces spurious multidimensional solutions. The implications of this result for research concerning the antecedents of compliance‐gaining message choices are discussed.