Testing persuasive argument theory's predictor model: Alternative interactional accounts of group argument and influence

Abstract

Persuasive Arguments Theory (PAT) is a noninteractional theory of group decision‐making that predicts postdiscussion shifts (polarization/choice shifts) from the cognitive arguments individuals generate prior to discussion. PAT relegates interaction to a “display channel”; it is considered an efficient, but not necessary, mechanism for displaying members’ cognitive arguments in group discussion. The principal PAT predictor model incorporates four primary elements: (a) cognitive arguments, (b) cognitive argument novelty ratings, (c) cognitive argument persuasiveness ratings, and (d) a weighted‐averaging combinatory formula. This model is utilized to predict individual and group shifts produced in group discussion. Following a review, critique, and alternative interaction‐based explanation of the argument‐decision shift link, two models, each increasingly more interaction‐oriented, are advanced to test the PAT perspective on argument. Results of the tests revealed that although the initial model was not a significantly better predictor of post discussion shifts than the PAT model, the second elaborated model demonstrated superior predictive ability at some levels of analysis. Implications for PAT and for the role of communication in determining decision shifts are discussed.