| Abstract: | This study tries to evaluate empirically the validity and generality of causal models based on alternative theories and the integrative frame‐work provided by Crowley and Hoyer (1994) to explain the underlying persuasive mechanisms of two‐sided messages in marketing. In addition to an empirical test of the models, the study theoretically discusses and empirically explores suggestions for model modifications. Applying meta‐analytic‐based causal estimation, the results show that the parsimonious model based on attribution theory provides high generality and affirms the prevalent use of the theory in previous studies. All other models could be meaningfully improved by integrating alternative processes of less‐effortful message elaboration as suggested by dual‐process theories. The results suggest that the impact of argument quality in two‐sided messages is in line with previous models assuming effortful message elaboration, while less‐effortful elaboration processes seem to rely on the number of arguments used in two‐sided messages. © 2007 Wiley Periodicals, Inc. |
