Abstract
MacKenzie, Lutz, and Belch have enhanced our understanding of the mediating role of attitude toward the ad (AAd). The current study replicates and extends the structural equation tests of the four competing models they presented. Two independent datasets are used to examine the role of processing involvement. Consistent with the earlier findings, the dual mediation hypothesis model provides the “best” fit of the data in both experiments. However, the hypothesized causal path between brand cognitions and brand attitudes that emerges for each of the datasets conflicts with the earlier findings. Contrary to expectations, processing involvement does not produce substantial differences in the specification or strength of the causal paths.