Negative Advertising Effects on Presidential Support Ratings during the 2012 Election: A Hierarchical Linear Modeling and Serial Dependency Study

In a literature where different research approaches generate inconsistent effects, we examine negative TV advertising effects on public support ratings of presidential candidates over time. Weekly national advertising and poll data from 2011 and 2012 were analyzed via (a) pooled time series analysis with growth curve modeling and (b) individual time serial dependency analysis with autoregressive integrated moving average. Study results support a social influence model derived from attribution theory, wherein negative TV ads did, in fact, influence poll results or public support ratings. In particular, spending on negative advertising increased the support for the sponsor of such ads. We also uncover a reciprocal effect of negative advertising between the competing candidates over time, such that spending on negative advertising attacking Mitt Romney predicted spending on advertising attacking Barack Obama in about 4 weeks at Lag 3 and Lag 4. Findings establishing the efficacy of negative advertising are d...