Understanding reactions to advocacy advertising through sentiment analysis

This thesis introduces a new measure to evaluate the effectiveness of advocacy advertisements to generate audience sentiment. While more traditional ads focus primarily on products, there is an increasing trend toward ads that focus on advocacy or specific stances related to public policy and political issues. Although they can be topical and attention-getting, advocacy ads also carry inherent risks such as eliciting contention from those with opposite stances and potentially offending members of the existing consumer base. In short, the ads can generate negative, rather than positive audience sentiment. Of further interest related to this phenomenon is the amplification of the resulting sentiment via social media channels. In this study, I use content analysis and audience sentiment analysis, a big data technique of machine learning and natural language processing, to explore online responses to the implicit and explicit messages present in sampled advocacy ads. The study finds that advocacy ads can be a feasible way to communicate the socially responsible aspects of a corporation, but the strategy needs to be conducted properly to arouse positive, rather than negative, sentiment.