Targeting Political Advertising on Television

Dec 8 '10

As we take a step back from the recent midterm elections, it is a good time to consider some of the questions raised by campaigns and elections that we might not have time to think about during the campaign itself (when we are largely preoccupied with who is going to win and who is going to lose). One such question is how campaigns decide how to allocate advertising dollars and whether or not they do so efficiently. Professors “Mitchell Lovett”:–research/faculty-directory/full-time-faculty-directory/mitchell-j-lovett/index.aspx and “Michael Peress”: of the University of Rochester have an interesting new paper (“ungated”: on exactly this topic. I asked Peress to summarize the research for the readers of the The Monkey Cage. Here is what he wrote:

bq. Television offers political candidates the opportunity to strategically target viewers to generate the greatest impact on election outcomes. Using a new data set on voting and viewing behaviors, we find that the primary effect of advertising is to persuade voters to choose a particular candidate. Hence, effective political advertising targets audiences with many likely swing voters, but advertising effectiveness must be balanced by cost. While ads on the most effective shows are up to four times more effective than on the least effective shows, advertising during prime time can cost two to five times as much as advertising during early morning or daytime. These low cost day parts have many highly effective shows, offering excellent targeting opportunities. By and large, the candidates do in fact advertise on cheap shows with high turnout rates and many swing voters. However, we uncover instances where candidate strategies could be improved– ads run on Fox local morning news are far less effective than ads run on NBC local morning news because Fox local morning news viewers vote at relatively low rates. Political candidates do not seem to appreciate this subtlety and overspend on Fox when more cost-effective options, such as Sunday morning news shows and daytime cable news shows, are available.

The following figure summarizes candidate spending in 2004, with red “R”s representing Republican spending and blue “D”s representing Democratic spending by type of TV show according to the voting and turnout behavior of its audience. The size of the letter is proportional to the amount of spending on the given show. I recommend you click on the figure to get the bigger version of it.


The “full paper is available here”: