Abstract: Television advertisements play an essential role in modern political campaigns with several billion dollars spent in the 2018 general election. For more than two decades, political scientists have studied TV ads by analyzing the hand-coded data from the Wisconsin Advertising Project (WAP) and its successor, the Wesleyan Media Project (WMP). Unfortunately, manually coding hundreds of variables, such as issue mentions, opponent appearance, and negativity, for many videos is a laborious and expensive process. We propose to automatically code political campaign advertisement videos. Applying state-of-the-art machine learning methods, we automatically extract various audio and image features from each video file. We show that our machine coding is at least as accurate as human coding for many variables of the WAP/WMP data sets. Since many candidates make their advertisement videos available on the Internet, automated coding can dramatically improve the efficiency and scope of campaign advertisement research. Joint work with June Hwang and Alex Tarr.
Kosuke Imai is a Professor in the Department of Government and the Department of Statistics at Harvard University.