Project Background

In early 2010, the U.S. Supreme Court paved the way for unlimited corporate and union investment in federal election campaigns with its decision in Citizens United v. FEC. Many politicians, political pundits and scholars fear that the decision will open the floodgates to millions of dollars from corporate and union treasuries on 30-second attack ads. Although special interests with ample funds can purchase proprietary information on campaign advertising, transparency and government responsiveness depend critically on public information. The Wesleyan Media Project seeks to enhance transparency in government by providing public information on how special interests are attempting to influence American democracy in general and political campaigns in particular. Through the use of unique technology provided by Kantar Media’s Campaign Media Analysis Group (CMAG), we track and analyze all broadcast advertisements aired by or on behalf of federal and state election candidates in every media market in the country. Throughout the course of the 2016 election cycle, we provided real-time, public information on the extent of corporate and union spending in federal election campaigns across the country, who specifically is doing that spending and which candidates are benefiting.

The long-term goal of the Wesleyan Media Project is to develop a definitive database that tracks all advertising by source (corporation, union, interest group, party, or candidate) in federal election campaigns, including the presidential nomination race and general election, as well as U.S. House and Senate races. Gathering such data, and making them widely available, is critical for understanding the impacts of recent changes in the campaign environment both in terms of the players involved and in terms of what ads people see on television.

The source for the ad data is Kantar Media/CMAG.  This commercial firm specializes in providing detailed, real-time tracking information to corporate and political clients.  These tracking data represent the most comprehensive and systematic collection on the content and targeting of political advertisements.  The data include two types of information. First, frequency information tells when and where ads aired.  It contains precise and detailed information on the date, time, market, station, and television show of each airing.

Second, the data provide information about each ad’s content in the form of a video file for each unique creative or individual ad.

CMAG gathers such data by using a market-based tracking system, deploying “Ad Detectors” in each media market in the U.S. In addition to all local advertising activity, these detectors track advertisements on the major national networks, as well as national cable networks. The system’s software recognizes the electronic seams between programming and advertising and identifies the “digital fingerprints” of specific advertisements. When the system does not recognize the fingerprints of a particular spot, the advertisement is captured and downloaded. Thereafter, the system automatically recognizes and logs that particular commercial wherever and whenever it airs. Studies that examine advertisers’ “buy sheets” obtained from television stations and compare them with this CMAG data find that the company’s system is highly reliable in tracking the universe of ads aired.

After receiving the data from CMAG, the Wesleyan Media Project processes and codes the ad tracking data from all 210 media markets in the United States. In this process, using videos of ads captured by CMAG, project staff first research the entity responsible for airing each political ad, distinguishing between those paid for by candidates, parties, and interest groups. Finally, the Wesleyan Media Project codes the content of each ad on an extensive battery of questions using a web-based analysis platform.

The extensive battery of coding includes: the tone of the ad, whether the ad uses magic words, whether the favored candidate appears narrating the ad, what endorsements (if any) are mentioned, music, gender of the narrator in voiceovers, and what issues are discussed among others. This extensive coding allows for the compilation of a massive database of the content of commercials and frequency of airing that can be used in a variety of ways by academics.