Wesleyan Media Project Expands Into Health Policy Analysis

WMP logoThe Wesleyan Media Project, which for the past two federal election cycles has tracked and analyzed campaign television ad spending, is expanding into the realm of health policy analysis with a new study examining media coverage accompanying the Fall 2013 rollout of the Affordable Care Act’s health insurance marketplace.

The question of inquiry: How did media coverage of the ACA (commonly called “Obamacare”) differ state to state—or even within states—and what impact might this have on new health insurance enrollments? Findings were published July 18 in the Journal of Health Politics, Policy and Law from Duke University Press.

The new project grew out of the Govt 378 Advanced Topics in Media Analysis course taught by Assistant Professor of Government Erika Franklin Fowler, co-director of the Wesleyan Media Project, this past Fall. Students in the course were given a choice to investigate either elections-related media or health media. Initially, a small group of students in the course began a limited exploration of advertisements and local news coverage of the Affordable Care Act in 12 states during the October 2013 rollout of the health insurance marketplace. They noted some interesting geographic differences in media coverage, and decided to investigate further. The project was expanded to include all eight students in the course, who together developed and tested a coding sheet, which would be used to code the content of local television news coverage of the Affordable Care Act.

The effort went into overdrive during Fall break, when Fowler and Laura Baum (project manager with the Wesleyan Media Project and a lab manager in Govt 378) hired 40 more students to contribute to the project. The eight Govt 378 students, plus a veteran Wesleyan Media Project student worker, took the reins to train the new student workers in how to code TV news segments. The team collected clips from a commercial content provider, and used its system to conduct a search of the closed caption text for words related to the ACA.

To refine the overwhelming list of potential ACA stories that the coders worked on, Leonid Liu ’14 and Syed “Mansoor” Alam ’15 worked on data analysis preparation, writing syntax to create a priority score for each of 300,000 local news keyword hits to identify the top two local news broadcasts within the highest-rated half-hour of local news in each of 210 media markets. Alumnus Ross Petchler ’12 contributed textual analysis to further refine and prioritize this list. He searched the closed caption information on every news clip for certain ACA-related keywords to give each an “ACA score,” indicating the likelihood that the story actually was about the ACA. This allowed the researchers to narrow their sample and weed out any “false positives.”

One coder on the project, College of Social Studies major Michael Linden ’15, previously coded campaign ads for the Wesleyan Media Project.

“We code the ads for a large range of different variables, from whether the ad is positive or negative to the types of emotional appeals the ad makes to the policy issues (or lack thereof) that it addresses,” he explained. “It was really a neat experience to apply that same kind of analysis to local news coverage of the Affordable Care Act, and it was particularly exciting to get to help new coders learn the ropes on such an important project. Campaigns can often be ugly and make viewers cynical about the democratic process, particularly watching for hours on end. Whereas with the news coverage product, I felt like we had the opportunity to reveal relevant information about political communication, which was particularly cool considering how important the Affordable Care Act is to the current political landscape.”

In addition to studying local news coverage, the team also purchased data from Kantar Media/CMAG (the company which also provides data for the Wesleyan Media Project’s election season campaign ad analysis) on the frequency of health insurance exchange ads, as well as political ads that mentioned the ACA.

The resulting study, published in the Journal of Health Politics, Policy and Law from Duke University Press was co-authored by Fowler; Baum; alumna Sarah Gollust ’01, currently assistant professor at the University of Minnesota School of Public Health; Colleen Barry of Johns Hopkins Bloomberg School of Public Health; and Jeff Niederdeppe of Cornell University.

The authors knew that the local media is a central vehicle for the provision of information and that coverage would likely influence public attitudes toward the Affordable Care Act, as well as the number and characteristics of those who sign up for insurance through the new marketplace. That is, if fewer than expected people sign up for insurance in a particular location, it might not necessarily be evidence that available insurance plan options weren’t affordable, or that the public wasn’t interested in them. Rather, it could be that the dominant media messages—whether in news or advertisements—contributed to lower public enthusiasm toward insurance options.

The analysis found dramatic geographic differences in the volume and tone of local news coverage, insurance product ads and political ads in 210 local media markets during the two-week period of Oct. 1-17, 2013 following the launch of the health insurance marketplace. Specifically, media messaging was significantly more negative in states where the political climate was generally hostile toward the ACA—those states which chose not to operate their own insurance exchanges, but instead run federal or partnership exchanges. The disparity in volume of coverage was also striking. For example, TV audiences in Kansas City were exposed to more than 1,000 ad airings promoting health insurance options in the two-week period, while those in Nashville, with a similar population size, were exposed to only 50 airings. The researchers even found striking variation in the volume and tone of coverage even within states. For instance, only four news stories mentioning the ACA aired in Syracuse, N.Y. during the time period studied, compared to 14 news stories airing in Buffalo, N.Y. And while local television audiences in Tyler and Lubbock, Tex. saw similar amounts of ACA marketplace-focused coverage, three-quarters of that coverage in Tyler was discouraging, while 60 percent of Lubbock coverage was encouraging. Given these regional disparities, the authors urge researchers and policymakers evaluating the implementation of the ACA marketplaces to consider the role of media influences.

Linden was especially interested in the notion that in more conservative states, which chose not to create their own health insurance exchanges, “the resistance to forming a state-based exchange has become somewhat of a self-fulfilling prophecy: When you want policy to fail, it often will. In Connecticut, however, we see a very successful state exchange—the most successful in the country, in fact, because our leaders were willing to set up a state exchange and advertise it properly.”

“Increasing evidence suggests that where you live matters for health care delivery and outcomes. What we document in this piece is further evidence to suggest that where you live also matters for what sorts of messages average Americans are likely to receive about enrollment options,” said Fowler. “It’s not altogether surprising that coverage of the exchanges might vary with political environments, but we also document large variations within states. Of course, what we have not yet demonstrated is whether and how much such variation matters for public attitudes and enrollment uptake. That’s a challenge we’re fundraising to take on next.”

According to Fowler, the team is applying for funding to both extend this pilot analysis and to combine its media content measures with surveys to assess the actual influence of media messages on enrollment and public attitudes.

In addition to its expansion into the health policy arena, the Wesleyan Media Project is gearing up for the 2014 midterm elections and will be tracking mentions of the health care law in campaign ads throughout the year. The team plans to release a new analysis in the Fall of campaign ad spending in the 2014 elections, thus far.