Moving forward: RQ1. Who were the mediated voters?


I love blogging, but it’s obviously not the right medium for me. I’m going to try to rectify that with this series of short posts using the chaos that currently ensnares us to develop some research questions for 2017 and beyond. Some of them are strongly journalism-focused, others about campaign organization, some about information systems. I’m going to dig at the necessary research designs a bit, but I’m not thinking too much here about doability; this is more about what we should be figuring out.

The first one came to mind reading this New York Times piece on voters and non-voters, and primarily black ones, in Milwaukee County:

“We went to the beach,” said Maanaan Sabir, 38, owner of the Juice Kitchen, a brightly painted shop a few blocks down West North Avenue, using a metaphor to describe the emotion after Mr. Obama’s election. “And then eight years happened.”

All four barbers had voted for Mr. Obama. But only two could muster the enthusiasm to vote this time. And even then, it was a sort of protest. One wrote in Mrs. Clinton’s Democratic opponent, Senator Bernie Sanders of Vermont. The other wrote in himself.

This sort of voter profile piece is a staple of post-election reporting, particularly when a candidate under or overperforms in an unexpected way, and is presented as a way to understand the broader scope of what happened in the election. At the same time, this piece managed to find just one reluctant Clinton voter in a city that cast 76% of its votes for her (“…as did many others here” is how the story puts it). This genre existed during the campaign as well, as noted frequently by Eric Boehlert of Media Matters for America:

In general, I understand the media’s desire to try to explain what’s driving the support for Trump, who’s obviously running a highly unusual campaign and marketing his run in openly bigoted language. For a lot of people that’s deeply troubling, so understanding the dynamic behind Trump represents an obvious story of interest.

What I’m baffled by is the media’s corresponding lack of curiosity about examining Clinton voters. After all, she has accumulated more votes than any other candidate this year and is leading a Democratic surge into key states. (Why hasn’t The New Yorker published an 8,000-word piece on why Virginia has turned into a deeply blue state over the last decade?)

And I’m not alone in noting the year’s long-running disparity. Journalism professor and Clinton supporter Jeff Jarvis recently admonished the media (emphasis in original): “I never hear from voters like me who are enthusiastic supporters. I never see reporters wading among eager backers at Clinton rallies to ask them how much they like her and why.”

So the question here is a simple one: Do the voters and areas presented in mainstream press profiles represent the actual electorate that votes in the election? If not, in what ways is the presentation biased? Some of these are fairly obvious — the view of Trump as a weird insurgent, at best, or a danger to the republic, at worst, make for a Man Bites Dog story regardless of what else is going on. However, “Former First Lady becomes first ever woman to win major party nomination” is also an unusual story. Projected swing states seem likely to have gotten more attention, but the Democratic movement of Arizona and Georgia is also compelling.

This sort of study would’ve been much easier to do 20 years ago. Identifying what qualifies as the national political press in 2016 is a study on its own, and then figuring out how to find all the relevant profile pieces from, for example, CNN.com is another extensive piece of work. The difficulty of systematic sampling and the breadth of how profiles are presented suggests a qualitative approach may be the most sensible, but any comparison with real election results will need more precision than that. Assuming we find a way through the sampling process, the work is a little easier. Coding for candidate support, enthusiasm, location, history, available demographic information, and anything else that helps form a picture of the voters being profiled can be aggregated up to a model of what the mediated electorate looks like. It’s very possible there aren’t enough data points available to do a true statistical analysis, but I think that capturing the picture in the coverage is really the goal of a study like this.

Share this post:

    Tags: , ,