Does Big Data Threaten Political Inequality?


Andrew Mayersohn in Boston Review:

Few Americans claim to follow “very closely” stories that are not above-the-fold headlines, even ones with a strong partisan valence and widespread coverage such as the Keystone XL pipeline (16 percent of Americans as of March, per Gallup) or last year’s IRS scandal (20 percent as of May 2013). That small population of news junkies likely overlaps the minority of Americans who are devotees of Fox News or MSNBC, further diminishing the number of citizens engaging in actual political dialogue. For better or worse, our politics are already fragmented.

The extent of the gap between the politically engaged and disengaged is what makes Anthony Fowler’s findings troubling. He and his coauthors report that get-out-the-vote operations “increase representational inequality” by bringing “more rich, white, educated, churchgoing citizens to the polls.”­ Knowing that their efforts are more likely to affect some than others, campaigns assign “propensity scores” to prospective voters in order to zero in on those who just need a nudge to vote.

This is where big data is most valuable. The Obama campaign’s major analytical accomplishment was to improve propensity scores by combining traditional voter rolls with consumer data and huge numbers of voter contacts, but even before Obama’s 2012 campaign, political operatives were getting much better at honing in on the best prospects among potential voters. For example, political scientist David Nickerson, who served as Director of Experiments for the Obama re-election campaign, has demonstrated that voter contact in Ohio was vastly more concentrated among high-propensity voters in 2008 and 2012 than in 2004—a triumph of intelligence-gathering from a campaign’s perspective, but one that reinforces political inequality. The better campaigns get at concentrating resources on prospective voters, the more they can focus on turning out their base and the less they need to worry about broad mobilization. Senate Democrats have apparently already adopted this strategy to some extent, according to Sasha Issenberg, who says that candidates’ strategy for this November is to “mobilize their way into contention, then persuade their way across the finish line.” In short, even if big data doesn’t inaugurate an era of personalized campaign messaging, it’s already fragmenting our democracy in another way by widening the gap between the engaged and the disengaged.

More here.