When Voters Are Reduced to Numbers
Data collection now allows campaigns to target voters like never before. Is that good for democracy?
Sometime this month President Obama’s counselor, John Podesta, will deliver a report to the President on “Big Data.” The scope of the report is vast: how Big Data affects “the way we live and work” and “the relationship between government and citizens,” among other things.
Thus the White House begins its systematic look at “Big Data,” the stunningly enormous aggregation of bits and bytes about individuals that has grown exponentially during the information age. Today, Big Data pervades our lives, affecting almost every single commercial and social interaction we have, as companies data mine their way to bigger and better business models.
In the months since Podesta launched the Big Data review—spurred in part by Edward Snowden’s revelations about the National Security Agency —a horde of industry, government, and public interest representatives have descended on the White House to have their say.
However, one group of Big Data users seem to have largely escaped attention. Yes, Google and Facebook have had their meeting with Obama.
But has Obama talked to Obama yet? Because he (and his campaign) have created one of the biggest data sets yet on the American voter. And how he, and indeed all politicians and political groups, use the data will have profound implications for the future of elections and our democracy.
In a recent piece in Talking Points Memo, civil rights leaders Wade Henderson and Rashad Robinson, warned that “without reform, high-tech discrimination will continue and grow in law enforcement, lending, employment, immigration, education, and other areas–in ways seen and unseen.”
Big Data is potentially, though not inevitably, a threat to civil rights because, without necessarily meaning to, it may both harden and accelerate pernicious trends. Are poorer neighborhoods being redlined when it comes to loans or insurance? Big Data makes redlining easier and faster. It is not only simpler to find poor risk areas and individuals, but it is also easier to justify the redlining—“The data made me do it.”
Big Data is collected and used by more than companies who want to micro-target advertisements or by the NSA tracking “suspicious” activity. It is collected by political parties and by politicians who use it to grease the wheels of the election system. And we know that the political machine is largely immune to self-regulation. After all, these are the people who exempted themselves from laws attempting to limit unsolicited phone calls and junk mail.
During the 2012 election, the Obama Campaign created a much-lauded and unprecedented data mining organization. (They recoiled at the thought of being reported on prior to the election or at having their work called data mining, fearing that either would turn off voters.) Their success has been chronicled breathlessly, after the fact, in Sasha Issenberg’s Victory Lab and by the New York Times Magazine.
Using the Obama website, which featured data-tracking cookies, and a vast database on voter activity, the Obama Campaign unleashed a remarkably effective targeted messaging drive on the American voter: the emails, the phone calls, the snail mail, the door knockers, the personal visits, the Facebook ads, the TV ads. Relentless. Exhausting. Effective.
The Romney Campaign’s data staff was about one-tenth the size of Obama’s. They were as relentless and exhausting. But were less effective and targeted. Its doubtful they’ll make the same mistake again.
On the one hand, there’s something gratifying about being studied and pursued by politicians with such acuity. On the other hand: ick.
During the election cycle, the Obama Campaign created a 100-point score for each voter. Four factors went into the score: likelihood of supporting Obama, likelihood of showing up to vote, and then likelihood that either of the first two could be altered.
Issenberg likens it to a political credit score, and those are never used for ill right?
The campaign matched those scores up with the vast amount of data it had on individual voters (or on voter segments in general) and focused like a laser. In the chronicling of the elections, it’s pretty rosy, idealistic even: they sought to change people’s minds, educate them, and get them to the polls. Indeed, Obama saw turnout rates higher than expected. A two percent uplift here or there, in this county or that, can make all the difference in winning an election.
But am I wrong to worry that the same data can be used for bad? We know campaigns or political organizations sometimes work to keep turnout down. Negative advertising and misleading pamphlets (wrong voting day, wrong poll addresses) are common in campaigns.
Could this data be used to harden and accelerate some of the more worrying trends in political participation in America? Turnout continues to drop. The income divide in voting persists, with voting turnout by the wealthy almost twenty percentage points higher than by the middle class. Latino turnout in 2012 was high, but it still lags behind other groups of voters. Same for young voters. I worry Big Data will be deployed to cultivate these trends depending on whether they benefit candidates.
Big Data, to be fair, is neither good nor bad. It all comes down to how we use it. It has great potential to help, to illuminate the gaps and deficiencies in our political system. The more we analyze and learn about election administration and voter motivation, the more we can work to improve the functioning of our democracy.
But the vast majority of money and with it the data collection and analysis of voting behavior goes to partisan campaigns today. In elections, each campaign has one goal: win. The overall health of the democracy is not at stake. Except it is.
The views expressed are the author's own and not necessarily those of the Brennan Center for Justice.
Follow the author's Tumblr page: Electoral Dysfunction