Research Note: The Trouble With Studying Big Data in Campaigns

What are political campaigns doing with our data?  How would we know?

Sasha Issenberg, author of The Victory Lab, gave a talk at GW last night.  The book offers a strong take on the impact of the Analyst Institute on American political campaigning.  It traces the emergence of more sophisticated (and more widely available) voter data, and also traces the emergence of rigorous social scientific experiments that help campaigns optimize their outreach tactics.  It’s well worth your time.

During Q&A, an interesting tangent came up: political campaigns won’t talk with reporters about their data practices.  They didn’t want to give anything away that their opponents could use.  The Obama campaign told its staff not to talk to Issenberg.  When other reporters write articles about campaign data mining, the campaigns don’t offer corrections if they’ve gotten it wrong.  What little public record we have of these activities is based on reporters’ best guesses, without the usual corrective of sources shouting them down via the blogosphere.

This morning, one of those potential sources weighed in.  Ethan Roeder, data director of Obama for America, wrote an Op-Ed for the New York Times titled “I Am Not Big Brother.”  Pushing back against some of hype, he tells us, “You may chafe at how much the online world knows about you, but campaigns don’t know anything more about your online behavior than any retailer, news outlet or savvy blogger.”

The truth is probably somewhere between Roeder and the underinformed headlines.  It’s true that campaigns don’t know anything more about our online behavior than retailers like Target, but what those retailers know is pretty disturbing.  And c’mon, the Obama campaign operates at a scale and complexity far greater than any “savvy blogger.”  That scale matters for what questions the campaign is going to ask, and what it is going to do with our information.

As a researcher who studies how organizations adapt to the digital environment, the real trouble here is that it’s nearly impossible to move beyond vague impressions.  Campaigns have an incentive not to talk to reporters.  They have an even greater incentive not to talk to academic researchers (at least without a non-disclosure agreement firmly in hand…).  When the journalistic coverage gets basic facts wrong, scholars have little way of knowing.  When campaigners disagree after-the-fact, we can’t tell whether they’re correcting the public record or trying to smooth away rightful mistrust.

Academics at our best offer healthy skepticism to the public discourse.  There are important conversations for us to have about the implications of refined digital marketing, management, and persuasion techniques for a healthy democracy.  But it’s going to be systematically difficult to engage in those conversations, because the underlying facts just aren’t going to be very clear.

 

One thought on “Research Note: The Trouble With Studying Big Data in Campaigns

  1. I haven’t seen a healthy democracy in this country in some time.

    What we have seen in the recent past is the equilibrium position of electoral democracy — the only thing that matters is winning. Obstruct, lie do whatever it takes. Thank you for demonstrating this so clearly Mr. McConnell and Mr. Romney.

    If people cared about losing their privacy they would do something about it. Vote in the last Facebook vote. Get do not track software. ETC.

Comments are closed.