The premise of this essay is that partisan affiliation---which is really the only thing that voter data contains for profiling purposes---is a valuable, independent "data point." That's not intuitively obvious, and while this draft is formally open to the possibility that there's less to the claim than meets the eye, you don't really explore it. You cite Guess for the proposition that people read less one-sidedly than the Republic.com hypothesis predicted. But that's not really the point.

Precisely because we are a very politically polarized society, partisan affiliation is really quite predictable. Residence patterns and consumption patterns correlate with people's beliefs well enough to help one subdivide partisan "independents" pretty helpfully. After all, except in partisan primaries, one isn't trying to reach people who fill out registration forms the same way, but rather people who are receptive to the same messages.

So the great independent value of the particular data point "how is this person registered to vote in party primaries?" is not something to be assumed, but rather to be shown. If the State is making accessible something that existing commercially-available data allows one to predict with, say, 80% accuracy, who cares? And if the data point is itself really only a proxy for a set of attitudes that are signaled elsewhere in the usual personal profile data set, again, who cares?

In the end, we are left with a really peculiar constitutional irony: we are supposedly forced to prevent society at large from being able to develop and trade this particular information---which is available to government agencies and officials very widely, and is also possessed by political parties---lest it fall into the hands of the press, who will misuse it. Fond as I am of irony, drifting into this conclusion makes me wonder if I might have been off course at an earlier stage.

I think making this stronger, then, depends upon nailing down a premise that looks unsupported to me, and offering a constitutional defense of your sense of the press as a unique danger to free information if it knows what we all pretty much know about one another anyway.

Feeding Us Ourselves: How Using Voter Data to Target Readers Threatens Truth, Choice, and Journalism

-- By DanielleTomson - 22 Mar 2017

I. Introduction

If you are on Facebook or Google in the United States, you are being harvested. Your hobbies, opinions, friends, interests, location, and purchases are all being collected, dissected, and sold to advertisers. Based on this this identity information, companies feed you more content and products they predict you will want. They are feeding more of you to yourself.

Now, add this digital identity to your public voting record and party affiliation, all of which is available for commercial use--at least in select states. Media companies use this data. As a, Jew I don’t eat meat with milk together. It seems sacrilegious to me to consume political content that has been harvested and reconstituted from my own social and political data.

In a media landscape enhanced with voting data, sensational partisan pieces are the new clickbait distraction. Is the press free when advantage goes to partisan outlets, fueled with state-supplied voter data? I suggest the press is not free from state interference when the state offers private voter data for commercial press purposes. This betrays our privacy, but also damages our civil discourse by giving unfair advantage to the most emotionally and politically resonant media.

II. The Problem: Voting Data, Psychographic Targeting and Media

Campaigns have hired companies doing psychographic targeting of individuals based off of voter and social media data, but now those companies are selling their partisan-wares and voter files to media outlets. This is a problem.

Voter registration, party affiliation, street address, email, phone number, and voting record – basically all but who one votes for—are available publicly either free or for purchase in many states for campaign and commercial use. Data specialists on both sides have used this data for years; private company NGP VAN services Democrats and Republicans have the GOP Data Center.

Companies now sell a hot new product: psychographic targeting based on “Big Five” personality traits, using voter lists matched against social media data. The most boastful salesperson this election was Cambridge Analytica, the American spawn of a British psy-ops project. With conservative billionaire donor Robert Mercer as major shareholder, Cambridge Analytica claims to have “profiled the personality of every adult in the United States of America—220 million people,” using Facebook data and surveys. Using their psychographic targeting, they claimed they won Iowa for Ted Cruz. Trump hired them later. Now, Cambridge Analytica has walked back its claims of using Facebook and scholars like Dave Karpf have critiqued the extent to which their methods actually hyper-target individuals. Regardless, the dystopian idea remains that companies predict your political preferences using social media data and voter registration. Cambridge Analytica now has a reputation for dark magic as they sell their products to media companies.

Tech-illiterate media companies desperate for readership and advertising find the idea of voter-data-enhanced hyper targeting attractive, even if the capabilities are overblown. This sales pitch does not exist in the European Union, where data protection legislation prevents the commercial use of voter and party registration. In the United States, we face a different truth—our voting data will now feed into the black box of our identities, spitting out more viral, extreme, and engaging content to hook us. In this new media paradigm, explicitly partisan outlets are winners and objective journalists are losers.

III. The Consequences: Partisan Winners, Objective Losers

The press is not free if the state offers a tool, wrought with privacy issues, that privileges certain media outlets, creating unfair advantages to the most partisan sources. Hyper-partisan targeted “news” on social media has begun to undermine trust in objective journalism. This weakens the Fourth Estate necessary to democracy.

Though political organizations funded many 19th and early 20th century American newspapers, post-“Golden Age of Journalism” Americans have come to expect truth telling from media sources. Lack of media literacy makes many Americans vulnerable to “fake news.” They don’t know they are targeted, are unable to stop receiving the targeted news, or are too captivated by it to know the difference. Craig Silverman of Buzzfeed has shown that fake viral news out performs real news on Facebook.

The nature and severity of impact that hyper-targeted content has had on our media and political landscape is debatable. Some like Upworthy’s Eli Parisier and scholar Cass Sunstein have cried “Filter Bubble!” as our electorate becomes more polarized and vulnerable to fake news. Conversely, researcher Andy Guess has suggested that news consumption of most Americans remains moderate (with the exception of an influential minority of conservatives).

Debates of severity and capacity aside, partisan-enhanced targeting benefits partisan outlets, while damaging faith in and access to objective journalism. With state voter and party data available for commercial use, the state legally violates our privacy and essentially gives an unfair advantage to outlets with hyper-partisan viewership who could benefit from that data. The truth is non-partisan. Of course, making a First Amendment Right claim here is beyond the scope or expertise of this author. That said, this data could continue to be used to undermine the objectivity and trust of journalism -- an affront to the possibility of a free press so valued in our Constitution. The targeting will only get better unless we cut off its source.

IV. Possible Responses

To counter this, states should end the commercial use of voter files, preventing firms like Cambridge Analytica from profiting off of political polarization and damaging the strength of the Fourth Estate. Data privacy scholar Colin J. Bennett suggests that American commercial and even deregulated party use of data amounts to voter surveillance. Adopting European data privacy standards around voter data would eliminate this problem.

There might be another option for news organizations, perhaps in law, perhaps in marketing, to band together and make the case that releasing personally identifiable voter information to partisan news sources creates unfair and dangerous competition. Alas, this Communications Ph.D. knows little about the law.

• #Set ALLOWTOPICVIEW = TWikiAdminGroup, DanielleTomson
• #Set DENYTOPICVIEW = TWikiGuest

Note: TWiki has strict formatting rules for preference declarations. Make sure you preserve the three spaces, asterisk, and extra space at the beginning of these lines. If you wish to give access to any other users simply add them to the comma separated ALLOWTOPICVIEW list.