"Data-driven" Politics
Be wary.
If you’re a paying subscriber, in our Monday Political Brief, we pointed to a new NYT article on how political parties use data to target voters for ads — “Why Am I Seeing That Political Ad? Check Your ‘Trump Resistance’ Score.” Here is an illustrative excerpt on the matter:
Voter-profiling systems like the Covid-19 scores may be invisible to most people. But they provide a glimpse into a vast voter data-mining ecosystem in the United States involving dozens of political consulting, analytics, media, marketing and advertising software companies.
In the run-up to the midterm elections next month, campaigns are tapping a host of different scores and using them to create castes of their most desirable voters. There are “gun owner,” “pro-choice” and “Trump 2024” scores, which cover everyday politics. There are also voter rankings on hot-button issues — a “racial resentment” score, for example, and a “trans athletes should not participate” score. There’s even a “U.F.O.s distrust government” score.
Campaign and media consultants say such political-issue scores make it easier for candidates to surgically target messages to, and mobilize, the most receptive voters.
“We’re seeing not only U.S. congressional races, but State Senate races that are diving into this, and consultants using it to help them find those perfect targets,” said Paul Westcott, a marketing executive at L2, a leading voter database firm. He added that even some county campaigns were using scoring models to target voters on local ballot measures.
But the same nano-targeting that may help mobilize some people to vote could also disenfranchise others as well as exacerbate political polarization, political researchers say.
I’ve worked on two major political campaigns — in 2008 and 2012 — and so this laser-like precision in categorizing and sorting voters comes as no surprise to me, though the technology has clearly and significantly advanced over the last decade.
I wrote about the use of data in our politics in my first book, Reclaiming Hope. I also lamented the capacity of “data-driven decisionmaking” to distort and displace principled decisionmaking, and to reflect, misconstrue and reinforce voter impulses and preferences that should not be encouraged.
Data-driven politics is incompatible with aspirational politics. It is willing to sacrifice a broader coalition for a few bucks, a dozen hours of free airtime, and an angrier base. It is the type of politics that allows the NRA to have an outsized impact on our national debate about guns through targeted appeals that maximize their power. And it is the type of politics that leads Democrats to send specific women—women who subscribe to a particular magazine, for instance—mailings that warn Republicans will “drag women back” with anti-choice policies while telling national press that it is Republicans who are obsessed with reproductive issues.
When you’re led by data alone, you can make some pretty foolhardy decisions. To take one example, the campaign to reelect the forty-fourth president of the United States decided it would be good to include profanity in its digital outreach. One e-mail included the word damn in the subject header…
…I did not bring the issue up, that is, until I received an e-mail from a nationally known Christian leader. She liked the president, and I believed she was a supporter. She forwarded to me the campaign e-mail with “damn” in the subject line, along with a message that she would like to forward these e-mails to personal friends to encourage their support, but given their vulgarity, she obviously could not do so. I thanked her for letting me know, and I forwarded the e-mail to other staff.
Just to make this clear: a nationally known, supportive leader was telling us that she and others like her would be able to be more helpful to our campaign if we stopped sending out e-mails with vulgarity under the auspices of the reelection campaign for the sitting US president.
The response I received could have been delivered by Siri rather than a human possessing judgment and intuitive reasoning: I was informed that e-mails with profanity have an open rate higher on average than e-mails without profanity—by a matter of percentage points.
This kind of targeting leads to what the NYT article identifies, an earlier version of which I critique in Reclaiming Hope:
Politicians have always tailored their messages to reach certain kinds of voters. What is different today is that with a niche-focused news media and the tools of big data, political campaigns can privately deliver messages to individual voters that contradict messages sent to other voters. This is how two voters can go to the polls and vote for the same person for different, even opposing, reasons.
It seems even more prudish and futile to be concerned about the use of the word “damn” in an email now, but this is the point. Technological “advances” are used to distance oneself from and justify moral choices which affect our politics and voters in ways that open up occasions for further distancing from and justifying moral choices. Often, these choices themselves are framed as positive, only to lead to obviously negative consequences. How wonderful and democratic it would be to use data to go around “traditional gatekeepers” in media and reach voters directly? Wouldn’t it be great to have politicians who speak in the language and vernacular of everyday voters?
We have an opportunity in politics, as we do in all of life, to not simply reify the bad decisions we’ve already made, or the negative impulses we have not yet transformed, but to make decisions in the here and now that reflect our better impulses. Data is ubiquitous, and we should never seek to remove data, understood broadly, from political decisionmaking. But we should not allow data or technological capability make our decisions for us. Data is not synonymous with knowledge. We should not allow our ability to do something, to get away with something, to serve as the only permission we need to actually do it.
The political technologies that are developed today will emerge from the kind of politics we have today. We have a deeply troubled politics today. Be wary of what it naturally produces.
As someone who has "data-driven" in his LinkedIn profile, I co-sign a lot of what I see here as it relates to political campaigning. The issue I cut my teeth on, "climate change", has been polled in so many variants ("global warming", "climate crisis", etc.) that seasoned activists begin to sound like robots; the candidate campaigns I have chosen to work on have purposefully all been state-level races where the communications have not yet been hijacked by AI.
I also appreciate you mentioning that data is ubiquitous and something we shouldn't remove entirely. To offer something more surgical: a helpful distinction here would include the difference between using data in "politics" versus in "policy". Principled policy without data to back it up can become ideology-driven destruction.
Furthermore, although you don't speak about religious organizations in this specific piece: I do think that while a prophetic voice is necessary against abuses of data, we who work in this space also need to simultaneously demystify some of the fears religious groups have about data, as it possible to steward data in order to improve missional efficiency, accountability, and resilience. (As a fun trivia fact: the word "data", which derives from the Latin for given, was first extensively used in the early modern era to refer to "scriptural truths or revelations given by God".)
Hmmm, I don’t think the data-driven-ness is the real problem in each of the cases you cite.
1) In the example of profanity in a campaign email, the problem was which metric they chose to maximize: open rate. Any number of other metrics could have led in a different direction, and your concern could have just as well been described quantitatively as with the “forward rate.”
Generalizing beyond the specifics of email, profanity attracts attention, and the problem was not that they had quantified attention but that this is what the team chose to maximize.
2) Slicing and dicing the electorate, and tailoring different messages to different subgroups of the population, is honestly as old as democracy itself. Sure, fine-grained data can make that potentially easier, but the decision to target your message to a subgroup in the first place is the real issue.
To take another example from 2012, the lack of micro-targeting didn’t make it any less bad for Romney to talk disparagingly of the 47 percent of the population who receive government assistance in a closed-door fundraiser with rich folks. It was the same problem of talking out of both sides of his mouth.
Data is just a tool, the real issue is what we’re trying to do with it,