Allow me to offer a claim using the same logic used by Chuck Todd and Carrie Dann in their recent article “How Big Data Broke American Politics”:
Epicurious broke Americans’ diets. Its big data approach to food has allowed people to combine fats and starches in ways never before conceptualized, and the result is the rising and dangerous obesity rate. Just look at how obesity has risen since 1995, when Epicurious went online. Our cooks aren’t responding to balanced diets anymore because they’ve concluded that they don’t need vegetables or whole grains to win.
Yes, this is a silly story. Growing obesity is a problem in the United States, but there are many systematic reasons for its existence, including persistent poverty, the subsidizing of unhealthful foods, poor food selections in public schools, an overreliance on cars, etc. A website like Epicurious simply compiles information, which can be used to create healthful or unhealthful meals.
But substitute “big data” for “Epicurious,” and you basically get the Todd and Dann article. They argue that the use of big data (which they define loosely) by campaigns is contributing to the polarization and paralysis of American politics:
Big Data — a combination of massive technological power and endlessly detailed voter information — now allows campaigns to pinpoint their most likely supporters. These tools make mobilizing supporters easier, faster and far less expensive than persuading their neighbors.
Now, to be sure, voter data and the computer power available to process it have improved considerably in recent decades, and this same time period has seen substantial partisan polarization in American politics. But Todd and Dann seem to be confusing this correlation with a causal story. It’s not.
Polarization in American politics is right now about as strong as it was in the late 19th century. There’s probably a mediocre steampunk sci-fi novel out there about the McKinley campaign hiring an analyst to build some Charles Babbage analytical engine to perfectly target prospective Republican voters across the country (and if not, dibs), but really, big data was not the story then.
Partisanship was strong then because it typically is strong when you don’t have a large swath of white conservative voters aligned with the more liberal of the two major parties, as was the case in the mid-20th century. That was the weird era in American politics. Now that white Southerners are largely aligned with the Republican Party, most Democrats are pretty liberal and most Republicans are pretty conservative. And Republican-leaning districts tend to nominate and elect pretty conservative representatives, while Democratic-leaning districts send more liberal representatives to their legislatures. The campaigns may use technology in service of those elections, but we’d have pretty much the same level of polarization even if we were lighting our homes with candles and commuting on horseback.
But beyond that, Todd and Dann’s piece seems highly confused about modern campaign approaches. The authors blame big data for moving campaigns away from persuasion and toward partisan activation:
[T]hanks to the advent of what was first known as "micro-targeting," campaign consultants realized that the easiest way to win wasn't to persuade the folks in the middle at all. Instead, data could be used to activate every possible base voter and build a partisan firewall.
But that’s not what microtargeting does. Microtargeting means identifying small groups of voters who plan to vote one way but might be persuaded to vote another way because they are cross-pressured on some key issue. When Democratic campaigns use sophisticated consumer profiles and polling methods to try to identify the suburban women who might vote Republican but don’t want to see Roe v. Wade overturned, that’s microtargeting. When Republican campaigns identified those who largely agreed with the Democratic agenda but were leery of the Clintons, that’s microtargeting. What’s more, it’s persuading the folks in the middle — precisely the thing Todd and Dann claim is no longer happening.
They suggest the 2016 election was a perfect example of big data helping campaigns construct partisan firewalls, but that’s not what happened at all. Where did most of the presidential campaign visits and advertising occur? In the swing states. Among whom? People who were undecided until very late in the campaign. Yes, both Democrats and Republicans convinced their bases to turn out, but the way Trump won in several key states was, arguably, by convincing downscale white voters who often vote Democratic to take a chance on him. One could certainly make a claim that those who were persuaded toward the Republican ticket were misled, but that’s beside the point. Persuasion happened. Firewalls fell.
Todd and Dann are right that advanced data analysis can be used both to unite Americans and to tear them apart, but it’s silly to claim that it’s responsible for the polarization we see today. It’s like claiming that our lives are overscheduled because our watches are getting more accurate. Big data is just a tool for dealing with the political world in which we already live. It didn’t create that world.