Friends ping Mark Grebner about their love interests. Politicians rely on him to help target voters. Reporters ask whether the suspect in a mass shooting is Democrat or Republican.
In us-and-them America, where many people are eager to render flash judgments on the basis of a person’s political choices, Mr. Grebner tells them the score — literally.
For two decades, Mr. Grebner, 72, has been a pioneer in assembling vast databases that can generate a single number to predict which party a voter is most likely to support.
As campaigning for Tuesday’s election entered the final stretch, dozens of candidates for mayor and city councils across Michigan were using Mr. Grebner’s political scoring to target persuadable voters by knocking on their doors and bombarding them with mailings.
With a few keystrokes, Mr. Grebner can call up the score of just about any adult in Michigan. The rapper Marshall Bruce Mathers III, better known as Eminem, scores as 58 percent likely to vote Democratic. Gov. Gretchen Whitmer comes up as 99 percent Democratic. William Clay Ford Jr., the executive chairman of the Ford Motor Company, scores 86 percent Republican.
“If you give me a name of someone you know in Michigan, I could appall you by how much information I have about them,” he said one afternoon, as he pecked away at his keyboard in his ground-floor apartment in East Lansing. Scout, his deaf, 17-year-old Australian cattle dog, meandered nearby.
Americans have never been so thoroughly dissected and judged by data. Credit scores are key to determining whether someone gets a loan and at what interest rate. Students have SAT scores. Adults have cholesterol scores. Everyone has a Social Security number.
Political scores, assembled from a variety of voting and personal data, are the rating that many Americans don’t even know they have.
Yet lists of voters with reams of data attached to their names are relentlessly bought and sold in the world of politics, a niche of the mammoth market of Americans’ personal data.
In their drive to divine how people might vote, analysts seek out correlations across broad fields of data, said Paul Westcott, the executive vice president of L2, one of the country’s leading vendors of voter lists. Boat owners, for example, tend to be more conservative, while wine drinkers trend progressive, Mr. Westcott said.
L2’s political partisanship database has 850 fields for each voter, a vast trove of information on everything from ethnicity, education and income to gun ownership, donations and more obscure categories like “interest in woodworking” and ownership of cats, dogs or horses.
All of these markers of identity help decipher who we are. But in recent decades, party affiliation has reached tribal status. For some people, knowing whether a friend, colleague or potential lover votes red or blue is a make-or-break piece of information.
Americans also increasingly want to understand the political affiliations of suspects after acts of political violence like the murder of Charlie Kirk in September or the attempted assassinations of Donald Trump last year.
“Everyone is grasping for a reason and trying to say, Oh, it’s left wing, or Oh, it’s right wing,” Mr. Westcott said. His company’s data includes what it calls a partisanship score for 217 million American voters. The score is a number from 0 to 100 that tries to predict which party a voter will choose in the next election.
L2 said it sells its lists to political campaigns, advocacy groups and researchers, but has a blanket ban on selling data for commercial purposes.
Mr. Westcott also said that he receives queries about individuals, which he does not answer.
In addition to L2, which is nonpartisan, a handful of other companies compile national lists of voters, including i360, which specializes in conservative clients, and Catalist, which has largely Democratic clients.
How long political scores will remain concealed from the public is an open question.
“I suppose it would be naïve to think that this kind of information will remain inaccessible forever,” said Costas Panagopoulos, a professor of political science and expert in voting behavior at Northeastern University.
The Trump administration is already trying to link the data about individual Americans that is held in various government databases. It has also demanded sensitive personal data from states to help it build a national voter list.
Artificial intelligence is likely to offer even greater detail and precision to data about voter preferences, Mr. Panagopoulos said. By vacuuming up social media posts, an algorithm could take into account whether someone attended a protest, or whether they were photographed with Make America Great Again gear. Candidates for office might be able to use an A.I. model to decide which issues would resonate with a voter and the best time of day to reach them.
For the friends and acquaintances of Mr. Grebner, the data is just a phone call away.
On a Sunday afternoon in late September, a reporter from a local Michigan news service called him to ask about the political leanings of the suspected gunman in an attack on a Mormon church that had killed four people that morning.
Social media was filled with conjecture on the motive for the attack in Grand Blanc Township, a suburb of Flint, Mich., as people sought to understand the violence.
Mr. Grebner told the reporter the gunman’s score: 81 percent likely to vote Republican.
Further reporting by journalists in the days after the attack revealed that the suspect, a former Marine named Thomas Jacob Sanford, did not appear to have been motivated by politics, but by a deep hatred for Mormons.
But the rush to view the shooting through the prism of politics highlights how these databases reduce voters to a number — and the risks of oversimplifying complex views.
Still, the demand for the information remains.
Calculating political scores for voters, which began around two decades ago, is both art and science. Algorithms take into account the obvious, like party registration or whether a voter requested a ballot for a Republican or Democratic primary.
But in some 20 states, including Michigan, voters do not list their party preference, increasing the importance of other data to divine a voter’s inclinations.
Mr. Grebner’s database looks for patterns among voters in the same household — they often vote alike. And he obtains copies of petitions that circulate in the state and sends the thousands of pages to Bangladesh, where data entry specialists record each signer.
Mr. Grebner revels in his own eccentricity. A staunch Democrat, he has a collection of Make America Great Again hats that he wears among family and friends when he knows it will rile them. He said he does not spend time thinking about the pros and cons of gathering and selling data on his fellow Michiganders.
“I want better ways to understand people,” he said.
Asked whether it is potentially harmful for every American to be assigned a political score, he sounded flustered.
“Is it a good thing or a bad thing? Well, it just is,” he said. “Is it good to refine arsenic? I don’t know. But if I were to refine arsenic, I would want it to be really pure.”
Mr. Grebner, the son of a school superintendent who traveled from job to job across the Midwest, has a made a career of using data to tweak the establishment.
As a student at Michigan State University in the 1970s, he riled the administration and faculty by creating a booklet that ranked professors based on surveys of students, a pre-internet version of websites that rate professors.
In addition to running his company, Practical Political Data, he is a longstanding member of the Ingham County Board of Commissioners, and his election slogans over the years prove that politics and humor can successfully mix.
“No worse than the rest,” says one of his bumper stickers.
“We can’t get rid of him,” reads a yard sign.
He began ranking and categorizing the partisanship of voters in the 1990s, and his scoring system came in the early 2000s. His methods were widely admired among politicians and those who worked in the field.
“I would describe him as an irreverent genius, very idiosyncratic and interested in upending conventional wisdom at every opportunity,” said Donald P. Green, a professor at Columbia University who has worked with Mr. Grebner on research projects.
While most of Mr. Grebner’s clients are political candidates, he is happy to help out a friend in need. Four years ago, a fellow graduate of Michigan State, Lawrence Kestenbaum, sent him a query: “What do you have for Andrea Curtis Rothney?”
Mr. Kestenbaum, a committed Democrat who is the Washtenaw County clerk, had spotted Ms. Rothney on Tinder and was cautiously courting her.
Mr. Grebner reported back that Ms. Rothney was not only a Democrat, but a dedicated one, “serious about voting and signing petitions.” His model scored her as 93 percent Democratic.
“I was pleased and maybe a little relieved,” Mr. Kestenbaum said.
Three years later, the couple were married.
Ms. Rothney, who only learned about the political due diligence her husband had conducted when a New York Times reporter reached out to her about it, says she thinks it was a “good idea.”
A match made in Democratic heaven?
“Hmmm,” Ms. Rothney pondered the question. “Well, purgatory, anyway,” she said.
Thomas Fuller, a Page One Correspondent for The Times, writes and rewrites stories for the front page.
The post The ‘Political Score’ You Might Not Even Know You Have appeared first on New York Times.
