Here's a sampling of headlines today from what are generally respected outlets:
Buzzfeed: "Taylor Swift's Instagram Post Has Caused A Massive Spike In Voter Registration"
Daily Beast: "Taylor Swift: Voter Registration Spike Follows Star’s Political Awakening"
Washington Post: "Taylor Swift’s endorsement of Democrats causes spike in voter registrations"
Variety: "Taylor Swift’s Political Comments Lead to Spike in Voter Registration"
I'd give the Daily Beast a little credit for hedging some, but then there's the lede: "Taylor Swift’s first foray into politics appears to have produced impressive results." All the outlets made the same argument: that, a.) there was an increase in voter registration in the time period following Taylor Swift's Instagram post urging people to register to vote for a Brett Kavanaugh supporter and b.) Swift caused that surge.
But of course correlation does not imply causation, much though the media tends to struggle with that concept. The evidence presented in all these pieces doesn't actually go a step further and prove that the surge of registrations on Vote.org was caused by Swift. On the contrary, some of the pieces allude to a compelling alternative explanation. Here's the Daily Beast again:
"We are up to 65,000 registrations in a single 24-hour period since T. Swift's post," Kamari Guthrie, director of communications for Vote.org, told Buzzfeed, which reports, for context, that 190,178 new voters registered nationwide in the entire month of September, while just 56,669 signed on in August. …
"Vote.org saw [Tennessee] registrations spike specifically since Taylor's post," Guthrie told Buzzfeed, saying the organization has seen 2,144 registrations in the last 36 hours, up from 2,811 new Tennessee voter registrations for the entire month of September and just 951 in August.
Today, as Swift's post sought to inform her fans, is the last day on which Tennesseans can register to vote.
Uh, what was that last part? Tennessee registrations spiked on … the last days people can register to vote? Wouldn't that be exactly what researchers should expect to naturally occur? It's not just Tennessee either: seventeen other states have the same deadline, including populous states like Texas, Pennsylvania and Florida. Deadlines in states like New York and Virginia are also quickly approaching.
The analysis Vote.org is putting forward here seems flawed. It's akin to noting that a certain number of people filed taxes in March, that number surged on April 15, and concluding it must be because Carrot Top tweeted April 14 that people should file their taxes. The far likelier explanation is that many people tend to put off things until the last minute.
The only slightly compelling evidence that Swift might be responsible for voter registrations comes from WaPo, which notes that "Among the new registrations since Sunday, more than half were by those ages 18 to 29." But even that doesn't tell us anything without comparing the demographics on previous days. My suspicion is the category of voters who have to register to vote in the first place—people who moved, changed their names, or didn't vote in the previous four years—would tend to naturally skew young. Millennials are probably also likelier to register on sites like Vote.org than older Americans given that they don't understand how stamps work.
The correct way to judge the impact of Swift's post would be to compare last-minute online registrations this year to the numbers in previous midterm years. Even an increase would of course still just prove correlation; it's possible voter registrations could be up this year for unrelated reasons, such as general enthusiasm, backlash/support to the weekend's Kavanaugh confirmation, improving ground games and technologies by political parties, etc. But journalists didn't even bother with that basic step.
It's of course entirely possible that Taylor Swift did lead to an increase in voter registration; it'd be shocking if an Instagram post from someone with 11 million followers had zero impact after all. But that Swift's post definitely had an impact and that that impact numbered in the tens of thousands is speculation, and it's sloppy journalism to suggest otherwise.