On one of the most important nights of the 2020 Democratic primary, politics junkies went to bed disappointed.
By now, you probably know why — the Iowa Democratic Party commissioned a smartphone app to streamline tabulation of the much-hyped Iowa caucuses, but the program created more delays (and confusion) than it solved. Pair that with other new bells and whistles, including three different ways to present returns, and it’s not hard to see why we still don’t fully know who finished where.
That didn’t stop the candidates from quickly spinning the chaos in their favor, as each took turns delivering late-night stump speeches — many of which read like repurposed victory addresses — to a national TV audience. Bernie Sanders and Pete Buttigieg went a step further, with both campaigns releasing unverified internal projections showing their candidate ahead.
Online, the scene wasn’t much calmer; some journalists hung around individual caucus sites and tweeted estimated headcounts. Others scrambled to make sense of the few official returns that did exist, despite the information coming from fewer than 40 precincts statewide. A few even tried to crowdsource their own results by retweeting collaborative spreadsheets for precinct captains to fill out. So much happened, I honestly didn’t see until two hours after voting started that The New York Times had revived its infamous Upshot needle…before shutting it down for the night shortly after.
In the aftermath, most blame has rightfully been put on those tasked with running the election, including the developers of the mobile application, the Iowa Democratic Party, and the state of Iowa itself, with the “Should Iowa be the first to vote?” debate starting anew. Yet, many election watchers and reporters are also at fault for how the night was covered.
Avid readers hopefully know my opinion on analyzing surveys: it’s better to look at an average than just one. Thanks in part to the law of large numbers, with some help from common sense, election night returns operate much in the same way. And it’s here that many journalists failed.
In fairness, some likely only read into early precinct data with the expectation that they’d receive supplemental information later. Nevertheless, even if it had arrived on-time, relying on incomplete, unofficial, and anecdotal evidence can be irresponsible and misleading — to be clear, what I’m describing is different from analysts judiciously using sparse data to make good-faith forecasts or inferences; early returns have utility, but only when handled responsibly.
Other figures, like the projections from the Sanders and Buttigieg camps, had no right making such substantial media waves. While I’d wager the numbers aren’t fabricated, they should nevertheless be approached like internal polls, with skepticism and an open mind: if a campaign releases something that was initially meant to be private, it probably has an ulterior motive.
Many, myself included, dropped the ball last night, but there’s still time to improve before the New Hampshire primaries (which are only a week away!) and beyond. Though the most obvious solution would be to never trust anecdotes, that advice is neither realistic nor true.
Instead, I caution you to not get swept up in the excitement. Much of political journalism relies on the horse race, and when that race occurs over a few hours — rather than an entire election cycle — the superficial punditry is often more damaging and specious than ever. Especially for those who cover politics as a career, maintaining a cool head in unfamiliar times is a must.
For all involved, last night’s ideal outcome was a quick and efficient caucus, but we didn’t get that. The next time an election goes wrong, no matter how frustrated you are or how badly you need a story, the first instinct shouldn’t be to make hasty assumptions about incomplete data or to sleuth out your own returns. Just wait out the anecdotes, because, I promise, full official results will come. Eventually.
First Past the Poll 