Harry Enten, a forecaster at FiveThirtyEight, believes he isn’t a favorite of Bernie Sanders supporters; they “don’t exactly love me,” Enten recently wrote. That may be true, as Enten’s coverage of Sanders during the Democratic primary has, for the most part, been extremely dismissive. When Sanders entered the race in April of 2015, for example, Enten asserted that Sanders had “pretty much no shot of winning.” And despite the massive growth in Sanders’ support in the time since, Enten had barely changed his tune nine months later – at the beginning of March, in an article entitled “Hillary Clinton’s Got This,” he argued that “Clinton eventually will win the nomination with relative ease.”
Enten’s underlying analyses have generally been based on polling, endorsements, demographics, and other reasonable indicators. And Sanders supporters have long understood that Sanders’ path to the Democratic nomination is an uphill climb. The main objection with Enten-type coverage is not that it’s based on faulty data analysis – it’s not – but that it is at least as much a driver as a predictor of election results.
What does that mean? It means that two of Sanders’ biggest obstacles to winning the Democratic nomination – perhaps the two biggest obstacles – are the facts that many people aren’t familiar with him and that, of the people who are, many don’t think he has a chance to win. The latter has driven organizations and individuals who should be natural Sanders allies to, because of a misguided sense of pragmatism, endorse Clinton instead. It has also driven highly unbalanced media coverage. These outcomes make it harder for Sanders to introduce voters to his extremely popular ideas, his consistent record of fighting for social justice causes, and his history of legislative and executive success.
Unsurprisingly, the more voters hear about Sanders, the more support Sanders gets. The problem is that, because of Enten-type assertions that Sanders doesn’t have a chance, fewer voters get to hear about Sanders than otherwise would. That lack of exposure ensures lower levels of support, thus often confirming the forecasters’ original predictions and emboldening them to make new ones that further feed into this loop. Forecaster predictions, in other words, are frequently self-fulfilling prophecies.
That’s one reason Sanders’ historic win in the Michigan primary is so important – by exposing how wrong pundit predictions have been, it has the potential to break this self-reinforcing cycle. The results in Michigan illustrate more than the inherent uncertainty in data analyses; they show that pundit assertions about which candidates have a chance should always be viewed with skepticism.
To Enten’s credit, he responded to Sanders’ victory by owning how wrong he’s been – “Bernie Sanders made folks like me eat a stack of humble pie on Tuesday night,” Enten wrote. Vox’s Matt Yglesias, who openly mocked the Sanders’ campaign’s chances in Michigan six days before its victory, also admitted his error.
Where did the Sanders campaign get this idea that he can win Michigan?
— Matthew Yglesias (@mattyglesias) March 2, 2016
I thought Bernie would lose because polls showed him way behind. Clearly I was wrong! https://t.co/SH7SfBPQH1
— Matthew Yglesias (@mattyglesias) March 9, 2016
Yet while these one-time corrections are a good start, they’re also not nearly enough. Rather than having humility as their dessert, Enten, Yglesias, and the rest of the punditry would do well to consume a little more modesty before – not after – their future political meals.