https://newrepublic.com/article/154124/ ... try-crisisThe polls were unequivocal. In 2016, two days before Michigan’s Democratic primary, the respected Marist poll predicted that Hillary Clinton would win in a landslide, with 57 percent of the vote to Bernie Sanders’s 40 percent. On the morning of the March 8 primary, James Hohmann of The Washington Post wrote, “Michigan should have been fertile territory for Bernie Sanders’s populist and protectionist message, but he’s expected to lose the Democratic primary there today by double digits.” That night is mostly remembered for the marathon coverage of a bizarre press conference in which Donald Trump, who had just won three primaries of his own, shamelessly hawked Trump Steaks and Trump Wine. But if TV viewers had squinted at the crawl, they would have noticed that Sanders was pulling off a stunning, poll-defying upset—defeating Clinton by 1.5 percent. At the time, Nate Silver called it “among the greatest polling errors in primary history.”
And yet the media (liberal and conservative alike) continue running story after story using polling data that is not worth much at all...and people like our own poll expert here CS continue defending them...
So, how can an industry that has had so many big misses continue to thrive...and grow...To take roughly one week in May, they produced such headlines as, “Biden dominates Dem rivals in new early primary polls” (Politico), “Warren’s rebound? Massachusetts senator gains footing in polls” (Fox News), and “Brutal new 2020 numbers for Beto O’Rourke and Bill de Blasio” (The Washington Post). What these stories didn’t mention is that today, seven months before Iowa, polls are about as accurate as a blunderbuss, and most pollsters know it.
Well, the 24 hour news cycle is a major reason...the need for political news is inordinate and “polls” can be done by each side to show their “strength “ with voters...
To explain the wild fluctuations in their numbers, pollsters and their eager enablers in the press have created an artificial narrative of candidates bouncing up and down as if the campaign were conducted on pogo sticks. As Karlyn Bowman, a polling analyst for American Enterprise Institute, said, “Pollsters think that it’s good for business to have a new poll almost every morning.” Some are experimenting with online and texting polls. But until they come up with a reliable substitute, almost all of them are forced to fudge the numbers, artificially inflating the responses they get from, say, young voters.
The press pack, for its part, cannot shake its addiction to polls any more than candidates can resist the lure of TV cameras. Even if reporters could return to practicing retro journalism—divining the leaders, not with polls, but by talking to voters, county party officials, and other insiders—all those people are now glued to the polls, too, and their responses are as canned as an MSNBC talking head’s. We’re left with no other choice than to endure horse race journalism while the horses are still doing their morning workouts.