Unreliable Polls
New York Times, January 1, 2025
RE: “Polls Strength in 2024: Spotting the Big Trends” (News Article, December 15)
Nate Cohn claims “a decent case” can be made that the 2024 polls were “the best they’ve ever been”, indeed that they “excelled”.
Seriously? That analysis — relying on data about certain voter sub-groups — might be accepted by a few statisticians in a laboratory, but it is totally divorced from what the general public encountered. The polls portrayed an election that was unbearably close, on the knife’s edge, hanging by a thread (pick your favorite cliché). The reality was the opposite: an undisputed one-party sweep of all three branches of government, all seven swing states, the popular vote and the electoral college.
Unfortunately, the pundits and the media treated the polls not as fallible predictions but as virtually flawless portrayals of the inevitable outcome. Accordingly, they projected a post-election scenario of divided government, protracted litigation, and threatened democracy. This characterization dominated media coverage of the campaign and overwhelmed serious reporting about the candidates and the issues.
In future campaigns, the media must revise the way it covers polls. Their significance should be downplayed and their uncertainties exposed. They should never replace good, old-fashioned reporting by skilled journalists.
Moreover, the media should no longer sponsor its own polls. To do so creates a substantial conflict of interest. How can a media outlet objectively report on, question, or give proper weight to a poll that it has sponsored and financed? Instead, it publishes exercises in self-justification like Mr. Cohn’s article
.jpg){kind=spacer}