Con Psarras: We need a different kind of opinion polling
Julie Drainer watches election night coverage at Ector County Republican Headquarters on Tuesday, Nov. 3, 2020, in Odessa, Texas. (Odessa American/Eli Hartman)/Odessa American via AP)
The question of why polling projections were so off in critical voting districts for the second straight presidential election now begs a bigger question: Why should news media continue to invest in showcasing the kind of “horse race” polling that can’t be counted on to pick out the fastest horse in a given race?
There has long been concern about how the media’s obsession with head-to-head candidate polling has skewed political coverage toward something more suitable to sports coverage, detracting from thoughtful analyses of candidate policy positions and voter sentiment on specific issues. That’s especially the case with cable and broadcast outlets.
There are valid arguments in favor of such polling, but they fall apart when the polling is wrong, especially in ways that disrupt the electoral process and leave voters frustrated with what they got as opposed to what they were led to expect.
The polling establishment is now writhing in a harsh exercise of introspection. What exactly went wrong and why? What lessons were not learned in 2016? The autopsy will be ongoing, but any blame brought to bear shouldn’t lie only at the doorsteps of the polling firms.
When polls are wrong, the news organizations that sponsor them own a share of that wrong. At a time of diminishing trust in traditional news sources, this can’t be good.
In my tenure as a newsroom manager, I hired and supervised the work of a number of pollsters and found them invariably to be professional and proficient. The late Dan Jones, as an example, was fastidious to a fault, a perfectionist who was meticulous about data to the point of paranoia. Over a respected career, he missed only a handful of high-profile races, but the curse of a pollster is being remembered for the ones you got wrong.
That curse is now upon much of the polling establishment and, by extension, its news media cohort. “You just can’t trust the polls” was a common refrain during the recent campaign. Once trust is gone, it may not return. Would you retain the services of an accountant whose assessment of your tax obligations was off every year, by an expanding margin of error, plus or minus?
With successive election cycles of notable polling misses, news organizations should rethink the practice of fashioning political coverage around pollster projections. It’s done largely because it’s easy, and because polls are content winners. Audiences love a race and a poll is an “add-water-and-stir” meal for news producers. Draw up some graphics, send a reporter to get a few person-on-the-street interviews and, boom, you have a lead story.
Not that horse-race polls are without value. It’s important to know which candidates are succeeding, particularly in crowded primary races. Academic research, though, has found that head-to-head polling coverage accounts for as much as three-quarters of all attention given to political races.
Polling resources would be better used to delve into voter attitudes about specific candidate proposals. Fleshing out the alignment between voter sentiment and a candidate’s position on an array of issues is of greater public service than a breathless barrage of numbers showing a race narrowing, widening, or too close to call. Pollsters are capable of asking nuanced and complex questions beyond, “Whom do you plan to vote for?”
They should be put more to that task by their journalistic stewards, whose essential job is to chronicle the present, not predict the future.
Con Psarras has been a television news director and has taught journalism at the University of Utah.