In many ways, 2016 is poised to be a defining cycle for pollsters. Survey research firms now face tighter calling restrictions thanks to the FCC, the Brexit results from across the pond have resurrected the specter of survey research uncertainty, and some within the industry are predicting widespread changes to how pollsters do business within the next four years.
Enter a Republican nominee working overtime to sow doubt with baseless warnings of wide-scale voter fraud and “rigged polls” (plus a campaign that claims it has identified a crucial group of voters who aren’t being picked up in surveys), and the pressure to get it right for pollsters is only ratcheted up.
As the race between Hillary Clinton and Donald Trump tightens a bit in the final stretch, some recently released public polling has shown significant swings. Take Pennsylvania, widely considered a must-win for Trump. The latest Franklin & Marshall College poll has Clinton up 11 points compared with a Remington Research Group statewide conducted in a similar window that has Trump within two points.
“Lots of uncertainty,” Nate Silver tweeted Oct. 31.
That uncertainty will be compounded if Trump surprises and wins on Nov. 8. An unexpected result would place pollsters on a very hot seat (not to mention the Trump camp is now famous for spending more on hats than on polling). Even some of what the Republican nominee has spent on polling ($766,756.67) he’s disputing with his vendor, Fabrizio, Lee & Associates.
Some pollsters are responding to the current environment by trying to position their firms to capture the growing demand for data analytics. What the future holds for political survey research is still uncertain, but at least one Republican pollster is openly predicting its demise.
“I believe polling will be gone in two-to-three cycles,” Chris Wilson, CEO of Wilson Perkins Allen Research, told C&E. “That’s why we’ve gone so aggressively into analytics. My challenge is convincing people they need to do analytics.”
Wilson isn’t actually as emphatic as he initially sounds. Polling will undoubtedly survive in some form, he added. “I think it has a place. It’s a place for message testing.”
In other words, you still need a benchmark poll, but brushfire and tracking polls may increasingly become obsolete (at least in the manner most campaigns currently conduct them).
In regard to the latter, Wilson said: “Why do you want me to tell you there are 12 percent undecided [after an ad runs] when we can cookie-match them and go and track them on their devices?”
Wilson isn’t the only one trying to anticipate the movement of the market. Some pollsters have for years been shifting away from traditional survey research. For instance, GOP pollster Patrick Ruffini has long talked about social listening-style services surpassing traditional polling — whether for campaigns or media outlets.
“Most media polls are simple random digit dialing in which they really don’t check if somebody’s on the voter rolls necessarily,” Ruffini recently told the Harvard Political Review.
Meanwhile, polling accuracy can be improved by using the voter file, Ruffini argues. “I think the future of polling is going to be more rooted in hard data about the electorate, in voting files, around what we know about the electorate from large scale analyses, from what the actual demographic and population and likelihood to vote makeup is from actual voter behavior.”
Of course not all pollsters think they’re on a sinking ship. But even those who are optimistic concede that more experimentation is a must.
“There will always be a place for good polling,” said Stefan Hankin of Lincoln Park Strategies. “We are big believers in analytics and believe all good decisions are based on sound data. However, analytics are not going to tell you what the right words and narrative are to persuade movable voters. The days of the traditional polling approaches are coming to an end, but that does not mean an end to polling.”
Hankin admits the presidential race may very well be ripe for a miss with the number of new voters who have emerged this cycle. There are now some 200 million registered voters and it’s the most diverse group in U.S. history.
“There are a lot of different views of what is going on in this race, from a Clinton blowout to a tight race,” Hankin said. “Clearly both can't be true. So we will see who is right, but anyone basing their decisions on public polls should have their head examined in the first place.”
Brent Buchanan, whose firm Cygnal surveys for down-ballot Republican candidates, said he doesn’t disagree with Wilson, but warned that no one knows for sure what direction survey research moves in next.
“It's much like the transition from door-to-door polling to phone polling,” he said. “We just don't currently know what the next reliable step is.”