Politicians may tout polls that cast their candidacy in its brightest light. TV stations have a tendency to report on polls they view as newsy, even if they aren’t necessarily accurate.
The Washington Post compared polling and reporting results from the 2008 presidential campaign and found TV newscasts tended to report polls that showed a tight race between Democrat Barack Obama and Republican John McCain, even though the average results of all polls reflected Obama was clearly ahead.
“On average,” according to the Post story, “viewers likely heard more about polls that portrayed the race as close than polls that more accurately showed Obama in the lead.”
TV news gave more air time to jumps in poll results, which gave the race that worn on for months spark of “breaking news," even if the actual poll changes were mere aberrations.
The researchers who prepared the Post report conducted an exhaustive search of all the polls that TV reporters could have cited. They found time after time polls showing Obama gaining or holding the lead were bypassed in favor of polls that showed the race as very tight.
Polling can be slippery enough in revealing the actual mindset of the electorate without any help from people trying to put their finger on the scale. Pollsters use different samples and apply different techniques. The timing of when a poll is conducted can be before or after a significant event in a campaign. How questions are framed can shade the results, too. With all that built-in variation, selecting what you might call outlier polls makes the reporting even more suspect.
As Nate Silver, founder of FiveThirtyEight.com, has said, looking at the averages of all credible polls provides a more reliable picture of where a race stands and what voters are thinking. Even the average of all the polls can be wrong, Silver admits, so the takeaway is don’t take polls too seriously, especially when they are still weeks to go before the election.
Looking at the 2016 presidential election, polls have played a large part in both the campaigns and the reporting of the race. GOP nominee Donald Trump routinely bragged about his lead in the polls during the primary season and has complained that poll results showing him trailing Democrat Hillary Clinton are media fictions. Trump’s campaign manager, who is a pollster, says Trump is actually ahead, but his secret supporters are unwilling to tell pollsters they plan to vote for him.
The media, especially TV news, gave a lot of coverage to the post-convention polling bumps that both Trump and Clinton received. And, there has been nonstop media discussion about GOP fears that a slumping Trump presidential campaign will have negative effects on down-ballot Republican candidates, especially in a handful of tight Senate races that could decide which party is in control in the next Congress. And TV networks have begun reporting on what CBS News calls a “poll of polls” to avoid cherrypicking results.
Coverage of the 2016 race has drilled down into the demographics of the race. The simplest narrative is that Trump appeals to white, older, non-college-educated men while Clinton appeals strongly to women, minorities and people with college degrees. This narrative fits neatly into the tight time frames for TV news coverage and may be this campaign season’s newsy fixation, even if it is over-simplified and nearly stereotypical.
Skepticism of political polling remains a smart move, as 2012 GOP presidential nominee Mitt Romney learned when he headed into election day thinking, based on polls, he was about to be elected president. Believing in polls is a little bit like believing your own press releases. Believe at your own risk.