The Fall of Poll News
"Do you ever get the feeling that the only reason we have elections is to find out if the polls were right?" - Robert Orben (comic writer)
Conventional wisdom says most folks take a campaign time-out in the summer. It's not until September that interest picks up, before it gets intense around October. Yet, according to the latest News Coverage Index, published weekly by the Project for the Excellence in Journalism (PEJ), news about McCain and Obama made up 26 percent of last week's newshole (the amount of print space or air time devoted to news amid the avalanche of ads).
If a quarter of all news being focused on the race for the White House seems like a lot, that's half of what it was during the Democratic primaries. That breeze you feel is what PEJ describes as an early summer "period when media attention to the race is cooling noticeably."
But even with campaign coverage "cooling," it was still the top story in all five media sectors (newspapers, online, network TV, cable TV, and radio). On cable TV, 49 percent of airtime focused on McCain v. Obama.
To break it down a bit further, one of the most covered campaign stories of last week was how McCain and Obama matched up in polls. Even though the primary season is over, journalists and news consumers can still expect to be hit with a constant flurry of presidential poll news, leading up to presidential poll-a-palooza as we approach November.
And while it may sound dignified and sophisticated to say: I don't believe in polls, the science of quantifying public sentiment has come a long way. That's not to say polls should be taken at face value. You have to learn how to interpret polls. If you're so inclined, a good place to start is the online course offered by News University and the American Association for Public Opinion Research. For free, you can learn about polling methods and how to crunch the numbers to see if the polls add up. Even if you're not a journalist, imagine the intriguing cocktail party conversation you can have, sharing with your company how to know when nine out of 10, really isn't nine out of 10.
Meanwhile, your curmudgeonly neighborhood columnist has done you the favor of coming up with the five things you need to know about polls -- a little list I distilled after talking with polling expert and political science professor at MIT, Steve Ansolabehere. Author of "Going Negative: How Political Advertising Alienates and Polarizes the American Electorate" and former co-director of the Caltech-MIT Voting Technology Project, trust me -- he's wicked smaaht.
5. This far away from the November election, polls about which candidate is ahead don't tell you much. "If you take a survey this far from an election, the intention measured in the poll may not be a good indicator of how people will vote." (Poll history note: When Dukakis was the Democratic contender running against Bush Sr. in 1988, Dukakis was ahead in the polls at this point in the campaign).
4. Over the course of a general election campaign, fluctuations in the polls actually show what voters are learning about the candidates. But, "people usually end up learning more about the candidate they were probably going to vote for anyway. The underlying, fundamental indicator of how someone will vote is party identification. That tends to be where most people end." The seminal treatment on this is considered to be Gary King and Andrew Gelman's paper, published in the British Journal of Political Science where they explore the question: "why do polls vary so much when elections are so predictable."
3. The big debate among pollsters right now is about how to conduct polls. The good old-fashioned way was to do it going door-to-door. With the rise of communication technology, came the advent of the "random digital dialing (phone) survey." The cheapest and supposedly lowest-quality polling is done with "robo calls" - when a computer calls and conducts the survey.
The problem with phone surveys, Ansolabehere noted, is that response rates have dropped over the years. Also, it's not clear how much the polling picture is distorted because pollsters don't call the growing number of voters who only have cell-phones. Yet, if you compare all the various polls with actual voting results in the primaries, you'll find that "robo-calling" Survey USA, "beat the pants off everybody." Still, in terms of overall quality and accessibility, the Pew Research Center for the People and the Press is the best in the business.
2. There are things even polling experts don't understand about polls. For example, when it comes to predicting winners, Gallup polls "has a systematic GOP bias, while exit polling data has a Democratic bias. It's a mystery why that is, really."
1. While polling data is used by the ruling elite to manipulate and massage public opinion, we can all gain a valuable insight from stepping outside of the campaign polling arena and consider one of the most consistent findings of public opinion across decades. In poll after poll, year after year, the overwhelming majority of Americans say they want universal health care, despite considerable propaganda about "socialized medicine" lobbed from pure "free-market" ideologues. But now that big business is worried about health care costs, suddenly the dominant conversation isn't about whether there ought to be universal health care but what policy is the "right" one to deliver it.
That we don't have universal health care -- despite a consistent, majority public opinion in favor of it, going back decades -- tells you something about power in U.S. society, not the least of which is that when politicians like W say they don't pay attention to the polls, they're not joking. Apparently, what "the people" want doesn't matter. What business people want, does.
Now, summer on -- before poll news falls on us again.
Sean Gonsalves is a columnist and news editor with the Cape Cod Times. He can be reached at firstname.lastname@example.org