Political poll takers have had a difficult time of late. Traditionally, they have used robo-phone calls to households to measure the likelihood that the respondent will vote for a certain candidate. But more than 40 percent of people, especially the under-30 crowd, have given up their landlines. They rely solely on cellphones, and the 1991 Telephone Consumer Protection Act prohibits autodialing cellphones. This confronts pollsters with a dilemma, because it limits queries to those who are middle-aged or older and lean conservative. Their interests and preferences are vastly different than those of younger respondents who tend to tilt left.
Compounding this issue is the prevalence of caller ID. If these older folks resemble my household, they don't pick-up on unrecognizable numbers and names, figuring that if the call is important, the caller will leave a message. Most don't, thus reinforcing our behavior. This helps explain why only an average of 9 percent of people contacted respond to telephone polls.
Pollsters are also using Internet surveys to count votes, but again this favors a younger audience. And what's more troublesome, it is left to individual computer users to respond to the survey on their own time — not a dependable strategy for people who would rather be on Facebook. And so, this reliance on a mixture of landline and Internet polling gives a natural edge to older, more conservative voters. Among Republicans, the unintended consequence of all this is to give Donald Trump a built-in advantage in the polls.
An analogous technology gap occurred back in the early days of polling. Beginning in 1916, The Literary Digest counted potential votes by mailing out ballots to names garnered from subscriber lists, and later from telephone directories and automobile registrations. By 1932, this amounted to 20 million people. The Digest had used this straw poll methodology to correctly pick the presidential winners from 1920 to 1932. But in 1936 their respondents chose Republican Alf Landon, only to be proven wrong when Franklin Delano Roosevelt won with 57 percent of the vote. The Digest had failed to realize that Great Depression-battered Democrats were less likely to own telephones and cars than were Republicans. Are Trump's poll numbers similarly suspect?
In my "Webster's Unabridged," the first definition of "poll" is "the head; especially the crown, back, or hair of the head." Hence, polling is nothing more than head counting, and that's how this practice got its name.
Fox News relied on the average of five national polls to select who took the stage for their first presidential primary debate, and subsequent debate sponsors continued to use poll results to choose participants. Again, this surely skewed the anointing to Trump, whose strongest following is among the over 50, angry white male cohort still relying on landlines. What's more, the candidates' position on stage was poll-determined, thus placing Trump at the center of TV screens on wide shots, enhancing his perceived importance. And once his superior status was established, it became self-perpetuating.
To succeed, a demagogue requires an angry and frustrated core constituency to whom he promises swift and easy remedies. They will brook almost anything he says, no matter how extreme. The next prerequisite is a big stage from which he can win over even more adherents. Trump's reality show and proclivity for affixing his name to everything from buildings to fragrances gave him a leg up here, and the poll-driven debate selection process has nurtured his rise. The media's saturation coverage sealed the deal.
In the not-so-distant past, our political parties selected presidential candidates in smoke-filled rooms where sketchy pacts were made. Election polling was initially hailed as a way of changing this — of giving the nominating process transparency and awarding candidate selection to the voters. But are current polling practices legitimate, and are they good for representative democracy? New Yorker writer, Jill Lepore, thinks not. She claims that polls disenfranchise all those who can't or don't wish to answer the phone. That's the vast majority of the population. Hence, the all-important winnowing down process of candidates is now left to an unrepresentative demographic, and we end up with candidates who may or may not be best for the country.
Ireland, France, Turkey and Scotland are seeking to limit the use of political polls, especially in the run-up to elections. This has First Amendment implications here, but may be worth the debate, especially if it helps save our democracy.
Frank Batavick writes from Westminster. His column appears Fridays. Email him at fjbatavick@gmail.com.
