Election Day is the moment of truth for the Four P’s: presidential aspirants, political parties, pollsters, and of course, the American public.
As Mark Ambinder of the publication The Week notes, today’s pollsters face strong head winds as they try to bring their cargo of voter-intention data safely into harbor by accurately predicting the actual election results. Any polls that run aground will face an embarrassing public reckoning, and probably a close inquiry into what went wrong.
Here are some insights from Ambinder and from Pew Research Center white papers from 2012 and 2016 that focus on how the changing landscape for telephone surveys is impacting political polling (and opinion surveys in general):
- Americans’ willingness to answer pollsters’ questions on politics (or, for that matter, any topic) has fallen “dramatically,” from 36% in 1997 to 9% in 2012, according to Pew. Among cell phone participants, the response rate was just 7%.
- Consequently, Pew noted, “greater effort and expense are required to achieve even the diminished response rates of today.”
- Half of American households don’t have landlines, making them harder to reach for the telephone polls that are standard for presidential pollsters. And, per the Week, “the younger you are, the less likely you are to have ever used a telephone with a cord attached to it.”
- This suggests that polling will face ever greater challenges as we go deeper into the smartphone era. The Millennials who’ve come of age with diminished (if any) connections to landlines will account for an increasingly large slice of the electorate for decades to come. The oldest members of Gen Z, for whom landlines are even more alien, will be eligible to vote in the 2018 midterm election.
- Noting that more than 90% of U.S. adults now have mobile phones, Pew has responded by going to extra lengths to reach them: in its 2015 “Survey on Government,” the balance of respondents was 65% cellphone interviewees, and 35% via landline. Pew sees “an impending transition to 100% cellphone surveys,” even though they cost 30% to 50% more than the no-longer-representative landline method.
- Pew says it takes “high-effort surveys” to land more of those necessary mobile interviews. That includes hiring “elite interviewers” who are experienced and adept at persuading reluctant contacts to take part, and paying cash incentives of as much as $20 per respondent.
- But there are still enormous challenges in establishing a truly representative polling sample. As The Week puts it, “it’s much easier to get certain demographic groups to respond to a pollster’s telephone call, regardless of the circumstance. If you’re a white woman over the age of 50, you’re more likely to respond to a pollster than if you are black, or Hispanic, or a millennial. Pollsters try to control for this in their weighting, but their sample sizes [for the under-represented groups] are often so small that tiny changes compound error rates….every interpolation creates additional room for error.”
- On top of everything else, Ambinder notes, when a race is particularly close, heated, and impacted by volatile events, as Clinton vs. Trump has been, an extra degree of unpredictability can emerge: “If your candidate has a bad news cycle, you aren’t as likely to be as enthusiastic about your vote, and you’re less likely to respond to a pollster, or to be honest with a pollster about your partisan or ideological affinities.”
So will the head winds election pollsters face turn out to be hurricane-force gales that blow their predictions onto the shoals of inaccuracy? Soon you’ll know which of the polls have safely passed the big test, and which will have some explaining to do. And you’ll see whether one of the upshots of this unprecedentedly strange, strained presidential election will be an intense re-examination of political polling methods.