Any farmer who has had a crop fail despite doing everything right can sympathize with public-opinion pollsters. Even the pollsters who used the smartest, most sophisticated techniques -- the ones who did everything right -- botched the 2016 election. Almost all of them thought Hillary Clinton would win.
Why did they get it wrong? Sometimes pollsters flub the wording of questions, but that wasn't the problem this year. Often pollsters have a hard time getting responses, but the best of them did what was necessary to achieve a valid statistical sampling.
The most convincing explanation for this year's failure is a factor beyond the pollsters' control, akin to the bad-luck weather that can make crops fail. The polls, it seems, could determine how people would vote if they voted but couldn't accurately predict who would vote. For that the pollsters had to extrapolate from exit polls from previous years' elections. From history, in other words. The problem is that history doesn't necessarily repeat itself; sometimes, to contradict Mark Twain, it doesn't even rhyme.
There's another reason pollsters go astray, one that may also have played a role in this election. It's certainly one to keep in mind when thinking about survey results on food and agriculture issues. Simply put, it's that people often do not tell pollsters the truth. They may or may not be lying -- more on that shortly -- but the way they answer polls and the way they actually behave are sometimes at odds.
In this election, for example, there were people who told pollsters they planned to vote for someone other than Donald Trump because they were embarrassed to admit their true intentions. In surveys on food, huge majorities of Americans affirm their solemn belief in healthy eating, yet many of these people clearly do not practice their faith.
Big food companies know better than to trust what people tell pollsters. Some of them have had trouble finding the right balance between serving those who fill their shopping carts with healthier food and those who consume a lot of potato chips and candy bars, whatever they told the pollster.
The Wall Street Journal recently chronicled PepsiCo's inability to meet its own internal targets for increased sales of healthier food and beverage offerings. According to the Journal, "high-fat, high-salt standbys such as Doritos and Cheetos" account for much of Pepsi's revenue growth (http://tiny.cc/…).
Apparently some people just want to give the pollster the "right" answer -- in this case, "healthy." Others have inaccurate memories or unrealistic aspirations for doing better in the future.
A surprising number simply have their own idea of what constitutes healthy.
For example, in one poll nearly 90% claimed to eat healthily (http://tiny.cc/…). Yet asked in that same poll about their vegetable consumption, most checked a number of servings below what nutritionists recommend and admitted they weren't cutting back on salt, sugar or fat. Most claimed to have a healthy weight, yet when pressed for detail in the same poll provided body-mass numbers indicating they were overweight or even obese.
In another poll the public agreed with nutritionists about the healthiness of many foods, including the 70% of both groups who rated hamburgers unhealthy (http://tiny.cc/…). And yet hamburgers remain among the country's most popular foods. We tell pollsters they're unhealthy, but we eat them. Oh, do we eat them.
Should we totally ignore polls, then? No, of course not. Well-constructed polls tell us many interesting and important things. It's just that they're not reliable guides to how people behave.
So pity the poor pollsters. Like farmers, they can do everything right and still fail.
Urban Lehner can be reached at firstname.lastname@example.org