Medical Errors

Why is polling about health care reform so unreliable?

Quick: Do you want the government to tax your employer health benefits to pay for health care reform? When asked this question straight up, most Americans—54 percent—say they do not. But that number drops to 36 percent when people are reminded that right now, higher-wage workers get a bigger tax break than lower-wage workers. When they’re told that taxing benefits would mean that fewer employers would offer health insurance, opposition surges to 73 percent.

So the answer to the original question is anywhere from “Yes, please!” to “Hell, no.”

Public opinion is always malleable when it comes to policy. Ask someone whether they support a cigarette tax. Then ask them whether they support a cigarette tax to help kids with cancer. You get two very different answers. But health care polling is especially variable, depending on the wording, the context, and the momentary angle of the sun. Part of the reason for the polling fuzziness is that health care is such a personal issue. Everyone has or knows someone with a nightmare story, which makes emotional appeals especially effective.

But it’s also the issue’s complexity. We all know—or think we know—what a “public option” is. But do we really understand the full ramifications of government-run insurance? Would a public option drive our current private insurer out of business? Or would it be a safety net in case our employer stopped providing coverage? There’s no certain answer to these questions, and the result is that polls on the public option, not to mention the other thousand aspects of health care reform, can therefore be confusing at best, and deeply misleading at worst.

“On any issue, public opinion is a mix of well-informed to uninformed to misinformed,” says David Mermin of Lake Research Partners. “Health care happens to be a very complicated topic, with more things in the mix than usual.” His colleague Celinda Lake put it more bluntly: “If you ask stuff about comparative effectiveness, people have no idea what that means.”

Poll questions typically break down into two categories. There are questions that measure core beliefs and values. Do you like your health care plan? Do you think it costs too much? Does the system need to be reformed? In general, responses to these core beliefs questions are steady over time, and respondents know exactly what they think about them.

But then there are complex policy questions. Should we force employers to offer insurance? Should individuals be forced to buy it? Do you support a “co-op” plan as opposed to a full public plan? With these questions, responses vary widely depending on the presentation. The phrase “tax your benefits” sounds unappealing in a vacuum. But when people are told that taxing benefits could pay for health care reform and more, they’re more receptive.

As a result, health care polling can offer unhelpful data. Republicans often point to high levels of satisfaction with current health care plans as evidence that reform is unnecessary. Indeed, a whopping 83 percent of people say they are satisfied with the quality of their health care. But dig down into specifics—particularly cost, continued coverage, and the overall health care system in this country—and they have nothing but concerns. For example, many people like their health care plan but support a public option just in case they lose their job. So when people say they’re “satisfied,” argues Mark Blumenthal of, “It’s, I’m satisfied, please dear God don’t take it away.”

The data can be manipulated other ways as well. Stat fiend Nate Silver argued recently that if public opinion won the day, “Congress would pass a robust plan with a public option that was funded by raising taxes on cigarettes, booze, and people making over $250,000, and we’d live happily ever after (or not).” While it’s true that most people support a public option, and it’s true that most people support vice taxes and progressive taxation, it doesn’t necessarily follow that they would support all at once, or at all, Blumenthal says. Just because people articulate these beliefs to a pollster doesn’t mean they’re firmly held. And if the health care debate really played out in the court of public opinion, those attitudes could easily change.

What we can take away from health care polling is the limits of the debate. Even if public opinion on taxing benefits ranges from percentages in the 70s to percentages in the 30s, that’s still useful information. It shows where people begin and where they’re persuadable. “It tells leaders how hard a job they’ll have in communicating with their constituencies,” says Mollyann Brodie of the Kaiser Family Foundation. It also reveals their vulnerabilities. For example, polls show that relatively few seniors—about 30 percent—support a new government-run health care plan, since they’re already covered by Medicare. Democrats are therefore trying to figure out ways to give seniors added benefits.

Polling can also give us a preview of solutions to come. For example, Mermin says there’s “tons of support” for paying for health care by rolling back the Bush tax cuts. But few politicians have put the proposal on the table—so far.

Of course, gut reaction matters. Bill Clinton’s health care plan collapsed partly because insurance companies appealed to voters’ visceral feeling that it wouldn’t work. Politicians don’t look to polls to see what their constituents think of minor policy details. They want to see the broad brush strokes, because in elections, as in polls, you usually go with your gut.