The science of polling has taken some hits over the last few years with everything from Republican attempts to “unskew” them, to the way they’ve been blamed for being wrong about the 2016 presidential election. I know that during election years I can get enamored with poll-reading and have been working on catching when I focus in on the ones I like and disregard the ones I don’t. But given that I actually like math, there are a couple of things I always keep in mind.
The first is that I tend to dismiss the results of one particular poll and tend to look for trends over time—specifically when there are changes from one pollster who asks the same question and implements the same methodology. In other words, if Gallup has been asking about Trump’s job approval over the last seven months and we all of the sudden see a drop that is greater than the margin of error, it’s worth paying attention to. Otherwise, it’s probably noise.
When it comes to polling on issues, I almost always ignore those. The main reason is that a lot of people simply provide an answer that they think is socially acceptable or in line with their partisan leanings. But we’ve just had a recent example of the fact that how a question is worded makes all the difference.
You might have heard that in a recent NPR/PBS/Marist poll, 62 percent said that statues honoring the leaders of the Confederacy should remain as a historical symbol. A lot of people assume that means that Trump’s position has a lot of support. But as Paul Waldman wrote, it’s all in the wording they used to ask the question.
But the truth is that the question asked in that poll could barely have been designed better to produce that result (though I’m not saying they did so intentionally). It asked whether “statues honoring leaders of the Confederacy should remain as a historical symbol, or be removed because they are offensive to some people.” That posits a choice between value-neutral “history” on one hand, and “some people” who are offended on the other. In other words, a reasonable understanding of our past, or a small group of censorious, politically correct agitators?
That was yesterday. Today, Public Policy Polling released their most recent national poll. It included the following question.
Would you support or oppose relocating monuments honoring the confederacy from government property and moving them to museums or other historic sites where they can be viewed in proper historical context?
The results are that 58 percent supported that idea and only 26 percent opposed it.
This is why it is a fool’s errand for political parties and politicians to craft an agenda based on one particular poll. The results of these two flipped based on how the question was asked. In writing about single payer, Joshua Holland pointed to an even bigger problem with crafting an agenda based on polls.
Don’t be lulled into complacency by polls purporting to show that single payer is popular—forcing people to move into a new system is all but guaranteed to result in tons of resistance. And that’s not even considering the inevitable attacks from a conservative message machine that turned a little bit of money for voluntary end-of-life counseling into “death panels.” Public opinion is dubious given that nobody’s talking about the difficulties inherent in making such a transition.
As Democrats explore how to go forward in the future, they should craft solutions that can be demonstrated to solve the problems Americans are facing—and take it to the people. They shouldn’t rely on issue polling to make their case.