I belong to the camp that thinks that pollsters generally have done a pretty good job in the “era of Trump” of mostly getting polls right. For all of the talk about 2016 being wrong, Hillary’s national polling average lead was 3% on the morning of the election and she won the popular vote by 2%, a margin of error difference that tilted the electoral college. In 2020, Joe Biden was averaging about 51% and got 51%, Trump simply got a margin of error bump to 47% that made the swing states closer than some hoped. If there has been a problem, it has been with down-ballot races- Congressional and state legislative- and on issue polling. It would seem that some people don’t really believe what they say they do, or something else is off.
For what it’s worth, I think we get two things wrong in polling. One is intensity on the issue we’re asking. The other is determining if an issue position is really all that important to how someone is going to vote. Basically, we do fine at figuring out what most people say they think, we do a terrible job of figuring out if they have any emotional attachment to said position.
The first question we should be asking is if you actually care about what we’re asking. John Kerry won almost every issue in the exit polling, besides being a decisive leader and keeping the country safe, and those happened to be what people cared about. Donald Trump was rarely on the right side of issue polling and yet turned out to be a very competitive national politician. Telling us you are pro-choice is great, if it actually impacts your vote. My sense is that a much higher percentage of people who say they are pro-life actually vote based on that issue than the overwhelming majority of the country that says it’s pro-choice. If something doesn’t impact your vote, it probably doesn’t matter how you feel about it. Basically, it would be nice to know how much a voter cares about something.
The second question we should be trying to answer is if you’re willing to sacrifice anything in order to achieve something. The majority of the country would say that civil rights are important and they believe everyone should have the right to vote, yet they would also tell you they achieve things like affirmative action, that require some level of sacrifice, in order to achieve those goals. Almost everyone would say fund public education, unless it means higher taxes for them. They want guns off the streets, but they don’t want to surrender their gun. A far higher percentage of the country will tell you they support the Ukraine against Russia than will say that they support the funding we are giving Ukraine in that war. National health care is a great goal that most people support, but they’re unwilling to pay higher taxes or give up their current health care to achieve it. You can see the point, I’m sure. The actual support an idea has is not the number of people supporting it, but the people willing to pay the price for it.
This is the huge caution flag I’m throwing out in front of Democrats who are feeling better about this Fall’s election than they were before. There’s a big reason why Bernie Sanders was more popular when not running than he was during the 2016 and 2020 primaries. There’s a reason Hillary Clinton’s approval was much higher when she wasn’t a Presidential candidate than when she was. People may like ideas in the abstract, but those ideas may be low priorities or simply too expensive for them to stomach when they’re real choices. The abortion issue seems to break the way of Democrats, and it’s rather clear that it has at least helped close the gap for Democrats, but believing it will win them the election is to assume a much larger percentage are voting based on that issue alone than has ever been the case before. It would seem that raw emotional appeals on the matter may have sway, but issues like inflation and crime will at least temper some of that.
“What we believe” is a very “2-D” view of who we are as voters, and unfortunately sometimes polls only show us that. I’d really like to see some polling that reflects more on what matters most, and how intense we feel about the issues. That would give us a much more complete look at the electorate, and what’s really likely to happen in this election and beyond.