Polling has been struggling with an image problem following some high profile pre-election problems in recent years, so it’s little surprise that more than half of Americans don’t trust public opinion polls.
A new McClatchy/Marist poll finds that only 37 percent of registered voters have a great deal or a good amount of trust in public opinion polling. Only seven percent say they have a great deal of trust in comparison to 22 percent of voters who don’t trust polling at all and 38 percent who don’t trust it very much.
Democrats have more confidence in public opinion measures than other groups, with 47 percent saying they have “a great deal” or “a good amount” of trust in opinion polls. Only 26 percent of Republicans agree, though, with 73 percent saying they don’t trust polling “very much” or even “at all.” Independent voters echo the overall views.
It’s understandable that many people wouldn’t trust public opinion polls after the run of election polling errors ― or perceived errors ― in recent years.
Critics point to Brexit polls, which mostly indicated that staying in the European Union would eke out a narrow victory in the U.K. referendum last year, and to the U.S. state-level polls that failed to indicate President Donald Trump would win in enough key states to take the Electoral College majority and win the presidency.
In light of these recent issues and known difficulties in identifying who will turn out to vote ― a key source of uncertainty in the election polling enterprise ― more skepticism by the public and the media regarding election polls is probably warranted.
But even with extra skepticism about election polls, there’s no other measure that can give us a broad view of what the public thinks. Plus, most polls on public policy don’t have the same issue of trying to identify who will vote, so measuring opinion in this case is much simpler.
Of course, reporters and politicians can talk to people anywhere, but the people they talk to won’t be representative of all Americans.
Only public opinion pollsters go through the necessary steps to contact a sample that is as close as possible to representative of the American population (or Americans who are registered to vote, as many polls place their focus on this population for political matters). That means randomly sampling people from all over the country, or getting people from all over the country to sign up for “panels” and carefully selecting a representative sample out of those groups to complete the survey.
Polls, when done carefully, are pretty good representations of American opinion ― with some room for error, of course, since they’re based on samples of a few hundred or a thousand people. Additionally, they can be tested against and weighted to known benchmarks, such as census data to check their representativeness.
Yet, because of election issues, and because people choose what they prefer to believe based on their own experiences and ignore that other perspectives exist, people would rather not trust the polls.
But consider that the president of the United States reportedly still consumes polls as much as he did during the election. Of course, he often dismisses polls that don’t look good for him as “phony” or “rigged,” but the fact that he looks at the data means he thinks polls are indicative of opinion. It also means that Americans have a direct avenue to getting their voices heard by the president.