What the polls are getting wrong about Trump
- Oops!Something went wrong.Please try again later.
Poll after poll continues to show former President Donald Trump with substantial leads over his primary opponents. The dozens of criminal charges brought against the former president have seemingly not harmed his standing in the Republican presidential primary but appear to be boosting his candidacy.
Russell Berman, writing for The Atlantic, says that the narrative that multiple indictments are helping Trump may be premature.
Berman writes about a new study, arguing that “pollsters have been asking the wrong questions” when trying to measure how the indictments are affecting Republican voters. Most traditional polls, researchers say, “have asked respondents directly whether the indictments have changed their attitude about Trump or their likelihood to vote for him.”
Matt Graham, one of the authors of the new study and an assistant professor at Temple University, says this type of query leads to biased answers. Republicans “want to say, ‘Well, I still support him regardless of the indictment.’ And if you don’t give them a chance to say that, they’re going to use the question to say that.”
Related
Graham and his fellow researchers reframed their questions and asked half of the respondents to a SurveyMonkey poll, “Suppose you did not know about the indictment. How would you have answered the following question: How likely are you to vote for Donald Trump?” They asked the other half questions that pollsters more commonly use.
The experiment produced statistically significant differences. The poll that asked the commonly used questions found, like other polls, that the indictments increased Trump’s support among Republican primary voters. However, the poll based on the counterfactual framing found that the indictments slightly reduced the chances of a Republican voting for him.
Even though statistically significant, the result is still small, reducing the likelihood that Republicans would vote for him by just 1.6 percentage points.
Graham told the Deseret News that although 1.6% is within the typical margin of error, the study’s estimates are more precise, partly because of the large sample size. They calculated statistical significance based on the same rule of thumb as the margin of error but were able to increase the specificity. The margin of error in the Graham et. al study is 1.2%.
Berman writes that “A drop of 1.6% suggests that charging Trump with multiple felonies is akin to tossing a pebble at a fast-moving train.” Berman also writes that “The dynamic has turned an infamous example of Trumpian bravado — his 2016 claim that ‘I could stand in the middle of 5th Avenue and shoot somebody and I wouldn’t lose voters’ — into something approaching a prophecy.”
Researcher Graham isn’t arguing that his team’s findings should fundamentally alter perceptions about Trump’s chances of becoming the Republican nominee. Even if the difference in Trump’s support is considered statistically significant, it’s also not big enough to matter. “The indictments clearly didn’t put much dent in Trump’s lead,” Graham said.
But, the false narrative that indictments actually boost a candidate is an important one to dispel. “I don’t think that survey researchers should be sending the public profoundly pessimistic messages about how their fellow citizens think and reason when those aren’t actually true,” Graham told Berman. “There’s plenty to be pessimistic about in our politics, but we don’t need to pile on by acting like people think that indictments are good.”
