The Washington Post published a poll Tuesday alleging that a majority of Americans favor the impeachment inquiry into President Donald Trump (58 percent approving and 38 percent disapproving) however the poll’s methodology is seriously flawed, which seriously undermines its credibility.
“The findings indicate that public opinion has shifted quickly against Trump and in favor of impeachment proceedings in recent weeks as information has been released about his efforts to pressure Ukrainian government officials to investigate former vice president Joe Biden, a potential 2020 campaign rival, and Biden’s son Hunter,” the Post reports.
However, the polling’s methodology problems beg the question about whether the opinion has actually shifted. The biased poll appears designed to create public opinion rather than to measure it, and it doesn’t acknowledge–as we learned from 2016–that public polls have been notoriously bad at accurately tracking Americans’ opinions of President Trump. If you read the wording carefully, it’s clear the Post’s questions are written from a Democratic point of view and are designed to get the results and headlines the left-leaning Post was looking for.
The Post didn’t ask any questions from the Republican point of view, such as whether there should there be an investigation of Joe Biden’s pressure on Ukraine to fire the prosecutor investigating the company his son worked for? Or should Congress vote publicly on allowing the impeachment inquiry to move forward? Or should Congress move forward with an impeachment inquiry or should the people be allowed to decide in the next election?
Critics also note that this is a poll of “adults” not registered or likely voters, which could mean a leftward skew. The poll under samples Republicans badly compared to actual 2016 turnout (25 percent in the poll, compared to 33 percent turnout in 2016). The Post’s polling also oversamples Independents even worse compared to actual 2016 turnout (44 percent in the poll compared to 31 percent actual in 2016)
The Post polling was also a random digital dial sample, meaning the Post actually had no idea who they talked to.