Opinion: Don’t automatically discount poll predictions
“Fish gotta swim. Birds gotta fly.” Add to Jerome Kern and Oscar Hammerstein II’s timeless wisdom from “Show Boat” that before every election, “(alleged) pundits gotta predict.” You can be sure that once again, this campaign year at least one long-shot candidate whose inevitable defeat, long before Election Day, has been smugly assumed by the pundits and/or pollsters will respond by quoting President Harry Truman, the patron saint of all political underdogs, who rebutted discouraging polling numbers before his own historic upset victory in 1948: “I wonder how far Moses would have gone if he’d taken a poll in Egypt.”
Truth be told, there is some very good public polling available, including the Wall Street Journal/NBC News poll, which, let it be noted, after a pre-Election Day survey in 2016, forecast that Hillary Clinton — who would actually receive 2.86 million more votes than Donald Trump and capture 48.2 percent of the national vote, compared with Trump’s 46.1 percent — would win by 4 percentage points. Readers and viewers were openly told that poll came with a margin of error of plus or minus 2.7 percent — well within the actual election results.
The most recent WSJ/NBC News poll contains some welcome news for President Trump and Republicans: Nearly two-thirds of Americans are satisfied with the nation’s economy, one of the highest ratings since Bill Clinton left the White House in 2001. Trump’s approval rating has climbed to 44 percent, the highest it has been since the first month of his presidency and almost identical to Barack Obama’s and Ronald Reagan’s scores at similar junctures in their presidencies.
But before members of the GOP put the Champagne on ice in expectation of a big win in the midterm elections Nov. 6, they should listen to the cautionary wisdom of two professionals who have, through decades of national elections, jointly conducted that Wall Street Journal/NBC News survey, Democratic pollster Peter D. Hart and Republican pollster Bill McInturff.
Hart reminds us that in 2010, when the Republicans took away 63 Democratic House seats and reclaimed the House majority, the WSJ/NBC News poll had shown that 63 percent of Republicans were expressing a strong interest in that election campaign, compared with 49 percent of Democrats. Hart points out that Democrats in 2018 are benefiting from an “intensity advantage” — interestingly by the identical 63-49 percent margin Republicans enjoyed eight years ago — which has historically been a reliable indicator, especially in lower-participation midterm elections, of actual voter turnout.
McInturff, writing with his polling colleague Dave Wilson, has found a direct relationship between a president’s job approval rating and the actual vote for members of Congress. It’s “so strong it accurately predicts the net margin of the election.” Using the average vote of the past three midterm elections, McInturff found that consistently, 85 percent of voters who approve of the job the president is doing will vote for the congressional candidate of that party, while 83 percent of those midterm voters who disapprove of the president’s performance will vote against the congressional candidate of that president’s party. So even with his climbing 44 percent favorable rating, Trump is sporting a tank top when it comes to political coattails for GOP House candidates, which may explain why, in the latest WSJ/NBC News poll, voters, by a solid 50-40 percent margin, would prefer that Democrats control Congress.
Even with positive economic news and the promised tax cuts, doubts and misgivings about Donald Trump’s temperament and policies remain a millstone for Republicans on the 2018 ballot and a wind at the back of challenging Democrats.
• Mark Shields is a columnist with Creators Syndicate.