Thursday, July 14, 2016

Rollercoaster polls take voters for a ride -- July 14, 2016 column


A headline in Politico this week read: “Swing-state stunner: Trump has edge in key states.” The only thing missing was an exclamation point.

Commentators online and on TV chewed over the news that three Quinnipiac University polls found Donald J. Trump slightly leading Hillary Clinton in the battleground states of Florida and Pennsylvania. The two candidates were tied in Ohio.  

The next day, a headline across a full page of The Wall Street Journal read: “Polls: Clinton, Trump Close in Key States.” Clinton and Trump were in a statistical tie in Ohio; she had a 3-point lead in Iowa and was ahead by 9 points in Pennsylvania, the latest Journal/NBC News/Marist polls found. 

And The New York Times reported the same day that Clinton and Trump were tied nationally, each with 40 percent of registered voters, in the latest Times-CBS News poll. 
But a different national poll a few days earlier had showed Clinton with a double-digit lead over Trump. Yet another put Trump ahead by single digits.

A reader could get a headache trying to parse the polls. But do polls matter? Not really. Not in July.

Everyone needs to remember that polls are a snapshot in time. If there’s anything we know about this presidential campaign, it’s unpredictable. 

Yes, Democrats would rather see Clinton on a positive trajectory, leaving Trump in the dust. And Republicans would like to see Trump steadily gaining ground on Clinton, although so far, while she seems to be sliding, he’s not rising.

But neither camp should get too exercised about polls this far out. They rarely predict Election Day.

Better to sit back, take a deep breath and ponder how Britain can change prime ministers in days while our presidential elections drag on for years. Here’s a poll tidbit that rings true: Six in 10 Americans are worn out by the presidential campaign.

Part of what’s driving the poll frenzy is news organizations’ trying not to miss the Trump story – again. Many political reporters -- I include myself -- thought Trump was a flash in the pan. Obviously, we were wrong.

But whether Trump or Clinton wins in November, some pollsters will be able to say they saw the incipient victory during the summer.

That’s fine, but voters need to know that analysts can’t even agree on polling methods.

Some analysts fault Quinnipiac, contending its sample size favors Trump by including larger percentages of white people and fewer minorities than voted in various states in 2012. Since minority voting is rising and white participation falling, Quinnipiac’s polls are biased, these critics say. We won’t know who’s right for nearly four months.

Naturally, Trump brags about his positive poll numbers and discounts those he doesn’t like. The Clinton campaign tweets that it always expected battleground states to be tight, and supporters just have to work harder.

When it comes to polls, though, you ain’t seen nothin’ yet.

After the Republican National Convention concludes July 21 and the Democratic convention wraps July 28, we’ll be bombarded by polls. The conventions traditionally bring the largest swings in polls during the campaign, say political scientists Robert S. Erikson and Christopher Wlezien. They studied polls from 1952 to 2008 for their 2012 book “The Timeline of Presidential Elections.” Traditionally, first, one party’s candidate gets a bounce and then the other.

In simpler times, most voters were just learning about presidential candidates by watching convention coverage on TV, and the conventions were spaced weeks apart. 

The exposure traditionally gave the nominees an average 5-point increase in the polls, Gallup reports, but the convention bounce has declined since 1996.

In 2012, a year like this one with back-to-back conventions, Republican Mitt Romney saw a 1-point dip after the GOP convention, and President Barack Obama got a 3-point bounce after the Democratic convention. Polls tightened by Election Day, as they usually do.

As always, the people who cast ballots Nov. 8 are the only poll that matters. And there’s something else to consider: One in 10 voters for both Clinton and Trump say they could still change their minds before Election Day.

©2016 Marsha Mercer. All rights reserved.

No comments:

Post a Comment