http://dyn.politico.com/printstory.cfm?uuid=6B977FDC-6CD2-44C9-8DD8-F100929CC527 Who else had a bad night? Pollsters
By: Steven Shepard
November 5, 2014 01:26 PM EST
The polls were wrong.
And so were the predictions about how the polls would be wrong.
Instead of being biased against Democrats, who still hoped they could overcome modest deficits in a number of states, the surveys of the 2014 Senate landscape turned out to have significantly underestimated Republicans.
The GOP won resoundingly in races where it led only narrowly in the polls, and the party also put into play contests its own strategists thought would be out of reach. Democrats, who had insisted the polls were stacked against them, were crushed in races they thought were competitive.
The results were another black eye for pollsters in what are already some tough times. Just five months after then-House Majority Leader Eric Cantor’s surprise ouster, it was another out-of-nowhere Virginia race that left political observers scratching their heads.
And as Americans become even harder to reach by phone – and emerging methodologies, such as Internet polling, remain unproven – the poor performance of pollsters this year casts serious doubt on the reliability of surveys during the 2016 presidential race.
Republicans nearly swept the 10 Senate races rated as toss-ups by the Cook Political Report on Election Day, winning seven of the 10 outright – they lost New Hampshire, Alaska remains undecided and Louisiana is going to a runoff.
To be sure, Republicans were favored to win back the Senate on Tuesday. They led in six of the 10 contested races – and hold the advantage in a seventh, Louisiana, in the runoff next month.
But it wasn’t that Republicans won so many of the most competitive races — it was how much they won by.
Tom Cotton (Ark.), David Perdue (Ga.), Joni Ernst (Iowa), Pat Roberts (Kan.) and Mitch McConnell (Ky.) all blew out their opponents, despite polls showing much closer contests.
Republicans have long claimed that public polls, usually conducted by randomly dialing phone numbers rather than only contacting voters with a history of turning out in midterm elections, include too many people who won’t ultimately cast a ballot – a group that tends to lean Democratic.
Those public surveys, they say, also weight, or peg, their demographic data to known Census parameters, ignoring historical trends of the midterm electorate – which is usually older and more white.
“I think the media polls were dramatically off because too many media pollsters use Census weights,” said Republican consultant Brad Todd, whose firm, OnMessage Inc., conducts polls and creates TV ads for GOP candidates. “In a midterm electorate, using the Census as a reference point would have the same value as using a grocery list as a reference point.”
In Arkansas, Cotton beat Democratic Sen. Mark Pryor by a 17-point margin. That capped a startling end to Pryor’s once-charmed political career – Republicans didn’t even run a candidate against him six years ago. The breadth of Cotton’s victory was all the more impressive when stacked against the freshman GOP congressman’s 7-point lead in the final RealClearPolitics average on Election Day. (Credit goes to the University of Arkansas, whose 13-point spread for Cotton represented the only survey to show him leading Pryor by more than 8 points.)
In Georgia, Perdue cleared the 50-percent threshold to avoid a runoff and romped over Democrat Michelle Nunn, 53 percent to 45 percent. But the final RCP average showed Perdue with just a 3-point lead in the polls.
While polls didn’t project Perdue’s margin, they did show him closing in the final weeks of the campaign. Nunn was tied or led in seven consecutive mid-October polls, but Perdue claimed the late momentum, leading in nine of the final 10 polls. Still, no survey showed him with a lead larger than four points.
Iowa’s Joni Ernst was yet another Republican underestimated by the polls: The final RCP average had her up by 2.3 points, far lower than the 8.5-point margin by which she beat Democratic Rep. Bruce Braley.
The average is a little misleading, however. Most polls showed the race tied, or Ernst with the slightest of edges. But the Des Moines Register’s Iowa Poll, conducted by West Des Moines-based pollster J. Ann Selzer, put Ernst up by a 7-point margin, 51 percent to 44 percent. Ernst won, 52 percent to 44 percent.
Democrats pilloried Selzer’s poll, pointing out – with some justification, given the other seven surveys conducted over the final week showed results that ranged from Braley ahead by 1 to Ernst up by 3. But Selzer, who has polled for the Register for decades and most recently accurately predicted President Barack Obama’s victory in the 2008 Iowa caucuses, proved again why she is the gold standard in the Hawkeye State.
Kansas was always a difficult dynamic for pollsters. Republicans haven’t lost a Senate election there since the Great Depression, and independent Greg Orman — who doesn’t have a natural constituency — presents a unique challenge.
But the polls, both public and private, missed badly. Orman entered Election Day with a negligible, 0.8-point lead in the RCP average. Not only did GOP Sen. Pat Roberts win by nearly 11 points, but Republican Gov. Sam Brownback – whom some GOPers had left for dead – also won reelection.
Polls in Kentucky had moved toward Mitch McConnell in recent weeks, but no survey showed the likely new Senate majority leader up double digits, let alone leading by the 15.5-point margin by which he beat Democrat Alison Lundergan Grimes on Tuesday.
In Louisiana, the RCP average showed Democratic Sen. Mary Landrieu 5.7 points ahead of GOP Rep. Bill Cassidy. While both candidates still made the runoff, Landrieu’s margin was just one point.
Polls in three other competitive Senate races – Alaska, Colorado and North Carolina – also exhibited a Democratic bias. While the difference was less dramatic, Sen.-elect Thom Tillis’ (R-N.C.) victory is considered a modest surprise because of Democratic incumbent Kay Hagan’s slight lead in the polls.
In only one of the 10 most competitive Senate races were polls too optimistic for Republicans: New Hampshire, where Democratic Sen. Jeanne Shaheen won with a larger cushion than polls had indicated.
Virginia wasn’t on anyone’s watch-list going into Tuesday, with polls — outside of a survey from GOP robopollster Vox Populi — showing Democratic Sen. Mark Warner with a comfortable lead. Warner had a lead of 9.7 points in the final RCP average; he’s currently up by less than a point as a possible recount looms.
Part of the problem with Virginia — and neighboring Maryland, where Republican Larry Hogan was the upset winner of the state’s gubernatorial race — was an overall lack of public telephone polling. Two colleges polled Virginia Senate in the final two weeks, finding Warner ahead by 12 and 7 points.
Virginia was the scene of the most shocking upset of the entire cycle, when Cantor went down — by a lot, and to the humiliation of his unprepared campaign — to virtual unknown Dave Brat in the GOP primary. Brat easily defeated Democrat Jack Trammell in the conservative district in Tuesday’s general election.
In Maryland, The Washington Post and Baltimore Sun surveyed the race in early October, finding Brown ahead in the high single digits. But the only other phone polls came from the state GOP or Hogan’s campaign.
Pollsters had the opposite problem in 2012, when the polls understated Democratic support in the national presidential race and many of the battleground states.
There were some bright spots for the industry, especially from in-state pollsters such as Selzer, who nailed the ultimate results close to home.
In the Wisconsin governor’s race, Marquette University Law School raised eyebrows when it showed GOP Gov. Scott Walker pulling away from Democrat Mary Burke, opening up a 7-point lead in the final week of what had been a neck-and-neck. Walker won by 5.7 points.
