Mark Mellman: About those Kaiser polls

A set of New York Times/Kaiser Family Foundation polls on Southern Senate races and ObamaCare created quite a brouhaha last week, eliciting a level of vitriol not seen since the “unskewers” rose up in reckless disregard of reality during the presidential race.

At some level, the invective is hard to understand. These polls lead to the same macro-conclusion most others have reached about these races: that anyone who writes off the chances of the Southern Senate Democrats at this stage is foolish. Clearly these polls reflect the same reality GOP groups see. Otherwise, they would not be continuing to pour millions of dollars into negative ads in these states.


Moreover, these polls were right in line with several others in at least three of the states polled. Since Feb. 5, other polls were conducted in North Carolina showing Democratic Sen. Kay HaganKay Ruthven HaganThe Hill's Campaign Report: Democratic Unity Taskforce unveils party platform recommendations Democrats awash with cash in battle for Senate The Hill's Campaign Report: Trump's job approval erodes among groups that powered his 2016 victory MORE between 4 points ahead and 2 points behind. A Republican firm gave her a 3-point lead, while, on average, the five earlier polls showed her 2 points ahead, exactly the margin reported by the Times/Kaiser survey.

Polling in Louisiana’s “jungle primary” is a bit less straightforward, but suffice it to say, the Times/Kaiser survey put Democratic Sen. Mary LandrieuMary Loretta LandrieuBottom line A decade of making a difference: Senate Caucus on Foster Youth Congress needs to work to combat the poverty, abuse and neglect issues that children face MORE at exactly the average of the other two public polls (one by a GOP firm) that have been released this year.

In Kentucky, a January Courier Journal/ SurveyUSA poll actually put Alison Lundergan Grimes ahead by 4 points, while Times/Kaiser gave Republican Sen. Mitch McConnellAddison (Mitch) Mitchell McConnellProgressive group launches M pro-Biden ad buy targeting young voters Pelosi signals flexibility on size of renewed unemployment payments Lincoln Project reports raising .8 million for anti-Trump efforts MORE a 1-point lead. (Full disclosure: Landrieu and Grimes are clients.)

Only in Arkansas were the Times/Kaiser results materially different from others. In the Natural State, Democratic Sen. Mark PryorMark Lunsford PryorCoronavirus poses risks for Trump in 2020 Tom Cotton's only Democratic rival quits race in Arkansas Medicaid rollback looms for GOP senators in 2020 MORE enjoyed a 10-point lead, despite having been ahead by 3 points, or tied, in two previous polls and being behind earlier. Both Pryor and his opponents have been on television for months in the race, though, so it’s not surprising we would see continued movement.

Listening to the GOP “unskewers” you’d think the polling units at The New York Times and the Kaiser Family Foundation are run by partisan hacks and incompetent boobs. They aren’t. These expert professionals enjoy impeccable reputations.

What really seemed to annoy Republicans, who uttered almost not a word about the horse-race results, was the recalled presidential vote. As my GOP colleagues Glen Bolger and Neil Newhouse noted early on, those results were crazy wrong. President Obama did not win Louisiana by 3 points or North Carolina by 7, nor did he only lose Arkansas by a single point or Kentucky by merely 3.

Taken alone, those numbers might suggest badly skewed surveys. But they should not be taken alone. First, asking people to recall their vote months later is always fraught and often yields results strikingly at odds with reality, and usually to the benefit of the winner. Answers to these recall questions are off so frequently, and by such magnitudes, that we urge clients never to even ask them.

More important, other numbers that should be correlated with underlying partisanship look quite reasonable. For example, according to the Times/Kaiser survey, the president’s approval in Louisiana is 40 percent. Averaged over all of 2013, Gallup polls had it at the same 40 percent. Gallup pegged Obama’s approval at 35 percent in Arkansas, while Times/KFF had a slightly more conservative 32 percent. In North Carolina, Times/KFF pegged presidential approval at 41 percent, Gallup at a slightly higher 43 percent, averaged across the whole years’ worth of interviews.

So, in the end, we have horse-race numbers that generally comport with others’ findings, political questions that comport with others’ findings and one question that seems far from reality and often is. In short, we have no real evidence these polls are flawed and a fair amount of evidence that the criticism, while filled with sound and fury, doesn’t signify much.

Mellman is president of The Mellman Group and has worked for Democratic candidates and causes since 1982. Current clients include the majority leader of the Senate and the Democratic whip in the House.