It’s kind of crazy to think that the 2016 election is already a year away, and we’re already seeing multiple polls in the U.S. Senate race before the second campaign finance deadline (and six months before the candidates can even file petitions to get on the ballot). Two days ago, Public Policy Polling (D) released a poll showing incumbent freshman Republican U.S. Senator Rob Portman with a narrow, two-point lead against former Democratic Governor Ted Strickland (43%/41%). Unlike an earlier Quinnipiac Poll, which showed Strickland attracting more Republican crossover votes than Portman attracting Democrats, PPP showed the opposite with Portman attracting 15% of Democrats to 9% of Republicans for Strickland.
Portman has a horrible 35% approval rating which ties his disapproval rating. Historically, any incumbent with an approval rating this low and below 50% in a head to head, even a year out, is considered in trouble. But with an $8 million warchest and a year to go, a lot can change. The question is whether that warchest is big enough to protect Portman if the presidential race favors the Democrats as it appears to be right now.
PPP shows the Democratic primary as a non-contest, with Strickland crushing Sittenfeld by 52 points (65% to 13%). Sittenfeld’s favorability rating is actually underwater. Sittenfeld is unknown to 69% of Democratic voters alone. 18% of Democrats have an unfavorable opinion of him to 13% holding him favorability, so he’s -5 underwater on that question alone. Clearly, if PPP is to be believed, Democratic voters have not taken a shining to his primary challenge. Strickland has a 60% favorability rating to 23% unfavorability among Democrats. Just like Quinnipiac, PPP shows Sittenfeld getting crushed by both Strickland and Portman by sizeable margins (although PPP showed only a 18 point margin for Portman with him still under 50%).
In a laughable attempt to paint a different picture, Ohio Republican Party Chairman Matt Borges pushed on social media this afternoon a new poll by Vox Populi Polling/Daily Caller that claims to show Portman with a ten point lead over Strickland. If you’re not familiar with Vox, you’re not alone as they haven’t really polled in Ohio before, and it’s a new pollster on the scene. It was founded last year by Mary Cheney and other Republican activists in order to, as Politico said at the time, “with the mission of helping Republicans compete in the field of automated polling.”
In other words, to give Republican candidates the potential to have polls to generate more favorable news of their campaigns. And any objective look at this poll shows that is what it’s attempting to do. First, the poll was conducted only over the weekend, which most pollster avoid doing because a) it’s REALLY hard to get people on the phone to poll, b) even harder when, like Vox, you use automated dialing prompts to conduct your polls instead of live interviewers, and c) it’s really tricky to weigh your results to make people answering the phone on a Friday night be representative of someone who is not.
The poll had only a sample size of 474 “active voters” which it defines as voters who voted in 2010 or 2012 or “registered since the 2014 election.” Anyone see the problem there? The sample only partially includes the last presidential election (which tends to be more Democratic) and includes an off-year election that is more Republican (in fact, arguably one of the most Republican turnout elections ever in Ohio). Whatever the 2016 turnout may be, it will probably look more like the 2012 turnout than anything resembling the 2010 turnout.
A quick review of their crosstabs shows how this sampling method led to a likely nonrepresentative sample. Their sample, presumably even after weighing, was only 9% African American, even though African Americans make up 12% of the Ohio population and were 15% of the electorate that voted in 2012. So, this sample underrepresents the African American turnout in Ohio in the last presidential election by 6 points. It under represents Democratic voters to the benefit of Republican voters by around 5 points. Those are ridiculous deviations.
On top of that, the poll asked a rather convoluted rating system of Strickland’s record as Governor, and then a question about Strickland’s age before asking the Portman/Strickland head to head question. The only question besides a head to head against Sittenfeld that involved Portman was a standard name recognition question. It never questioned voters as to Portman’s favorability or his job approval. After asking two questions on Strickland only, it then went directly to ask voters how they’d vote. I won’t go so far as call it a push poll, but I believe any pollster would question this practice as compromising the integrity of the head-to-head results. It asked no questions about Portman or Sittenfeld. By the way, even this slanted polls shows most voters don’t care about Strickland’s age.
Races involving an incumbent are normally a referendum on the incumbent, not a referendum on the last elected office the challenger held. I cannot recall ever seeing any pollster, in any race, ask to rate the approval of a challenge only, especially in an office he ceased holding five years ago. I get that’s Portman’s strategy because he cannot defend his record, but I’ve never seen a pollster not ask any question testing the approval rating of an incumbent, unless this outfit is not actually releasing all of the data it collected.
There are lies, damn lies and statistics. Polls a year out from an election out don’t amount to much. But if the Ohio Republican Party is already reduced to pushing Mary Cheney’s poll from a firm she created for the express purpose of trying to slant the news for GOP candidates, they’re clearly nervous. They know that Rob Portman is not well known, has not done much in his first term in the United States Senate, and he’s taken a number of unpopular, partisan votes and actions (Cotton letter and holding up the Cordray nomination) that will hurt him as the campaign really gets started next year.