What's Going On With Those Ohio Polls?

Discussion in 'Politics & Religion' started by pspr, Oct 25, 2012.

  1. pspr


    By Josh Jordan

    The race for Ohio is slowly tightening, but Mitt Romney does not hold a lead in a single poll in the current Real Clear Politics average (he is tied in two). Two polls from Time and CBS/Quinnipiac have grabbed headlines by showing Obama a five-point lead in each. Romney is chipping away at Obama’s poll lead, but the Democratic advantage in party-ID has increased across these polls. When looking at the polls in Ohio, it is becoming entirely possible that Mitt Romney should be able to win Ohio without ever showing a consistent lead in the polls, or any lead at all.

    In the past week Romney has trimmed four-tenths of a point off of his deficit in the RCP average, going from 2.5 to 2.1, but at the same time, the average party-ID advantage for Democrats in these polls has risen from 5.5 to 6.5. A big reason for the increase in Democrats’ share in the polls is due to early voting. If a pollster calls someone who says they voted already, they are automatically passed through the likely-voter screen since they have, after all, voted. The problem with this can be best summed up by Gregory House: “Everybody lies.”

    Pollsters can only work with what their respondents tell them, and this is the reason that likely-voter screens can be so tricky, though important, in polling. The preferable response is that you are going to vote or, in the case of Ohio, that you’ve already voted. Many respondents will say they are going to vote (or have voted) when in fact they may not end up doing it (this effect is known as social-desirability bias). For this reason, some likely-voter screens ask about previous elections and general political enthusiasm to gauge the actual likelihood that a voter will end up in the booth on Election Day. But that is where early voting throws the screen out the window — if a voter says they voted, there is nothing a pollster can do to but assume that it’s true.

    Enter Ohio, where the current estimates from compiling early in-person and absentee voting shows early turnout to be about 15 percent of voters. But responses in the current polls claim that 23 percent of registered voters have already voted. That means that polls are overstating early voting by eight percentage points on average. This could be in part because some voters have requested an absentee ballot and report that as voting, some have mailed in ballots that haven’t been counted as received yet, but some voters are also just flat out saying they voted when they haven’t. It’s impossible to know the exact reason, but it’s clear that more are claiming to vote than really have.

    In the polls’ early-voting results, Obama leads on average by 20 points. There are indications that the GOP has shrunk the Democratic advantage in this category significantly from 2008, but it is unclear how much. Either way, Obama’s early-voting advantage gives him a lead that Romney is only scraping away at with his Election Day voter lead. But if pollsters are finding more respondents who are claiming to have already voted than what the records show, some of this early-voter advantage is illusory.

    This is why it is increasingly difficult for Romney to show an lead in the Ohio polls. But even with Obama currently enjoying a 2.1 point lead, Romney is still in great shape to win Ohio on Election Day. Here are some of the reasons for the optimism coming from Boston these days:

    Romney’s strength with independents keeps growing: Last week when Obama led the Real Clear Politics average by 2.5 points, Romney led among independents by an average of 8.7 points. Romney has since increased that lead with independents to 12.3 points, which is why he’s been able to cut Obama’s overall lead even as the polls have leaned more Democratic. In 2008 Obama beat McCain with independents by eight points. It would be almost impossible for Obama to win Ohio while suffering a 20-point swing among independents.

    The polls give Democrats a better turnout advantage than they had in 2008: As I explained in my last Ohio post, in 2008 Democrats beat Republicans in turnout by five points. The current polls show an average of D+6.6. A D+5 turnout in 2008 gave Obama a 4.5-point victory, while he is currently leading by only 2.1 points on an even greater D+6.6 turnout. Again, we know it should be very difficult for Democrats to match their 2008 turnout, let alone increase it.

    History suggests late deciders will break against the incumbent: This is a rule that always receives some skepticism, but it’s very likely to benefit Romney at least some on Election Day. In 2004, late deciders broke against George W. Bush heavily, even though he was a wartime president. John Kerry beat Bush by 25 points among voters who decided in the last month, 28 points among voters that decided in the three days prior to Election Day, and 22 points among day-of deciders. Those voters were 20 percent of the Ohio electorate; while this year there are expected to be fewer late deciders, Obama cannot afford to lose among by those margins and still win.

    In Ohio, Republicans tend to outperform their share of the national vote: In the last nine elections, the GOP has outperformed in Ohio. With Romney currently running just ahead of Obama nationally, it seems much more likely that Obama’s lead in Ohio has more to do with the higher party-ID advantage than a dramatic shift in Ohio from the past nine elections.

    Strength with crossover voters in Ohio: In addition to Romney’s strength with independents, in the past two elections the GOP candidate has won over more Democrat votes than he’s lost Republican ones. Obama’s Ohio win in 2008 was based entirely on his strength with independents and the wave turnout, both of which are highly unlikely to be repeated in 2012. If Romney wins with independents by anywhere near the current average he has and takes more crossover voters than Obama does, Obama would need to exceed 2008 turnout greatly to win.

    So, with less than two weeks until Election Day we will all know the results soon enough, but as more Ohio polls come in, it is important to remember that the picture for Romney in Ohio is better than many pundits would have us believe. It only takes a quick look at Romney’s rallies to remind us it’s not 2008 anymore, as Republicans have reclaimed the enthusiasm advantage that led to such sweeping 2008 victories for Democrats. That GOP enthusiasm has become contagious since the debates, and it is exactly what has Team Obama so afraid these days. All they have left to hang their hopes on is a slim lead in the polls, and even that might not be enough on Election Day.

  2. Tsing Tao

    Tsing Tao

    Was a very interesting read. This morning, I was going through the polls and noticed massive differences in the amount of Democrats vs. Republicans they were considering. I was stunned at what passed for acceptable polls.

    Here is the latest Time poll.


    Note that it is using 290 democrats to 220 republicans (with Independents even higher that Republicans by 4). Really? Ohio was won by Obama over McCain by only 2.1 points. The fact that Obama can beat someone so lame as McCain by 2.1 points doesn't bode well for polls like Time, which is using a 10 point advantage???. Additionally, all the fanfare of electing the first black President doesn't exactly hold water this time around. Then there's that pesky economy issue that just won't go away.

    Here is the other recent poll showing Obama with a lead, by SurveyUSA:


    This poll is a 7 point advantage for Democrats over Republicans. What? Let me repeat. Obama beat McCain by 2.1 points in 2008. There is something inherently wrong with this methodology.

    Please note that if I saw a poll favoring republicans that far outside of the norm, I would absolutely call it out in the same manner.
  3. pspr


    Tsing, it seems at least some of the pollsters are drawing their weighting from their early voting questions to their sample. Since democrats seem to have an advantage here and it seems many don't want to admit they didn't vote for Obama it is skewing the polls too far toward Obama.

    Jem has been pointing out for some time a bias in the polling that seems to defy reason. I think you and the article I posted have valid points that things are looking better for Romney in Ohio than the polls are reflecting.
  4. Tsing Tao

    Tsing Tao

    I don't pretend to be a polling expert. I didn't even like Statistics in college. But something just doesn't seem to pass the mustard here, and I would happily change my thought process if someone could provide a countering opinion.

    Additionally, I have been looking for the sampling info on the Rasmussen poll, but cannot find it. Does anyone have it?
  5. pspr


    I remember seeing it in one poll about a month ago. It also oversampled dems but just slightly. Since then I haven't seen it. It may be provided only to subscribers to his polling data.
  6. Tsing Tao

    Tsing Tao

    It's hard for me to accept the Rasmussen polls on blind faith if I can't see their sampling.