On October 13, Public Policy Polling (PPP) released an Ohio survey finding that a shocking 19% of Ohioans had already voted – and they voted for Obama by an overwhelming 76-24 margin over Mitt Romney. Not surprisingly, many of us “poll truthers” questioned the likelihood that almost 1.1 million voters had already cast ballots in Ohio in such a short period of time (there were a total of 5.7 million votes altogether in 2008). After all, that was about triple the 5-6% ballot cast rate shown by most counties at the time. But logic aside, if PPP’s figures were correct this would have meant that Obama had locked in about 823,000 votes, while Romney was limping along with just 259,000. Boom. Game over Republicans.
Overall, Obama was given a 51-46% lead over Romney, almost all exclusively as a result of the disproportionately pro-Obama early vote. This helped drive the media narrative about Obama’s “big lead” over Romney in Ohio and how effective the Democratic early voting machine was. Again, all of this ran contrary to actual, real evidence about ballots requested and returned. But hey, why let actual, real, reported numbers get in the way of a “poll” of 165 early voters? Besides, according to this poll Ohio was well on it’s way to having 100% turnout before election day. Why would this seem odd?
Fast-forward a week. On October 20th, PPP released a follow-up poll in Ohio that showed some odd data, especially if the poll taken a week earlier was also to be believed. Apparently, no one felt like voting early this week as the percentage of early voters in their poll only increased from 19% to 21%, a shocking drop off for such a highly competitive state. If the polls were right, we would have to believe that 1.1 million peoplevoted in the first 12 days in Ohio, and then barely over 100K over thenext 7 days. (And remember, we are the ones ridiculed for questioning all of this).
But here is where their scientific, indisputable, unquestionable data really gets interesting. Remember that 76-24% early vote lead Obama had last weekend? This week, that fell to 66-34%, a 20 point swing in Romney’s direction. And when you do the math, Obama’s “locked in” early vote dropped from 823,000 votes last weekend, to 790,020 this weekend. Romney went from 260,000 votes locked in to 407,000 over the same time. Remember, PPP was measuring votes that are already cast. These people can’t change their minds, so such an excuse doesn’t fly. But if we trust PPP, Romney not only got 100% of every vote cast over the last week, but another 30,000 people walked into their local voting location and changed their already cast ballots from Obama to Romney.
Overall, the two polls taken together require another jump in logic. First, Obama’s lead drops from 51-46% last week to just a 1-point, 49-48% lead now. A 4 point swing might not seem crazy, but how we get there is: in order to maintain this narrow-lead, the PPP sample jumps from D+4 a week ago – when Obama was up 5 – to a not-going-to-happen D+8 today. Had the sample remained a steady D+4 week-over-week, Romney would have opened up a clear lead in Ohio.
When you look at the polls crosstabs, you find nothing but dominating news for Romney:
- Romney leads with Independents 49-42, continuing a dominant trend found in most swing states and nationally
- Romney Holds 90% of the Republican vote while stealing 14% of Democrats, taking almost twice the party-switch vote that Obama does
- Obama only holds 86% of Democrats while only taking 8% of Republicans
With these internals, the only way to show an Obama lead is to juice the turnout of Democrats. Also helping Obama in the poll is the current tally of 21% early voters, which still seems quite high. I know we are getting into “real” data again, but Ohio counties have reported an average of under 13% turnout thus far, roughly 40% below the figure that PPP is finding. In reality, if these internals hold Ohio is not even close. But then again, how much faith can we have in any poll at this point?
Follow Dustin Hawkins on Twitter
COMMENTS
Please let us know if you're having issues with commenting.