Don't recommend this post before you read the explanation. If you still want to, because it's an interesting example of failure in data presentation and tells us something about what happens when someone is not on the ballot, fine, but the initial implication I took from the Ohio SOS's results page is inaccurate. If you want to unrecommend, I'll certainly understand, but I hope that this update makes that unnecessary.
I'm leaving this post in its original form below this line. It appeared to me last night, based on reading the Ohio Secretary of State's results page, that Santorum would have won the election had his name been on the ballot in all counties. It now looks like the district-by-district breakdown of votes on the results page referred only to a separate vote regarding delegates for those districts, which is where Santorum's name did not appear.
The problem from my perspective (and that of the people who clicked the link, saw the same thing I did, and recommended the post thinking that I was correct) is that these contests are presented as if they were themselves popular vote totals.
A screen grab from the web page appears down below as an update to the diary.
There is nothing in the presentation stating "Delegate contest only, not popular vote" that would have been a clue to me and others. The counterargument may be that "well, everyone in Ohio knew that." (Probably not true, but that's still the argument I see from others.) To the extent that's true, the lesson here for Secretaries of State is that people from outside your state will read your results page, so don't presume that they'll understand what you're trying to say even though what you say is, on its face, something different.)
At a minimum, these results give us poli sci geeks some great information: this is a real life test of where votes go when they can't go to Santorum. We've been debating "who gets whose votes" here and elsewhere for a while; the data on the SOS page, once properly relabeled (by us even if not by him), give us good evidence of where they really do go, at least in Ohio.
Here's the post as it appeared for about six hours, during most of which time I was asleep on the west coast.
Here's the page of voting results by district in the State of Ohio. Look at it closely. Notice anything unusual?
All right, let me give you a hint: this diary is brought to you by the numbers 6, 9, and 13.
Still not clear? Look at the names of the Republican candidates for President in Congressional Districts 6, 9, and 13. Rick Santorum's name is not among them. Back when he was not yet considered a serious candidate, as the last of the "not-Romneys" to get a turn at bat, he failed to qualify for the ballot in those three districts. He couldn't get a single vote from there yesterday.
Ohio's other 13 districts were what we might call "competitive Ohio," where Romney and Santorum were in direct competition on the ballot. In Districts 6, 9, 13, Santorum voters had to either not vote or vote for someone else -- such as Gingrich.
One interesting question in Ohio is "who got the most votes?" Another is "who got the most delegates?" But a third question, more interesting if you're trying to predict what will happen in other states, is this: who would have gotten the most votes if Santorum, way back when, had been able to qualify for the ballot in all 16 districts?
This post addresses that question.
We might, of course, expect that the fact that Santorum didn't qualify for the ballot in Districts 6, 9, and 13 means that they were likely especially weak districts for him. Maybe, maybe not -- we have at least one way to check, as I'll describe later. But to have won the state of Ohio, Santorum would have only had to pick up about 10,000 votes out of about 1.2 million GOP primary votes cast. Could he have done it?
The answer is pretty clearly yes. Had Santorum been on the ballot in OH-6 (Eastern border with WV + some PA border), OH-9 (north shore, Toledo to Cleveland, the Kaptur-Kucinich district), and OH-13 (Youngstown-Akron area) -- 6 and 13 being especially good areas for Santorum -- then by my calculations he would have picked up about 25,000 to 50,000 votes, enough to beat Romney by 15,000 to 35,000. At the high end of the estimate, that's about a 3% win.
You can tell by looking at the percentage of votes that went to Gingrich and Perry, in particular, in those districts -- although note the caveat below. Gingrich's percentage in the 13 districts with Santorum on the ballot was about 15.4%. Without Santorum, his percentage of the vote was 30.3%. Perry got 0.8% with Santorum on the ballot, but 4.6% without him. Paul's percentages increased from 9.9% with Santorum on the ballot to 18.6% without him; even Huntsman picked up 1.5%. Romney picked up 7% as well -- perhaps from anti-Gingrich/Paul voters.
The caveat: or so it seems. I thought for quite a while tonight that I'd be able to have a definitive answer about these numbers, but the Secretary of State's breakdown of votes by district doesn't add up to the overall numbers. In some cases (notably Santorum, bit also Romney) the number you get up from adding the totals in the Congressional districts is lower than the statewide total; in other cases (like Huntsman) it's higher, which doesn't make a lot of sense. So I'm not going to post the pretty chart that I made yet; I'll revise it in the morning.
One thing is clear, though, barring some major misreporting on the statewide totals -- had he been on the ballot in those three districts (or frankly, probably even in the Appalachian district of OH-6), Santorum would have beaten Romney last night. Mitt's victory was an accident of fate; you can't read momentum into it -- except in the other states where Santorum isn't on the ballot.
I'll revise the diary (and republish it with my pretty spreadsheet) once the Ohio SoS's vote totals make sense. I'm publishing it now to try to shove a stick into the gears of Romney's spin machine, which depends on people not knowing that Santorum wasn't on the ballot in almost 20% of the state. Had he been, he'd clearly have won.
And that would have meant a very different storyline for today -- one we can still have if we explain it well.
8:45 AM PT: Warning: cryonaut and others below say that the premise of this post is wrong -- that Santorum was on the ballot and could and did get votes from these districts, just not delegates. I reply here. For all I know they're correct, but this post was written (1) with a general knowledge that Santorum could not get delegates from three districts and (2) in light of the following presentation on the Ohio SOS's election results website (I've copied only those for OH-06 to illustrate):
This post was written in good faith and I'm still trying to figure out who's right, because it seems simply unbelievable that a Secretary of State would present results broken down by district, without labeling them as "delegate votes" or anything, in this way.
If it turns out that I relied on the SOS results incorrectly and that rains on Mitt's parade this morning without factual basis, I do apologize, as follows: "this was not the means I would have used."
