March 06, 2008
Once again with Republicans in the Democratic primary

I see that Royal Masset seems to have found a balm for his anxiety about the huge disparity in turnout between the Democrats and the Republicans. He's noted the 700,000 vote dropoff from the Presidential primary to the Senate primary and concluded:


At least 500,000 Republicans voted in the 2008 Texas Democrat Primary.

The proof of this is simple and conclusive. In the Democrat Primary (Democrat Primary numbers cited here are with 16 precincts not yet counted. Republican numbers are 100% complete.) 2,856,813 votes were cast in the Presidential race. For the Democrat U. S. Senate race, which was heavily contested by 4 candidates, only 2,163,477 votes were cast. 693,336 fewer votes were cast in the US Senate race than in the Presidential race. A decrease of this magnitude is unprecedented. The only explanation that makes sense to me is that at least 500,000 of that 693,336 vote decrease was caused by Republicans voting in the Democrat primary for the sole purpose of influencing the outcome of the Democrat race to nominate a president.

In the Republican Primary 1,380,907 votes were cast for President. For US Senate, which unlike the Democrat Senate race has no credible competition, 1,216,732 votes were cast, mostly for John Cornyn. This is a much smaller drop off than the Democrats had.

[...]

There is normally a 7-10% drop off between the Presidential and US Senate races on a primary ballot. Some voters just don't like any of the candidates. But the overwhelming majority of voters vote in most races. They like to participate and make their ballot count. In the 2000 Texas Republican Primary, 1,126,753 votes were cast for President (mostly for Governor Bush.) Yet an uncontested Senator Hutchison, despite then not being a favorite of the pro-life wing of the party, received 955,033 votes, almost as many as the 986,416 received by Bush.


Well, I'd simply suggest that the Senate primary was as these things go fairly low-profile, as only Rick Noriega had any money to spend on voter outreach and name ID. I don't think it's crazy to suggest there was a higher than usual of voters who really didn't know anything about any race besides the Presidential. It may be true that the level of undervotes for a Senate race is unprecedented as Masset says, but what about this election, from the caliber of the candidates to the massiveness of the stakes to the vast disparity of the resources between the Presidential campaigns and those of the downballot statewides, isn't unprecedented? The latter, I think is the far more compelling reason.

Put simply, the Presidential campaigns took up all the oxygen. The AP reported on Monday how the rest of the candidates were struggling to get attention. I guarantee, you ask any Democratic consultant or activist for one of those candidates, they'll go on a rant about this. Rick Noriega was the highest profile person running statewide, but he only had so much money, and he was running against nobodies and do-nothings, meaning there was little earned media to be had outside the latest flap over 3 AM phone calls and other trivia. I don't see it as the least bit surprising that the dropoff would be unprecedentedly high in this situation.

Another way of expressing this: Had we gotten the Mikal Watts/Rick Noriega matchup were were originally promised, I guarantee you we wouldn't be having this conversation right now.

But let's check some empirical data, since Masset points to a 43% undervote in heavily Republican Denton County. Here's a chart examining the Senate vote as a percentage of the Presidential for all counties with at least 100,000 registered voters (all data courtesy of the SOS):


County Senate Pct
===================
Overall 75.74
HARRIS 78.06
DALLAS 70.52
TARRANT 74.46
BEXAR 81.05
TRAVIS 75.18
COLLIN 57.92
EL PASO 80.91
DENTON 72.67
HIDALGO 74.14
FORT BEND 77.59
MONTGOMERY 71.46
WILLIAMSON 75.73
NUECES 81.47
GALVESTON 73.08
CAMERON 81.87
BRAZORIA 79.44
LUBBOCK 73.93
JEFFERSON 61.87
BELL 75.65
MCLENNAN 72.63
SMITH 66.30
WEBB 89.06

In other words, Collin is an outlier, and however Republican it may be, there ain't enough of them there to make up a half million mischief-makers. The full spreadsheet is here (XLS) for your perusal. You'll note that Collin scores the lowest of all counties other than King, with its 205 registered voters; only Starr at 59.52% joins them in dipping below 60%. I submit to you that there's nothing remarkable about any of this.

If Texas' Democratic turnout blockbuster were unusual on its own, that might be suggestive of shenanigans. But this primary has led to record-setting Democratic performances all over the country, including such strongholds as Alabama and Utah. And if you include caucus states, Idaho, too. Again, given the extreme amount of attention given to the Texas primary, why is this so hard to comprehend?

Finally, I'll note again that the judicial races in Harris County had only slightly higher dropoffs from the Presidential race compared to 2004, when pretty much everyone participating was hardcore. That still gave them all vote totals that swamped their Republican counterparts, and it suggests to me that there were both a lot of new voters, some of whom only knew or cared about the one race, and a lot of others who really were invested in the whole ticket. Which is what we seem to be seeing elsewhere in the state as well.

I see Dr. Murray is discussing this as well.


[Sen. Hillary] Clinton ended up with a victory of just less than 100,000 votes in the Lone Star State. We heard tales Wednesday of many Republicans being asked to vote for Clinton in the primary. Those voters not only secured a win for Clinton, but may have done themselves a disservice in the runoff.

"Here we had about 400,000 Democratic primary voters, I think about 60,000-80,000 were normally Republican voters, possibly as many as 100,000," said KTRK Political Consultant and blogger, Dr. Richard Murray.

Dr. Richard Murray says surveys taken in the days before the primary indicated Senator Barack Obama would be the one to win votes from self-described independents and Republicans 58% to 39%.

"Something turned the vote from Obama in the early voting to Clinton in the late voting and I think Rush Limbaugh had something to do with that switch," Dr. Murray said.


Do we have exit polling data available? I haven't seen any yet. I can say that Obama still won the Election Day vote in Harris County, albeit by a lesser percentage than he won the early vote. How much of that is Republicans doing Rush's bidding, and how much of that is Team Clinton stepping up its ground game after seeing the way early voting was going? I think the latter notion deserves a little more respect here. It doesn't sound to me like Dr. Murray is claiming all Republicans who voted in the Dem primary did so for the sole purpose of boosting Hillary Clinton. The presence of crossovers has never been in dispute, and again we can tell how many are true-red Republicans from their voting history (has anyone compiled that data yet?), but we still don't know what all of their motives were.

Look, I'm not saying there weren't any Republican infiltrators. I know, logically and anecdotally, that there were. I'm just saying that I can think of other explanations for the downballot dropoff level besides the one Masset has come up with, and I think the one I've presented here is far more likely.

UPDATE: This Chron story, which says rural Texas won the day for Hillary Clinton, includes the following sidebar:


CROSSOVER VOTE?

Some Republicans had been urged by radio talk-show host Rush Limbaugh to vote in the Democratic primary for Hillary Rodham Clinton to keep the race alive. But there was not evidence of that occurring on a large scale. Network exit polls showed about 9 percent of the Democratic primary turnout was self-identified Republicans, who cast 52 percent of their ballots for Obama.


It also says that a majority cast their ballots for Obama in the Republican suburbs. Let's check that:

Brazoria = 14,632, Clinton, 14,192 Obama
Collin = 31,994 Clinton, 40,000 Obama
Denton = 24,045 Clinton, 30,374 Obama
Fort Bend = 25,670 Clinton, 43,893 Obama
Montgomery = 17,277 Clinton, 13,429 Obama
Williamson = 21,289 Clinton, 27,906 Obama

Brazoria (which is as rural as it is suburban) and Montgomery both went to Clinton, each with a spike on Election Day. The others, not so much, and E-Day totals are not very different from early voting. Make of that what you will.

UPDATE: Dave Weigel and Todd Beeton weigh in. And remember when stories before last week were about how some Republicans were voting in the Dem primary so they'd be sure to get a chance to vote against Hillary Clinton, since they figured she wouldn't be on the ballot in November? Boy, those were the days.

Posted by Charles Kuffner on March 06, 2008 to Election 2008
Comments

As I noted at my post on the last day of early voting, there were people who showed up just to vote for Obama and leave. After waiting in line for two hours, I could understand their fatigue.

The knowledge level for down ballot races might understandably be low for groups caught up in Obamamania. I'd look closely at precincts where Obama dominated to test my theory as a contributing factor.

Posted by: racymind on March 6, 2008 7:31 AM

I don't know if we should listen to someone so dumb that they don't understand basic grammar. "Democrat" is a noun. "Democratic" is the proper adjective to describe nouns, like "party" and "primary," that relate to Democrats. Someone send Royal Masset back to school.

Posted by: KB on March 6, 2008 10:01 AM

Wouldn't the surprising thing have been if there weren't a massive dropoff? I'm looking at Dallas county data, because it's the area I know the most about, and because Dallas voters really, deeply care about the Democratic primary for sheriff for some reason.

In 2008, 297,336 votes for president were cast in the Dallas County Democratic primary, 209,677 votes were cast for senator, and 228,233 were cast for sheriff. So the dropoff from president to sheriff was 23%. In 2004 (when Edwards had already dropped out before the primary), 51,749 (meaningless) votes were cast for president, there was no Senate race, and 44,244 votes were cast for sheriff. That's a 17% dropoff. In 2000, there were 46,105 votes for president, 35,104 votes for senate, and 34,643 votes for sheriff, a 25% dropoff. So even for a race that's (in Dallas) as important as the Senate race, there's a substantial dropoff from a meaningless presidential primary vote. Of course there's going to be a substantial dropoff when you send huges numbers of people who have never voted in a primary in their life to their first primary - I'd say 500,000 is a better estimate of the number of people who didn't realize there were votes for anything except president.

Posted by: Chris Koeberle on March 6, 2008 11:09 AM

It just shows they are scared to death of Hillary AND Obama!

The idiots, if they really did vote in the Democratic primary, they probably cancelled out their vote.

They realize they have a real shitty, old, dull, and dangerous candidate, John McLain and they have nothing else to do but to try to get the Democrat to lose instead of their candidate to win.

Posted by: John Cobarruvias on March 6, 2008 11:13 AM

In previous primaries he's won a much higher percentage of self-identified Republicans ... closer to 60 or 70%. So the most you could say is that 15% of the 10% or so of voters who were Republicans were following Rush's marches orders. And maybe some of the independents as well, though the evidence there is sketchier.

At most we're talking about 3% of votes, and I think that's high-balling it. It's probably just under 1%, which would have taken the race to something like 50-48 and maybe moved a few delegates.

Posted by: Nicholas Beaudrot on March 6, 2008 11:53 AM

The problem I keep having with some of these analyses is that they seem to treat "Democrats" and "Republicans" as immutable groups when all they really are is "people who voted in the Democratic (or Republican) primary four years ago." I know several people who have voted Republican in the last few elections, but who wouldn't consider themselves "Republicans," who have now voted in the Democratic primary. Technically, they're now "Democrats," but they don't consider themselves Democrats either. So you may get a large cross-over vote without there being any nefarious purpose or deliberate sabotage. Like you, Charles, I'm sure that has occurred in some instances, but I'm not convinced it explains the vast majority of ticket-switchers.

Posted by: Kenneth Fair on March 6, 2008 1:21 PM

While the republikanz have their noses pressed up to the candy store window they might want to gaze over at the general election voters who don't vote in the primaries.

Posted by: Charles Hixon on March 6, 2008 6:04 PM

** I don't know if we should listen to someone so dumb that they don't understand basic grammar. "Democrat" is a noun. "Democratic" is the proper adjective to describe nouns, like "party" and "primary," that relate to Democrats. **

Then again, a real stickler for accuracy in language might consider putting quotes around "Democratic," since a primary system that gives so much power to superdelegates and that repudiates the principle of one-man/one-vote by giving some "special" areas disproportionate weight for delegate purposes doesn't seem wholly democratic. :D

Posted by: Kevin Whited on March 8, 2008 8:37 AM