Suitably Flip

Was Florida a Closed Primary Or Not?

Color me flummoxed.

In my attempts to self-soothe in the wake of Romney's Florida defeat, I'm poring over the details of the exit polling in search of encouraging morsels, thus far with sporadic success.  Mitt beat McCain soundly among conservatives (37-29), among "issues" voters (35-27), and even edged him out among voters for whom terrorism was the most important issue (29-26).

One breakout that puzzles me though is the vote share by party identification.  Romney and McCain were tied among Republicans at 33-33, while McCain won independents 44-23.  (See page 4 of the exit poll.)

If I'm reading the poll correctly, it suggests 17% of Republican primary voters identified themselves as having no major party affiliation, while 3% identified themselves as registered Democrats.

One of the features of the Florida primary that was supposed to distinguish it from the other contests to date was that it was a closed primary, meaning only party-registered voters could vote in their respective primaries.  This was one of those alluring intangibles that hinted at a Romney advantage relative to the earlier open primaries.

And indeed, the exit polling suggests that Romney likely matched McCain among Republicans in Florida (rounded to the nearest percentage point).  From where I'm sitting, it appears Mitt Romney might actually have won the Florida primary.

Despite my support for Romney, I'm loathe to hunt for extra innings in this or any contest.  That's a game usually better left to the Gores, Kuciniches, Paulnuts, and the other dark-minded conspiracy buffs of the world.  And to be clear, I'm not suggesting conspiracy is afoot, only that Florida's record of electoral execution is... well, checkered.  You don't have to conspire to screw something up in order to screw something up.

At the risk of being lumped into the ranks of the aforementioned, and given that the impact of participation by non-Republicans in this primary appears to account for the entirety of McCain's margin of victory, I'm tempted to swallow my loathing and press for some answers as to what might have happened.  Exit polls rely on the accuracy of the verbal responses given by the pollees and therefore aren't definitively dispositive of anything at all.  But 17% of 1,500 respondents self-identifying as independents who voted Republican means something irregular took place, unless hundreds of voters chose to tell pollsters a similar lie.

As we explored earlier (via Michelle Malkin), at least one voter in Broward County reported being forced to fill out a party-specific ballot, despite identifying himself to poll workers as an independent.  (As you might imagine, he voted for McCain.)  The report claimed that two poll workers confirmed that they had been instructed to give party-specific ballots to independent voters, despite this being in contradiction of state law that requires voters to establish party registration at least 29 days before an election in order to cast a party ballot.

I e-mailed that county's supervisor of elections at around 2 pm on primary day to inquire about the reported irregularity, but have yet to receive a response.

Given the possibility that as much as 20% of the 2 million GOP votes cast today came from independents and Democrats (enough to change the outcome), I think we do now need some answers.

For reference, the exit poll also shows non-Democrats voting in the Democratic primary.  17% identified themselves as independents and 4% as Republicans (see page 3 of the exit poll).

No wonder turnout was so high...

---

Update:  More at Hot Air, Michelle Malkin (from whom I swiped the image up top), and Captain's Quarters.  Thanks all for the links and attention to this story.

Update:  Captain Ed offers this possibility:

Having lived most of my life in closed-primary states, I can tell CapQ readers that the exit polling should surprise no one.
...
In my experience in California, independents would often re-register as either Democrats or Republicans in order to participate in primaries, and then re-register again as independents for the general election.
...
In this case, exit polls show "party identification" statistics that put 20% of the voters outside of the Republican Party. That's their stated personal identification, not their actual party registration for last night's primaries.

I see a couple problems with this theory accounting for the whole irregularity.  First, between the two parties' primaries, roughly 700,000 people (20-21% of primary voters) would have to fall into this category of strategic, temporary re-registerers and/or voters who no longer identify with their party of record.  One could argue that the majestic Democratic tide of 2006 left a lot of Republicans feeling this way, so we shouldn't be surprised to see 20% of registered Republicans disavow the party label.  But if that phenomenon is at work, why did 21% of Democrat ballot casters also identify themselves as either independents or Republicans?

The representation of self-identified independents in the exit polling also approximates the representation of independents in Florida's electorate (17% vs. 19%), which is about what we'd expect to see if independents were unexpectedly being given party ballots (at some, perhaps not all, polling places).

Here's one other data nugget to kick around.  According to the Florida Secretary of State, 4.1 million votes were cast on the property tax amendment.  Since 1.9 million Republican votes and 1.7 million Democratic votes were cast, we're left with about a half million votes (actually 472,000) with no Presidential pick.  Those may not all have been independents (the occasional Democrat might've gone only to vote on the tax question, and skipped the irrelevant primary vote), but even if we assume they all were, that suggests a turnout rate among independents of 21%.  Among Republicans, turnout was over 50%.  Among Democrats (who had no meaningful Presidential contest), it was 41%.  You'd expect the turnout among independents to be lower still, but by half?

Well, yes, possibly.

But given the incredible electoral dereliction reported in the Sun-Sentinel article (poll workers saying they were instructed to give party ballots to independents, in defiance of state law), given Florida's history, and given how trivially easy it would be for the Secretary of State and/or one or more county supervisors to grab a random fistful of party-specific ballots and determine whether 0% or roughly 20% were cast by non-party members, this - to me - remains an open issue.

To that end, I've reached out to the SoS and the Supervisor of Elections from each of Florida's 67 counties to see if anyone can shed additional light.

Comments

Was Florida a Closed Primary Or Not?

Color me flummoxed.

In my attempts to self-soothe in the wake of Romney's Florida defeat, I'm poring over the details of the exit polling in search of encouraging morsels, thus far with sporadic success.  Mitt beat McCain soundly among conservatives (37-29), among "issues" voters (35-27), and even edged him out among voters for whom terrorism was the most important issue (29-26).

One breakout that puzzles me though is the vote share by party identification.  Romney and McCain were tied among Republicans at 33-33, while McCain won independents 44-23.  (See page 4 of the exit poll.)

If I'm reading the poll correctly, it suggests 17% of Republican primary voters identified themselves as having no major party affiliation, while 3% identified themselves as registered Democrats.

One of the features of the Florida primary that was supposed to distinguish it from the other contests to date was that it was a closed primary, meaning only party-registered voters could vote in their respective primaries.  This was one of those alluring intangibles that hinted at a Romney advantage relative to the earlier open primaries.

And indeed, the exit polling suggests that Romney likely matched McCain among Republicans in Florida (rounded to the nearest percentage point).  From where I'm sitting, it appears Mitt Romney might actually have won the Florida primary.

Despite my support for Romney, I'm loathe to hunt for extra innings in this or any contest.  That's a game usually better left to the Gores, Kuciniches, Paulnuts, and the other dark-minded conspiracy buffs of the world.  And to be clear, I'm not suggesting conspiracy is afoot, only that Florida's record of electoral execution is... well, checkered.  You don't have to conspire to screw something up in order to screw something up.

At the risk of being lumped into the ranks of the aforementioned, and given that the impact of participation by non-Republicans in this primary appears to account for the entirety of McCain's margin of victory, I'm tempted to swallow my loathing and press for some answers as to what might have happened.  Exit polls rely on the accuracy of the verbal responses given by the pollees and therefore aren't definitively dispositive of anything at all.  But 17% of 1,500 respondents self-identifying as independents who voted Republican means something irregular took place, unless hundreds of voters chose to tell pollsters a similar lie.

As we explored earlier (via Michelle Malkin), at least one voter in Broward County reported being forced to fill out a party-specific ballot, despite identifying himself to poll workers as an independent.  (As you might imagine, he voted for McCain.)  The report claimed that two poll workers confirmed that they had been instructed to give party-specific ballots to independent voters, despite this being in contradiction of state law that requires voters to establish party registration at least 29 days before an election in order to cast a party ballot.

I e-mailed that county's supervisor of elections at around 2 pm on primary day to inquire about the reported irregularity, but have yet to receive a response.

Given the possibility that as much as 20% of the 2 million GOP votes cast today came from independents and Democrats (enough to change the outcome), I think we do now need some answers.

For reference, the exit poll also shows non-Democrats voting in the Democratic primary.  17% identified themselves as independents and 4% as Republicans (see page 3 of the exit poll).

No wonder turnout was so high...

---

Update:  More at Hot Air, Michelle Malkin (from whom I swiped the image up top), and Captain's Quarters.  Thanks all for the links and attention to this story.

Update:  Captain Ed offers this possibility:

Having lived most of my life in closed-primary states, I can tell CapQ readers that the exit polling should surprise no one.
...
In my experience in California, independents would often re-register as either Democrats or Republicans in order to participate in primaries, and then re-register again as independents for the general election.
...
In this case, exit polls show "party identification" statistics that put 20% of the voters outside of the Republican Party. That's their stated personal identification, not their actual party registration for last night's primaries.

I see a couple problems with this theory accounting for the whole irregularity.  First, between the two parties' primaries, roughly 700,000 people (20-21% of primary voters) would have to fall into this category of strategic, temporary re-registerers and/or voters who no longer identify with their party of record.  One could argue that the majestic Democratic tide of 2006 left a lot of Republicans feeling this way, so we shouldn't be surprised to see 20% of registered Republicans disavow the party label.  But if that phenomenon is at work, why did 21% of Democrat ballot casters also identify themselves as either independents or Republicans?

The representation of self-identified independents in the exit polling also approximates the representation of independents in Florida's electorate (17% vs. 19%), which is about what we'd expect to see if independents were unexpectedly being given party ballots (at some, perhaps not all, polling places).

Here's one other data nugget to kick around.  According to the Florida Secretary of State, 4.1 million votes were cast on the property tax amendment.  Since 1.9 million Republican votes and 1.7 million Democratic votes were cast, we're left with about a half million votes (actually 472,000) with no Presidential pick.  Those may not all have been independents (the occasional Democrat might've gone only to vote on the tax question, and skipped the irrelevant primary vote), but even if we assume they all were, that suggests a turnout rate among independents of 21%.  Among Republicans, turnout was over 50%.  Among Democrats (who had no meaningful Presidential contest), it was 41%.  You'd expect the turnout among independents to be lower still, but by half?

Well, yes, possibly.

But given the incredible electoral dereliction reported in the Sun-Sentinel article (poll workers saying they were instructed to give party ballots to independents, in defiance of state law), given Florida's history, and given how trivially easy it would be for the Secretary of State and/or one or more county supervisors to grab a random fistful of party-specific ballots and determine whether 0% or roughly 20% were cast by non-party members, this - to me - remains an open issue.

To that end, I've reached out to the SoS and the Supervisor of Elections from each of Florida's 67 counties to see if anyone can shed additional light.