A major voting disaster Tuesday shows the pitfalls of having each of the state's 58 counties set its own rules and ballot designs. Voters in Los Angeles County who belong to no party ("decline-to-state" voters) and who wanted to vote in the Democratic presidential primary on Tuesday got a raw deal.
Where most counties simply give nonpartisan voters a party ballot at their request, Los Angeles County gives nonpartisan voters a separate ballot that requires voters to fill out a bubble for the presidential candidate of their choice – and a second bubble for a political party.
Many voters do not see and do not fill out the second bubble – and, thus, their votes do not count.
The scale of disenfranchisement is huge – 94,500 of 189,000 decline-to-state votes. That's half of the nonpartisan ballots. By comparison, in the infamous Florida "butterfly ballot" debacle in the 2000 presidential election, 19,120 Palm Beach County ballots went uncounted because of the bad ballot design.
The same number is reported by the SFChronicle, and it lines up nicely with my own guess yesterday, based on the fact that only 10% of the votes in the Dem presidential race came from DTS voters, when registration figures suggest that they should've been at least 20% of the votes. (At my own precinct, they were closer to 30%.)
There were also statewide issues with poorly-educated pollworkers either forcing DTS voters to vote provisionally, or turning them away entirely. And the system-wide problems with absentee DTS voters getting the NP ballot (with some of them submitting that ballot, not realizing that the proper procedure would be to surrender it at a polling station in order to vote there; if they submitted the NP ballot and then tried to separately vote at a polling station, unable to surrender their NP ballot, they'd have to vote provisionally, and the provisional would almost certainly be rejected).
My guess is that all of this tallied up would not overcome the difference between Obama and Hillary on the popular vote across CA. However, it would be enough to overcome the difference between their popular vote nationwide on Super Tuesday (which appears to have been under 100k -- the exact number depends on what source you look at and how many votes had been tallied when they reported; Time says it was around 50k, and I think I remember hearing 85k on the radio) and would be enough to shift the delegate count in CA towards Obama by a substantial number -- enough to mean that Obama won significantly more delegates on Super Tuesday than Hillary did. (He may've already just barely won more, but the exact count is still in flux as assorted caucus math is worked out.)
Courage Campaign is running a petition to the LA registrar to count NP votes in the Dem race, even if they failed to fill in the double bubble. After Florida, and after Donna Frye losing in San Diego because some GOP judge felt that writeins for "Donna Fry" weren't clearly enough for her, I'm sick of seeing elections tainted by systems that refuse to count votes where the voter's intent is dead obvious. We need to not only count all the votes in this race, we need a fracking Federal Constitutional Amendment guaranteeing the right of the voter to have their vote counted as long as the intention is clear, even if it doesn't conform to the exact procedure implemented by lazy, incompetent, or malicious registrars (I'm looking at you, L.A., and at Theresa LePore of Palm Beach) and Secretaries of State (Ken Blackwell and Katherine Harris).
ETA: OK, so I realized I should think through a little more math. I'm having trouble finding detailed exit polling covering the breakdown of DTS vs Dem Party votes in the CA primary, but the numbers I could find (pre-election polling and info on races in other states) says Obama was losing to Clinton among Dems by about 5 points, but winning among DTS voters by 9. If you figure there are a few votes for minor candidates, then we might see the 94k votes break out to 50 for Obama and 40 for Clinton, for a net move of +10k to Obama. Which is probably only enough to shift about two delegates, though the details would depend on where the votes were in LA, and how close particular congressional districts were to the point at which a delegate changes hands.
ETA: Actually, according to this, nationwide results had Obama up 20 points among independents. So that would mean moving his total about +20k, out of a pool of a million, which is about two percentage points... Maybe 3-4 delegates transferred. And if it's Obama up four, Clinton down four, that would put them tied for elected delegates -- the last AP count I saw had them separated by eight.