Kamala Harris Projected Win: 9,105 votes (11/11/10 Analysis)

[David__77]

I just finished reviewing the unprocessed ballot report available at the Secretary of State's website. Presuming that these numbers are accurate, and the remaining ballots at the county level split between Kamala Harris and Cooley the way the county-level returns have so far, when all ballot are processed, Harris would lead by 9,105 votes. This figure is premised in the presumption that all outstanding ballots are valid and will be counted.

I wanted to conduct some sensitivity analysis, particularly since the unprocessed ballots are categorized as outstanding votes by mail, provisional ballots, and "other" ballots, which includes "(1) ballots that are damaged or ballots that could not be machine read and need to be remade, and (2) ballots diverted by optical scanners for further review." If the proportion between the three categories varied across counties significantly, it could affect the projection.

I adjusted the projection by applying percentages to the various categories of ballots: for instance, perhaps only 50% of provisional ballots in each county will be found to be valid (I did not take into account potential inter-county differences in validity rates, as I have no empirical basis for doing so). In general, the "other" ballots negatively impact Harris' margin. That is, the lower the validity rate for these ballots, the better for her. Conversely, the higher the validity rate for provisional ballots, the better for Harris:

If 100% of VBMs, 25% of provisionals, and 75% of "other" ballots are valid: Harris is projected to win by 8,623 votes.
If 100% of VBMs, 75% of provisionals, and 25% of "other" ballots are valid: Harris is projected to win by 10,583 votes.
If, on the other (unlikely) hand, 75% of VBMs, 25% of provisionals, and 75% of "other" ballots are valid: Harris is projected to win by 1,258 votes.

I find the latter scenario extremely pessimistic, and still Harris would be projected to win.

The greatest uncertainty is associated with varying rates of ballot validity across counties, as well as potential errors in the numbers contained in the unprocessed ballot report. There is no empirical basis for assuming that the errors bias projections on way or the other; for instance, Orange County OR Los Angeles County could in reality have fewer or more outstanding ballots.

Overall, I see cause for optimism. Indeed, the result of Democrats coming into a lead, or increasing their lead in the late vote-counting process is "normal" in California, as the larger countries take longer to complete the ballot counting process.

[CreekDog]

1. Also the early votes tend to be more conservative, with a more liberal vote on election day

and it's those election day votes that tend to be counted later.

[Merlot]

2. I hope you're right

Kamala Harris is my newest hero. What I've read about her and hearing her speak makes me think she will be very good as Atty General.

She should have been campaigning in SoCal a lot sooner and that margin would probably be larger.

Thanks for your analysis.

[MrsCorleone]

3. Don't tease us, now.

Fingers are crossed for Kamala win.

[Iggo]

4. Yikes! That's tight.

[proud patriot]

5. fingers crossed

[David__77]

6. 11/12 Update: Good News

County, New Vote Share for Harris, Pre-Existing Vote Share for Harris, +/- % Change Harris
Kings 42% 33% 10%
Orange 43% 34% 9%
Riverside 50% 40% 10%
Sacramento 55% 49% 7%
Ventura 43% 42% 1%
Yolo 58% 58% 0%

There are four columns of data above relating to the 67,213 Harris and Cooley votes added to the vote totals today (for this analysis, I am excluding third party vote). The first row of data shows that Harris received 42% of the Kings County (Harris and Cooley, no third party) votes added today, while she had only 33% of the total Kings County vote before today, for an increased share of 10%. The table replicates these variables for the other counties reporting numbers today. Notably, Sacramento County represents about 70% of the votes added today, and Harris won 55% of the two-person votes added today for that county (my home), while she was narrowly trailing in the Sacramento votes reported before today. Of the six counties, only in Yolo County did the margin move against Harris, though by a tiny amount (0.4%) relative to the other swings.

In any event, the figures look good, and, if anything, appear to reinforce the estimate made above.

Have a good weekend!

[Merlot]

7. Thanks for the update!

Any idea when this will be finalized?

[David__77]

9. The deadline is 12/3, but we'll know before then. BTW Harris leads now.

http://vote.sos.ca.gov/returns/close-contests/

I may run the numbers again later since I have a cold and am couch-bound lol...

[Iggo]

8. Thanks for crunching the numbers, man.

I really appreciate it.

[bemildred]

10. Harris takes slim lead in attorney general's race [Updated] (11/13/10)

San Francisco Dist. Atty. Kamala Harris took the slenderest of leads in the race for California attorney general late Friday, moving ahead by 303 votes out of more than 8 million cast in one of the closest statewide races in decades.



Harris, the Democratic candidate, was buoyed by updated vote counts from several counties where she had outpolled Los Angeles County Dist. Atty. Steve Cooley, the Republican, on election day, including Los Angeles, Contra Costa, Marin and Santa Clara.

As of 6:30 p.m. Friday, Harris had 4,117,728 votes compared to 4,117,425 for Cooley, according to a Times review of website updates by all 58 counties.

http://latimesblogs.latimes.com/lanow/2010/11/harris-cooley-attorney-general-race.html

[David__77]

11. Kamala Harris Projected Win: 32,904 votes (11/17/10 Analysis)

I repeated an analysis of unprocessed ballots by county with a methodology identical to that I employed on 11/11: Presuming that these numbers are accurate, and the remaining ballots at the county level split between Kamala Harris and Cooley the way the county-level returns have so far, when all ballot are processed, Harris would lead by 32,904 votes. This figure is premised in the presumption that all outstanding ballots are valid and will be counted.

The reason that Harris' has developed a larger-than-expected lead is because the ballots being late have been favorable to her as compared with those that had been counted as of 11/11. In both Democratic and Republican counties, the votes counted late have trended her direction. It seems like this is pretty well sealed at this point, which is probably why the Republicans have been somewhat quiet about it.

To put it another way: in order for Cooley to win by one vote, he would have to systematically OVERPERFORM to-date results by 2.45%. Precisely the opposite result is intuitive, however. Late-counted ballots generally trend Democratic in California, even when controlling for county. The original 11/11 analysis, performed when Harris was behind in the count, was a conservative one.

Congratulations to the people of California in rejecting every state-level Republican candidate, in yielding not one seat to Republicans in the legislature or in congress, in defeating the Proposition 8 lawyer GOP candidate in picking up a Republican-held assembly seat, and in reelecting Barbara Boxer for senate!

[lazarus]

12. I'll be back doing monitoring again tomorrow

yesterday, we outnumbered Cooley's campaign something like 4-1.