. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Nap Time!!!

Wednesday, April 16, 2008
More elections thoughts

The voter file that's on the website does not include votes for candidates who dropped. Instead of doing the simple thing (dropping the candidates in the tabulation program), it appears the Elections Council tried to entirely scrub the voter file of any votes for those dropped candidates (so they show up as having 0 votes). The exact method they used is not clear, though a global find-and-replace seems plausible, because voter ID numbers are missing pieces that are equal to the candidate numbers of those who dropped. Given the way the Elections Council has been running things, though, I'm suspicious of how effective that method was, especially syntactically in relation to commas. I haven't been able to find consecutive commas (which seem likely if the voter number alone was removed), but there are hanging commas on the ends of the Senate preference list. Similarly, though DAAP candidate Vanessa Gathi was absent (dropped? Or not on the ballot?), her number was still present in other candidate numbers which are present. I'll try to check with Kozak and see if I can get the original version.

SQUELCH! did worse than last year, failing to quota their candidate.

9900 people voted for Senator, putting quota at 472.4 and 20th place at 347.8. Turnout for every office but Student Advocate is up from last year, even though overall turnout is down (probably because there were fewer referenda).

I've tried a few drop combinations without really changing the results. To increase their share in the Senate, a party would probably have to drop their highest vote-getters (and thus potential winners), and I'm not sure either party could handle the morale problems that might arise from such a strategy.

As for why Student Action went kerplop, I don't really know. Student Action gets votes by handing out favors from a position of power, and CalSERVE gets votes by calling everyone else a racist (which is easier from a position of less power), so I'm not sure how next year will go. Student Action has recovered from things like this before. I only hope they have the willingness to stand up to CalSERVE stonewalling when it comes to handing out money. If CalSERVE refuses to allow Student Action to reduce allocations to their pet causes, and Student Action doesn't have the balls to respond by refusing to vote for the original version, things might look grim.

posted by Beetle Aurora Drake 4/16/2008 07:21:00 AM #
Comments (16)
. . .
Does there exist data on turnout, grad vs. undergrad? It seemed to me that grad students were flying emails back and forth more than most years, mostly about killing the SUPERB referendum. But besides killing SUPERB, this may have also helped Calserve. It would be interesting to see some data on this.
I don't have that information, and I'm not sure if it was collected. I've seen information like that in the Daily Cal in previous years, though, so maybe somebody knows.
By the way, someone definitely did do find and replace to remove senator numbers 121, 123, 126, 129, 134, 140, 141, 155. Look at the voter IDs numbers in the 22100-22200 range, because they have these digits removed as well. Also look at the voter IDs 2155x to see that they became '2x' after the 155 was removed.

That is a weird thing to do.
"Student Action gets votes by handing out favors from a position of power, and CalSERVE gets votes by calling everyone else a racist"

Hahahah. So true! The Asian girl that kept getting people to sing would reject songs to sing from people because they were too white. I'm surprised she let them sing the theme song for Fresh Prince.
Hello, Jim Fung here, just checking back in on your blog from my new home in the great state of Texas because I'm so excited about these election results and excited about what they might mean in terms of real progress on issues important to students!

Why do these CalSERVE sweeps always happen right *after* I graduate? (I missed the 2003 sweep by 1 year too.) Darn!

I think CalSERVE gets votes because (in no particuar order) 1) it is a progressive party with consistent values and Berkeley students are progressive; 2) it is an experienced party compared to other progressive parties ("... since 1984"); 3) it truly listens to and represents its base constituencies, including but not limited to minority communities; 4) it is not Student Action.

I'm not on campus anymore, but I suspect this year it did even better than last year because it passed the test this year of being in charge of the ASUC without any major scandals (that I'm aware of). Also, for freshmen, a CalSERVE-led ASUC is what they are familiar with. Also, Van Nguyen seems like he was a charismatic, dynamic, innovative leader.

Also, I think the CalSERVE website got a really good, informative revamp this year.

Sarah Gold's endorsement probably also helped in terms of image since Cal Dems is a moderately progressive club. I don't think I remember the last time the Cal Dems president endorsed the entire CalSERVE slate.

It's funny to see people in the comments for other posts whining about CalSERVE throwing up anybody and winning, because whether that's true or not (I don't think that's true just from looking at bios and reading what the of-course-ever-unreliable Daily Cal wrote -- the winning candidates seem qualified to me), that's what the Student Action machine would do year after year and win. If CalSERVE has reversed that it is not just some "scary" accident, it is because they have earned a shift in students' votes.
Hey -

count the number of lines in the vote file - 10290. Compare that to the number of votes ASUCTabulator registers - 9900. The search-and-replace method of knocking out candidates actually corrupted the format of the ballots to the point that asuctabulator can't read them. This doesn't affect the senate results (just the ordering of some senators), but it might effect the exec results (particularly for curtis, who was very close).

Also, check out the last line in the vote file - it's been hand entered (the votes for senate/execs should be separated by tabs, this one is separated by spaces). It looks the same if you open it in a text editor, but you can tell by setting tab size to something huge. Also, the ballot id is 99999, whereas every other id is equal to the id on the preceding line plus one (the preceding ballot id is 31985). Removing the ballot doesn't affect the outcome of any election, but it does make me wonder how secure the votes were after the election was over.
The concerns raised by the last poster warrant a serious investigation
10,290 is the number of total votes, right? some voters only vote for senate and some only for President or EVP or whatever, so just because the there are only 9,900 Senate votes, that doesn't mean some votes weren't counted, just that some people only voted for exec and some only voted for Senate.
Strong evidence that it was grad students who killed the SUPERB referendum--

The primary "grad student" candidate for senate was Brad Froehle (#103). He also finished #21 in the senate race. Of the voters who ranked him #1 on their ballot, the same voters voted on the referendum 233N,36Y,13A.

That's 197 votes right there, more than the margin by which SUPERB lost (112).

And that's only 282 votes. Many more grad students voted I'm sure.
Correction: the number of lines in the vote file is 10459, which is what you get if you subtract the last voter ID number from the first.
If you get the original vote file could you post it? I'm trying to get it as well, and the ECC hasn't respond to my request.
Correction the correction - The number of lines in the file is definitely 10290 (open it in word or notepad, or if you're on unix call 'wc -l vote2008.txt'), but it is very interesting to see that the difference in ballot numbers is different from the number of lines.

There could be a good reason for this - maybe if people change their vote the original ballot gets removed? I'm not really sure, but it would definitely be good to ask about.
Okay, the comments above are very disturbing. However, from what I can tell in the Tabulator code (in the "asuc/" folder of the Zip file), the main issue regarding IDs is inconsequential.

Here is the code that processes each line in the vote file:

int skipNumber = 0;
skipNumber = 1;
}else if(model.isRaceType("Academic VP")){
skipNumber = 2;
}else if(model.isRaceType("External VP")){
skipNumber = 3;
}else if(model.isRaceType("Executive VP")){
skipNumber = 4;
}else if(model.isRaceType("Student Advocate")){
skipNumber = 5;
}else if(model.isRaceType("President")){
skipNumber = 6;
for(int j=0; j < skipNumber; j++){
tokenType = ballotTokenStream.nextToken();

As you can see, the first field in each line (the vote "id") is ALWAYS skipped. As far as I can tell, it is not used to organize, process, calculate, tabulate, or otherwise manipulate/understand the data in the vote file, so it would not have tampered with the counts to have messed up id's, even ones identical to one another. These are most likely ID's that are used for reference in the voting database, and not tabulation.

As for the mystery final line: this will have to be verified and determined by the Elections Council. As far as I can tell, the line should have been counted by the ballot tabulator since the StringTokenizer class in CandidateTable.java which processes each line, does not actually specify the tab as the only delimiter--so by default, the class will tokenize (parses each field out) by spaces as well as tabs and new-lines. There might be a story behind this mystery vote that we are unaware of, and which may require investigation by ECC.
No, some people vote only for the referendum, which is kept in a different file. There were 169 of those. Note that 10459 - 10290 = 169.
Additionally, the final line does appear to be entered manually, because as someone noted above it is organized by spaces rather than tabs, but also because it is missing a final field (which presumably is for the referendum since it is blank on every other vote as well). That line only has 6, while every other line has 7 fields between quotations. This would not have changed the count for any of the elections (the missing field, I mean), but it does indicate manual entry.
who am i?-
"it is an experienced party compared to other progressive parties ("... since 1984")"

Experience beyond 4 years really doesn't matter because that is the extent of institutional memory.
Post a Comment

. . .