Winter 2014 ASUCD Election

Notes

Despite SMART and NOW getting nearly the same number of first place votes, SMART came out with more seats then NOW. The result is that SMART will be over-represented and NOW will be under-represented in the Senate. This is partially a technical issue related to proportional representation. With Choice Voting, and proportional representation systems in general, increasing the number of people being elected also increases the accuracy of translating “percent of votes” into “percent of seats”. Having only six seats available in most ASUCD Senate elections can lead to situations where on voting bloc can be over or under represented. For example, if there were 7 seats open in the Winter 2014 election, NOW would have won that final seat and the results would have looked like this:

SMART
39% of first round votes
43% of Senate seats

NOW
39% of first round votes
43% of Senate seats

Independents
22% of first round votes
14% of Senate seats

You’ll notice that SMART and NOW are much more even but the Independents are now under-represented. As you continue this process, the numbers start to even out. If all 12 Senate seats were up for election at one time, chances are the percentages would be far more accurate. (The pros and cons of that change is an entirely different discussion. I’m not advocating either approach.)

Aside from that mathematical technical issue of a proporitional representation system, there are a couple other reasons why SMART won a third seat instead of NOW, despite their very close number of votes. One important point is that SMART gained 2287 first round votes with 5 candidates and NOW gained 2251 first round votes with 6 candidates. SMART had more support per candidate than NOW did: SMART averaged 457.4 votes per candidate. NOW averaged 375.2 votes per candidate. This means if all the first round votes for SMART were divided evenly (and NOW votes were kept as is), every SMART candidate would have placed above 5 of the 6 NOW candidates in the first round. In a Choice Voting election, this can give an important advantage. This is even more pronounced in ASUCD elections and it has to do with transfering of votes.

If a slate has extremely loyal voters, how their candidates place early in the election may not mean much. However, in ASUCD elections voters do not vote solely along slate lines. For instance, when Arya Shirani (NOW) was eliminated in the Winter 2014 election, it caused Robyn Huey (SMART) to be elected because 45 people who voted for Arya ranked a SMART candidate before the three remaining NOW candidates. Because of this, since SMART already had a lead in the first round votes, they continued to maintain top positions even though it was mostly NOW and Independent candidates being eliminated.

This points out a potential flaw of a slate running a large number of candidates. If the goal of the slate is to elect as many people as possible, they have to consider where a candidate’s votes could transfer if they are eliminated. If enough of those votes will end up going to another slate, it’s possible that candidate should not have been run at all. Not running that candidate could potentially move up your other candidates as voters who normally vote for you slate are forced to choose from fewer options. This would consolidate votes on fewer of your candidates, potentially moving one or two of your candidates above another slate’s candidate in those final, crucial rounds.

The other issue to consider with ASUCD Senate elections is the number of exhausted ballots. As the election continues, this number grows and makes it harder and harder to reach the threshold for election. This is why there are usually candidates elected who do not meet the threshold and are elected simply because they are the only candiates remaining. In ASUCD Senate elections, it is generally the case that the last seat or two is won without meeting the threshold. At that point, all that matters is staying one vote ahead of someone else until the election ends. Vote transfers are absolutely critical to being elected at this point. Those final seats can be decided by only a few votes. In the Fall 2010 election, the final seat was decided by 4 votes and there were 291 votes in the exhausted pile. If just 4 of those 291 voters had ranked Cameron Brown, the result of that election would have been different.

I think it’s entirely possible that a slate could be better off running 3 strong candidates than 6 candidates of equal vote garnering strength. The reason for this is that it puts your 3 candidates further ahead at the start of the election and lessens that chance that one of your weaker candidates helps elect someone else. The problem with this is that it’s based on trying to predict the voter support of each potential candidate. A slate may have more luck trying to build a more loyal voter pool by getting people to actually vote for all of their slate candidates before anyone else. For instance, at the end of the Winter 2014 election, 329 votes in the exhausted pile came from people who voted for a NOW candidate in the first round. If those voters had ranked the other NOW candidates, it could have changed the outcome of the election.