Crawl Across the Ocean

Monday, November 21, 2005

(N) P (A) Soup

There was a beautiful blue sky here Saturday morning, the first one in a while and the weekend forecast was for plenty of sunshine. But as I stepped outside to go vote in the municipal election, the sun went missing and the sky had started to turn grey. By the time I got to the polling booth (2 minutes later), the city was smothered in a heavy fog which still hasn't lifted.

Maybe the fog is just a special effect for a movie, or maybe life has imitated art and, after so many comic book movies being filmed in Vancouver, the city has turned into a comic book city, destined to be shrouded in fog until we are saved (in three years time at the next election?) by some hero in tights. Yes yes, maybe fog is just part of living by the ocean and the timing is coincidence, but really, where is the fun in that?

Anyway, the election results can be found here. In short, the conservative Non-Partisan Alliance has generally returned to it's former strength with NPA rep. Sam Sullivan beating Jim Green in the race for Mayor, and the NPA taking 5 of 10 seats on council, 5 of 7 on the Park Board, and 6 of 9 on the school board. Vision Vancouver took 4 of 5 council seats while COPE took the last council seat as well as the minority 2 seats on the park board and 3 seats on the school board.

The results website has an interesting map of the city showing which districts had more support for Jim Green as Mayor and which had more support for Sam Sullivan. People in Vancouver always talk about the split between the (rich, conservative) West side and the (poor, progressive) East side but the map shows more of a North-South split if you ask me. Perhaps the fact that Sam Sullivan himself is from the East side helped his vote there. I'd be curious to see the map overlaid with a demographic map showing the percentage of Asian residents per district or one showing the median income per district or one showing the median housing / condo price per district. I'm guessing you could accurately predict most district's preference based on these variables and the only ones you'd miss would be those with very local issues (the Cambie corridor and the RAV line, Hastings Park and the decision to put slots there).

One other note, the final mayor votes for the top 5 candidates were as follows:

GREEN, Jim VVN 57796
GREEN, James - 4273
WEST, Ben WLP 1907
YEE, Scott - 688

Note the similarity of the name between the second and third place finishers. Now it's possible (in the 'anything is possible' sort of way) that James Green just happened to be a particularly popular independent or that people deliberately voted for James Green instead of Jim as some kind of statement, but the most plausible explanation seems to be that a lot of people just were confused and voted for James Green when they meant Jim.

If we assume that the number of legitimate James Green voters was the same as the number of Scott Yee voters (Yee had the most votes of any of the remaining 15 independent candidates, so this is a fairly generous assumption), then 3,585 voters mistakenly voted for James. If we add those 3,585 votes back to Jim's total, then the final count would be:

GREEN, Jim VVN 61381
WEST, Ben WLP 1907
YEE, Scott - 688
GREEN, James - 688

Pretty close! Although, I should note that this analysis presumes that out of 61,381 people intending to vote for Jim, 3,585 (5.8%) voted for James by accident. I can see a few people being thrown off by the similar names, but more than 1 in 20?

Anyway, I think that next time there is a close two person race and there is third marginal candidate whose name is similar to one of the two-front runners, I am going to just ignore my personal preference and vote for the front runner who is afflicted with a doppleganger, just to make the final result match up better with voter intentions. Next time. Assuming, that is, that the fog has lifted by then and the city hasn't been overrun by evil (yet colourful) villains with silly nicknames.

Note: The Tyee has some commentary of the results here. The most insightful (in my opinion) is from Ian King (scroll a little).


  • I think it is quite possible that ALL of James Green's votes were mistakes... which means Jim Green should have won.

    By Blogger Andrew Spicer, at 1:45 PM  

  • I don't know about all but yeah, it could be higher than what I said. Another way to look at it is to wonder what the count would have been with James Green not running at all (considering the possibility he was put up to it by the NPA).

    Probably a Jim Green victory in that case as well. It kind of leaves a sour taste in my mouth, that is for sure.

    There was an article in the free daily Metro (pdf, bottom page 1) today about a UBC student who looked at statistically and figured it was pretty clear that most James Green voters were just confused, but couldn't really say for sure whether it was enough to tip the election ). He came up with an estimate that 5% of Jim Green voters accidentally voted for James similar to my guesswork.

    It would be interesting to chart the poll by poll results of Jim and James vote totals but I haven't been able to find the poll by poll voting results online.

    By Blogger Declan, at 12:29 AM  

  • Sounds like an argument for having a run-off election instead of FPTP.

    I was assuming that he only ran because he had the same name as another candidate. Has he any sort of history?

    By Blogger Andrew Spicer, at 6:23 AM  

  • The only history I heard about was a history of failed business deals. Still, his campaign has higher profile than most independents, partly because of the name issue, partly because he seemed to have more money to spend than most independents.

    Runoffs seems like overkill, but it would certainly solve the problem.

    By Blogger Declan, at 12:04 PM  

  • Also, the city did post a spreadsheet of the poll by poll results. I took a look at the correlation between % of people voting for Jim and James by poll and it was actually lower than I expected (0.38) but still significant.

    By Blogger Declan, at 12:09 PM  

  • Hmm, well, 0.38 represents a low-to-moderate degree of correlation, one that may not statistically significant. What were the correlations between those voting for either Greens and the other candidates? Additionally, what were the sample sizes of each poll, and how much did they differ?

    By Blogger JG, at 6:16 PM  

  • I controlled for sample size (measured % of total for each candidate rather than number of votes), I only looked in passing at correlation with a couple of other independent candidates but the ones I looked at were lower (near 0). You're right that .38 is by no means overwhelming evidence of anything. If I get more time maybe I'll look at it more systematically.

    The number of James voters per poll was pretty small so the pattern in James voting from poll to poll could be hard to pick up.

    I suppose it's also possible that the chance of misvoting could vary from poll to poll although I can't see why (other than perhaps due to age (careless younger voters? senile older voters?)

    By Blogger Declan, at 9:19 PM  

Post a Comment

<< Home