A Lefty Website for Strategic Donation

Swing33​.ca links to the donation pages of 33 can­di­dates that could be key to the outcome of the upcoming election. The idea came from Mitch Anderson, who rents a desk beside the desk I rent, and who has an op-​​ed in the Tyee today explaining. I helped build it.

We’re not fans of the Harper admin­is­tra­tion, and we’d like to do some­thing to boost the prospects of the leftist (and centrist) Canadian parties. There’s 308 ridings in Canada, but not all of them are likely to switch sides in 2011. Of those that are in play, not all are being seri­ously con­tested by a Conservative can­di­date. We had a look at the results of the last election and spread­sheeted out the ones where Conservatives came close to winning or close to losing. There are 33 ridings where there was less than a 5% dif­fer­ence between the Conservative and the top non-​​Conservative can­di­date. That’s our list. If you’re going to donate to a polit­ical campaign (and Mitch argues there’s good reason to), then you might want want to target it at cam­paigns that are likely to decide the number of Conservative seats in the next par­lia­ment. That goals feels impor­tant right now.

We did simple-​​ish math to pick out those ridings, and it’s not a perfect system. We’re using recent history as a guide and history doesn’t sit still. We’ve already spotted a couple of dis­tricts where local con­di­tions have changed since 2008 such that the best con­tender probably isn’t the one Excel picked. Saanich — Gulf Islands is the most obvious to me. (Donate to the Greens there, not the Liberals like the site says.) A com­menter at the Tyee has already called out Esquimalt — Juan de Fuca as another flub.

We’re con­sid­ering aggre­gating that kind of local­ized knowl­edge, and/​or bringing in con­tem­po­rary polling data and pro­viding an alter­na­tive curated list. Time is the lim­i­ta­tion. In the meantime we didn’t want to make it any harder to under­stand what the site was about by adding excep­tions to our algo­rithmic approach.

In sum: if you or someone you know can be con­vinced to pony up some cash to help facil­i­tate a good outcome in this year’s election, Swing33 could provide some rea­son­able guidance for giving that donation the most impact. The Conservatives are far better fundraisers than any of their com­pe­ti­tion, so give it some thought. Links to the donation pages of 33 likely can­di­dates are just two clicks away.

6 comments:

Thanks for doing this! Are the Liberal, etc. parties aware of this so they may promote it as well?

Greens’ polling inac­cu­rate on Saanich — Gulf Islands. Liberals have best chance of winning.

Green polling errors explained on their site here: http://​www​.green​par​tys​trategy​.com/​a​r​t​i​c​l​e​s​/​s​a​a​n​i​c​h​-​g​u​l​f​-​i​s​l​a​n​d​s​-​p​oll

hi Sharon, we haven’t explic­itly con­tacted any parties or can­di­dates. We’re aware of at least a couple that have heard about our site. Thanks for you thanks, it’s appreciated.

Hi SK, the article you point to is inter­esting if it’s accurate. It’s not clear to me exactly which poll it’s com­menting on, but appar­ently it’s not the poll that was con­ducted in fall of 2010 and showed May as the strongest non-​​Conservative can­di­date. We’ll look into it further.

Hi SK — I know this poll — it is from summer 2010. If you read the full report the OraclePoll results actually do show Elizabeth May in the lead. This is con­sis­tent with our SGI polling data over the past 6 months. All studies I know of show that SGI is clearly a two-​​way race between Elizabeth May and Gary Lunn.

DETAIL: The specific OraclePoll result you refer to as showing a Liberal lead is not really a valid test of Green support. That par­tic­ular party support question was run before most res­i­dents knew the names of can­di­dates in the riding, and did NOT mention that Elizabeth May or any other can­di­date were running; it simply asks about support for national polit­ical party.

This was a rea­son­able question for OraclePoll to ask back then because it allowed them to assess the baseline of support for parties prior to intro­ducing the can­di­date name.

And this is the key point: when the OraclePoll ques­tion­naire does ask about party pref­er­ence in con­junc­tion with actual can­di­date names, Elizabeth May is imme­di­ately thrust into the lead. OraclePoll reports a dis­tri­b­u­tion of 29% Elizabeth May, 27% Gary Lunn, with the Liberals, NDP and Other running behind.

Moreover, the OraclePoll uses a rel­a­tively simple approach to
weighting (under­stand­able due to a likely lack of budget). Oracle appears to have weighted for just 2 cells (Gulf Islands vs. Rest of riding). In com­par­ison, the McAllister poll cal­cu­lated precise weights for a full 36 cells within six SGI CSDs, accounting for gender, age and 4 divi­sions of geog­raphy in Saanich penisula and 2 divi­sions on the Gulf Islands. Oracle actually notes that they did NOT apply even their coarse weighting scheme in their report.

Bottom line: when you do look at the relevant question, OraclePoll shows Elizabeth in the lead or sta­tis­ti­cally tied. This is con­sis­tent with the polls we have run in the riding in the past 6 months. SGI is a two-​​way race between Elizabeth May and Gary Lunn.

SK: the article you’re refer­ring to is from last summer — August, 2010. I think things have changed since then.

leave a comment

Notify me of followup comments via e-mail.