A few weeks ago, I tried to do some analysis of how the outcome of the 2008 election might have changed under alternative vote. Today I ran the same process using data from the 2011 election (I used preliminary results where the verified results weren't yet ready) and voter second choice data (see slide 5 in the gallery section) from a slight more recent survey than the one I used in the 2008 analysis. The Procedure and Politics blog did a similar analysis recently, I'm rerunning the most recently updated numbers using the script developed earlier.
Here were the results in this updated simulation. See the earlier 2008 analysis for a description of the methodology.
|Party||Original||In AV Sim||% Seats won||% Seats AV||Popular Vote|
In this scenario, the Conservatives do not attain a majority, they get 22 fewer seats, (I'm reading slightly different reports on final numbers of seats so there may be a different of a seat or two based on validated final results). The Liberals pick up 14 and the NDP picks up 11. The BQ drops down to a single seat. Also note that while the percentage of seats won more closely mirrors popular vote for the three largest parties in the AV case, that's not true for the BQ, their seat percentage more closely ties with popular vote in FPTP. This is a good reminder that AV isn't necessarily of much utility to some smaller fringe parties with heavily concentrated support in relatively diverse ridings.
The ridings that changed in the calculations I ran:
|District Number||Name||Winning Party||Recalculated|
|12009||South Shore--St. Margaret's||Consv||NDP|
|35016||Don Valley East||Consv||Lib|
|35017||Don Valley West||Consv||Lib|
|35043||London North Centre||Consv||Lib|
|35079||Sault Ste. Marie||Consv||NDP|
|46014||Winnipeg South Centre||Consv||Lib|
|59031||Vancouver Island North||Consv||NDP|
Data is available here: link (see the "WithAV" tab for the numbers w/ the changes)