There were a lot of theories floated to explain STV’s crushing defeat two weeks ago in BC. Given there is no shortage of PhD students in the ranks of PR advocates, I’m sure we’ll be treated to several papers on this topic in the coming months.
In the meantime, I plugged the results from ElectionsBC into a spreadsheet and came up with the following findings.
1. The margin of victory in a riding had little impact on support for STV. One of the main reasons I decided to do this little exercise was to test the hypothesis that voters in ridings that are routinely blow outs would be more eager for electoral reform, to make their vote “count”. But that was not the case, with a non-existant -0.003 correlation between the general election margin of victory and support for STV in each riding.
2. Regionally, the big cities were more open to STV. This isn’t too surprising, as one of the biggest concerns about STV was always the massive ridings that would be created in some rural areas. Here are the region-by-region support levels for STV:
Greater Victoria 50.2%
Vancouver 46.5%
Vancouver Island 42.0%
North Shore and Sunshine Coast 40.9%
Kootenays 39.9%
Okanagan, Shuswap, Boundary 37.8%
Vancouver Eastern Suburbs 37.4%
Northern BC 35.1%
Richmond and Delta 33.3%
Thompson and Cariboo 31.3%
Surrey 31.3%
Fraser Valley 30.7%
3. Voting behaviours are closely linked to STV support. There is a strong correlation between support for the Green Party in a riding and that riding’s overall support for STV (r = 0.602). To a lesser extent, ridings voting NDP tend to be more STV-friendly (r = 0.314), while ridings voting Liberal prefer FPTP (r = -0.472).
Now, this isn’t to say that Liberals necessarily voted against STV and Dippers voted for it – just that Liberal ridings tended to be against the change (think of the situation in the US where states with large african american populations often vote Republican). Still, it wouldn’t at all surprise me if it was NDP and Green supporters pushing STV, with provincial Liberals more cautious.
And, for those of you asking, even after region is controlled for, vote intent is still a significant predictor of STV support. In the model I set up, regional and vote factors had a similar amount of influence.
