The Georgia Straight notes casually in its blog that of the 6 ridings in BC where the pro-STV vote passed 50 percent, “these ridings include some of the safest in the province.”
So I took a look at the preliminary results from election night for the referendum and the election, and with delusions of Nate Silver grandeur, I scatterplotted the results.
Victory Margin is winning party's percentage of vote subtracted by losing party's percentage, including only Liberal and NDP.
The results is a noticeable trend. Far more so than seat competitiveness, party mattered. In ridings that went for the NDP, STV support increased on an almost linear line with support for the New Democrats. In Liberal ridings, the results was slightly more chaotic, but the trend is still the same: the more votes the NDP got, and the less the Liberals did, the more people voted for STV in that riding. This is in line with Angus Reid’s exit polling reporting that 51% of NDP voters and 22% of Liberal voters supported STV (Greens supported STV the most, at 66%, but their candidates’ vote shares weren’t included, as no candidate made it past third place). BC overall fit perfectly into the trend, getting just about what would have been predicted by an equation if given the Liberal vote percentage.
That isn’t to say that seat competitiveness wasn’t a factor at all. This is pure speculation, but the reason that Liberal-won seats are less correlated between the STV vote and the party vote may be that in safe Liberal seats, feelings of neglect under First Past the Post may have somewhat offset the Liberal inclination to vote for FPTP.
However, most of the anomalies on the graph appear to be caused by a split between rural and urban British Columbians. Much hubbub was made during the referendum campaign over how STV would harm the larger Northern and Interior ridings, whose representatives would be drawn solely from the small (in area) population centres like Prince George or Kelowna, which would ignore the vote-deficient rural areas. Those fears appear to have translated in the currently rural ridings into votes against STV on the ballot.
The 2 ridings that the Liberals won by about 30 percent that bucked the anti-STV trend were Vancouver-False Creek and West Vancouver-Sea to Sky. They should theoretically have received around 33 percent support for STV, yet actually got 47.71 and 45.2, respectively. This can be explained with the rural-urban theory fairly easily, V-FC is a very dense, highly urban riding, and WV-SS, while large in area, has a population mostly concentrated in West Vancouver.
The Liberal ridings of Oak Bay-Gordon Head and Saanich North and the Islands over-performed their STV vote by about 8 percent, again possibly due to the urban nature of the Victoria ridings. Vancouver-Fairview and Vancouver-Point Grey both passed 50 percent in STV, beating most NDP ridings, despite being won closely by the Liberals. And while Point Grey is larger in area, both ridings would have been part of the Vancouver West STV riding, with 6 MLAs and little concern for community under-representation.
The Liberal’s biggest blowout was in West Vancouver-Capilano, which the Liberals won by 53.26 percent. It came nowhere close to passing STV at only 36.22, yet even that is overperforming the trend, which should have placed the riding’s support for the system at around 20 percent. Once again, a densely populated riding.
The NDP’s anomalies were more, well, anomalous, and don’t fit into the theory as well. Alberni-Pacific rim underperformed it’s STV vote by several percent, getting only 44.55 (as opposed to the trend predicted 60 or so). That can partially be explained by a rural nature, as the riding would have fit into the extremely large North Island-South Coast riding.
On the other hand, Surrey-Green Timbers and Surrey-Newton, which received around 26 percent STV support, drastically underperformed the average, and for no good apparent reason. Neither riding is particularly large, nor rural, so perhaps there were extenuating circumstances?
In any case, I’ll make another chart for population density vs. STV vote in another post so this theory can hopefully be more than babbling narcissim.
