* * *
I have covered three provincial elections so far: Ontario (124 districts), Quebec (125) and Alberta (87).
In Ontario, the model identified the right winner in 111 of 124 districts (90%), including all 53 "safe" districts. Out of the 13 wrong winners, 11 were within the margin of error. Only two were complete misses.
In Quebec, the election was quirkier (the CAQ won a majority, the first non-liberal or non-PQ government in my lifetime) and the polls were less precise than they had been in past Quebec elections. Nevertheless, the right winner was identified in 112 districts out of 125 (90%). Out of the 13 wrong winners, four were within the margin of error. Nine were complete misses ("not my fault liberal voters stayed home in Gatineau," I have grudgingly mumbled on a few occasions since).
In Alberta, the right winner was identified in 82 out 87 districts (94%), including all 38 "safe" districts and 29 of 30 "likely" districts. Out of the five wrong winners, three were within the margin of error.
By combining the results of these three election projections, we have the following breakdown:
The model correctly identified the winner in:
- 98% of "safe" districts;
- 94% of "likely" districts;
- 80% of "leaning" districts;
- 64% of toss-ups.
... for a total success rate of 91% (305 of 336). With such a success rate at the federal level, the model would miss about 30 of 338 districts.
Of course, I aim for perfection, but it is statistically unlikely - and would also require a fair amount of luck. I believe 90% is a fair benchmark.
There will be a full federal projection update on Sunday, as usual. Have a great weekend!
Philippe J. Fournier is the creator of Qc125 and 338Canada. He teaches physics and astronomy at Cégep de Saint-Laurent in Montreal. For information or media request, please write to info@Qc125.com.
Philippe J. Fournier est le créateur de Qc125 et 338Canada. Il est professeur de physique et d'astronomie au Cégep de Saint-Laurent à Montréal. Pour toute information ou pour une demande d'entrevue médiatique, écrivez à info@Qc125.com.