Thursday, April 28, 2011

Some thoughts on the polling in BC in this election

I am a political junkie and even worse a poll junkie.  It is truly like a drug and polls being the crack of political junkies.   I try not to pay attention to the polls, but I can not help myself.    Anyway, I am back with some thoughts on the polls in BC.

I talked with Frank Graves of Ekos on Twitter and pointed out their sample sizes from BC were small and often smaller than those from Alberta even BC has many more people.  I pointed out that BC has a lot of seats in play and Alberta has maybe two in play.   He took my point and they have increased their sampling in BC.

The latest Ekos poll has a sample size of 322 in BC.  This is means a margin of error of +-5.5%

  • Conservatives 35.7%
  • NDP 28.6%
  • Liberals 19.5%
  • Greens 13.4%
The NDP and Greens are both up, the Liberals are about even and the Conservatives are down a fair bit. Broadly these numbers were the same in the recent Forum poll, but I do not know the exact size of their BC sample.  

Meanwhile Angus Reid has some rather different results.   A week ago they did the poll with the biggest sample of anyone in BC in this election.   The results were broadly similar in their very small sample from their national poll this week.
  • Conservatives 42%
  • NDP 32%
  • Liberal 18%
  • Green 6.0%

The biggest difference between the two polling trends are with the Conservatives and the Greens.

Nanos is in a third trend, they have very, very small BC samples and they move wildly from day to day, but broadly the Conservatives are over 40% with the Liberals and NDP close to tied at  near 25%, the Greens are very low in BC according to them.

When I take account of some of the new questions that Ekos is asking about depth of support, the shift in the numbers gives me something like this:

  • Conservatives 41%
  • NDP 30%
  • Liberals 19%
  • Greens 9%

Keep in mind this is a very rough parsing of the numbers that has no real statistical legitimacy, it is simply an attempt to quantify how the depth of support data may impact party support levels.

Looking over the history of federal elections in BC, any party that has managed to win more than 40% of the vote has won the majority of the seats.  Several times, 2004, 1993, and 1988, the majority of seats were won by a party with around 37% of the vote.  In 2004, with 36.3% of the vote, the Conservatives won as many seats as in 2008 when they had 44.4% of the vote.  

In the last 13 elections only twice, 2006 and 1972, did no party win a majority of the seats in BC.

So what the polling and electoral history is saying to me is that the current prediction I have for BC, Conservatives 22, NDP 12 and Liberals 2, is broadly correct.

No comments: