Saturday, April 16, 2011

Some Latent Empirics that Should Have Conservatives Worried

I've been looking at EKOS second choice data and analyzing it in social networking statistical software {UCInet 6.198}. The idea is to examine the pattern of choices and second choices to see which party represents the "core" of the Canadian political zeitgeist. What drives coreness {as opposed to being on the periphery} is relative support and the ties through second-choice support of other parties.

I looked at the 13 April 2011 EKOS data [pdf] and found that the Conservatives represented the core, but the Liberals are just outside and the NDP is a little further out than the Liberals. The coreness numbers are on a 1-100 index with cutoffs for being in the core versus the periphery. Specifically, scores above 56.7 {mean + σ} place a party in the core and scores below 10.6 {mean - σ} place it in the periphery::


As stated above, the most recent EKOS data has the Liberals at 50.8 and close to the core. The Bloc, on the other hand, is very close to the periphery. The drivers of these numbers determining the Conservatives to be the "core" are the following::
  1. The overall Conservative support of 33.8% {over 28.8% for the Liberals & 19.0% for the NDP}
  2. The relatively strong loyalty of the Conservatives with 45% with no second choice
  3. The high degree of crossover between the Liberals and the NDP {43.2% Grits supporting Dippers & 40.5% Dippers supporting Grits}
So, if these numbers shift, the coreness numbers shift. The most likely to change dramatically between now and the election...#1.

What should alarm the Conservatives is a comparison of these numbers to those post-prorogue and pre-Olympics in February of 2010. These are taken from an EKOS poll with second choice data from that era [pdf] with coreness above 56.8 and periphery below 10.2::

               Bloc 10.7                   NDP 37.5         Liberal 53.5        Conservatives 70.5

  periphery< |----------------------- 33.5-----------------------------| >core
             10.2                            mean                                56.8

           Canadian Political Party Coreness, 11 February 2010, EKOS/CBC data
                                  mean= 33.5, standard deviation {σ} = 23.3

That post-prorogue era when the Liberals and Conservatives were statistically tied was a relatively dark era for the Conservatives. The Liberals were within a few points of being in the core. The current situation is similar, if not worse. Why? If the Liberals enter the core or if the NDP, Liberals, and Conservatives share the core, it is indicative of political fragmentation reaching a point where there is no dominant party, i.e., the three federalist parties lack a critical mass of support based upon their current positioning.

Moreover, the current situation is worse than where the Conservatives were right before the election in 2008 [pdf]::

  Bloc 11.76                              NDP 39.1      Liberal 46.3                         Conservatives 73.6

    periphery< |---------------------------- 39.7---------------------------| >core
                19.1                                mean                             60.03

                   Canadian Political Party Coreness, 13 October 2008, EKOS data
                                  mean= 39.7, standard deviation {σ} = 20.6

On the eve of the 2008 election, the Conservatives were alone in the core by far. The second choice data are not broken down by region, which would be really informative. {All of these analyses are based on Canada-wide numbers}. 

Strategically, this presents a challenge for all parties. The Conservatives have loyalty and plurality on their side right now. They need to maintain their base and prevent defections, i.e., predominately to the NDP in BC and to the Liberals in Ontario and Atlantic Canada.

The Liberals and NDP should be heartened that they are each close to the core, but nevertheless are bound by the first-past-the-post arithmetic. Catch-22.ca, an anti-Harper site has compiled a list of vulnerable Conservative ridings and opposition-held ridings being targeted by the Conservatives. This outlines where the battle lines are. TooCloseToCall.ca has a best-case seat projection for the Liberals based on current polls:: CPC 126 LPC 112 NDP 27 GPC 0 & Bloc 43One of the issues is that the NDP is much more efficient than in the past. Of course, if the Conservatives start slipping in the polls, all bets are off. 

I think the key takeaway here is that the current underlying preference structure as measured by coreness, which I freely admit is a crude metric, has all three parties converging. No one party can cut through the clutter. The Conservatives had a much stronger position right before the 2008 election. The current numbers are similar to those post-prorogue in February of 2010, with a twist—the NDP is close to the core, as well.

Twitterversion:: [blog] EKOS data shows curious emerging pattern of #cdnpoli zeitgeist {2008-elxn40, post-prorogue 2010, & now- #elxn41}  @Prof_K

No comments: