Saturday, April 09, 2011

Examining EKOS Canadian Federal Voter Intent 6 April Data:: Setting Up a Showdown

EKOS January13—April 6, 2011 poll with regional breakdowns,
click for larger image
The most recent EKOS poll [pdf] should have the Conservatives worried. {I consistently use EKOS since it's easy to get their reports and the last reported "house effects" aren't that substantial}. The Ontario numbers are closing and the Liberal are rebounding in BC and Québec. Atlantic Canada remains highly volatile.

The real problem, as pointed out by the EKOS report, are two clashing findings::
“'Everything is fine and why risk the adventure of rocking the boat with a new government' versus 'things may be fine but the government isn't and it's time for a change of management'”.
I think this sets up the debate as one of credibility, particularly on managing the economy, which I think will be important in bellwether Ontario. There are several realities. First, the public tends to recoil at talk of a Conservative majority. {Part of this is the campaign Harper is running. It's risk averse, which is to be expected, yet, gaffe prone with debate challenges that are later withdrawn, the campaign fortress or bubble, and the screening of rally attendees are creating the wrong kind of attention}. Second, if Harper is looking at minority territory, which is the increasing consensus, on the minds of the electorate will be a possible coalition. No matter what Michael Ignatieff says about coalitions, if the House of Commons has a similar seat composition that was present at dissolution, the same contentious issues are still to come—including the budget, which Flaherty said would be the same. Will this force a shotgun marriage? That would be interesting—and without Ignatieff actively seeking it. At any rate, voters may be thinking of the Liberals and New Democrats as paired and if they appear to be credible of managers of the economy, the Conservatives' fortunes may slip considerably. The Liberals and New Democrats need not be in lock-step collusion for the rest of the campaign, merely show that their platforms share more than are in opposition of each other. The Conservatives have few degrees of freedom, since the Conservatives like to spend money, want to appear to be on top of the deficit, and have a policy of being pro-corporate business. The less they say on specifics, the better, but I'm not sure the public will buy the standard spiel and the mantra of "trust us".

Second Choices
The last time I looked at second choices was with the March 10 data [pdf]. I look at second choices to see how voter preferences might change, as a second choice indicates a shared ideological space among parties.

ELOS Second Choice data, April 6, 2011
The pattern is similar to the March 10 data. I input the percentages into UCInet {v6.198}, a social networking software package, and looked at multiplicative coreness, which examines core-periphery. I also input the data into a graphing software {Netdraw 2.081}. Examining core-periphery gives a crude sense of where the zeitgeist of the electorate is. {I used the above percentage data in order to not let the variance drive the results when using a distance algorithm. I set "no second choice" as a party's second choice with itself. This captures a loyalty effect. I also controlled for relative party support, by multiplying the second choices for a given party by its the overall voter intent percentages, i.e., 37% CPC, 27.8% LPC, 16.1% NDP, 9.3% GPC, 6.9% BQ, and 2.9% Other}. The higher the coreness score, the higher a party represents Canadian political consciousness. The results of the analysis follow::

This shows how the Conservatives represent a core with its relative support and limited shared ideological space with other parties.  The March 10 coreness statistic for the Conservatives was .738 showing a consistency, which isn't surprising given the poll numbers. The Liberals and the NDP are further out from the core, while the Bloc is at the margins at 0.09. It would be interesting to see how this changes over the campaign. If the Liberals and NDP manage to increase their percentages of voter intent while still sharing a high level of support with each other as second choices {note Liberal support of the NDP is 41% and NDP support of the Liberals is 34.8}, they could represent a new core. 

Perhaps more interesting and useful is the network diagram of second choices::

Second choice patterns from EKOS 6 April 2011 Poll

Nodes represent federal parties. Inbound links (figures closest to each node) represent percent support of voters who support another party for first choice who support the focal party. Example: On the line between the Liberals (LPC) and the NDP, the 41.0 represents the number of Liberal supporters who consider the NDP a second choice. The size of the nodes represents the percent of federal vote intent. The thickness of the lines represents the second choice relationships between the parties. 
The pattern is similar, when compared to the March 10 poll, although the Conservatives are losing support as second choice. The Liberals and Conservatives are sharing less ideological space, while the NDP and the Liberals are sharing more. Almost 20% of Green supporters consider the Conservatives a second choice, while 24% and 25% consider the NDP and Liberals to be second choices, respectively. If strategic voting takes hold and Greens switch allegiances on election day, the current voting intent support of 9.3% may shake things up in a few select ridings. On the other hand, Bloc support of the NDP as a second choice may give the latter the four seats they're coveting in Québec, including holding on to Mulcair's Outremont. If the Conservatives go into freefall, second choices indicate the Liberals and the NDP would evenly split the defectors, with a slight edge to the NDP {20.1% vs. 17.3%). I'd love to get regional breakdowns, but even with access to those numbers, the relatively small sample sizes would make me nervous about making inferences.

The Conservatives need to hold their ground and its still their election to lose. They've currently slipped out of majority territory and the gap is closing in Ontario. The economy is the Conservative's strong suit, but if the Liberals and NDP appear credible, that advantage can evaporate. The x-factor is corruption and scandal. The EKOS poll indicates that the electorate is increasingly unhappy with the government. If allegations that the Conservatives are corrupt stick, approximately 27% of current Conservative support might shift to the Liberals and the NDP and their turnout might fall. Additionally, a dramatic fall in the polls may demoralize the Conservatives and their supporters, which could adversely affect their campaigning and their turnout, which is typically high.

Twitterversion:: [blog] Analyses & graphs of April 6 EKOS #canpoli #elxn41 data. #CPC still in driver's seat, but vulnerabilities abound. @Prof_K

No comments: