22 December 2007

The secret of their success: Roger Grimes

Polls is fun, as Pogo's official opinion survey guy used to say, and there are no polls with more fun in them than those available in Newfoundland and Labrador.

Professor Alex Marland, latterly at Memorial University and formerly a communications director with the Williams administration, uses Corporate Research Associates quarterly omnibus poll results as part of a paper published recently in the spanking-new journal, the Canadian Political Science Review. [Almost immediate update/clarification:  Marland was looking at the 2007 general election and used the poll results for context. What follows is purely the analysis from the Pogo Poll People.]

The figures are presently straightforwardly, except that Marland or his editors incorrectly state that the margin of error for the data is plus or minus 3.5% at the 95th confidence interval.  They aren't.  In the period since 2005, for example, all but two of the quarterly results have margins of error of plus or minus 4.9%.  CRA doesn't archive their new releases anymore back beyond early 2006 so there is no way of double-checking these sorts of details.

pc sptIn any event, while we may dispute the accuracy of the CRA polls, for the moment let's accept them as readily as they have been accepted publicly to date.
The results for polling conducted from the last quarter of 2003 to the first quarter of 2006 show one thing:  volatility.

Take a look at the table: up and down dramatically in very short spans of time.

The drop from the first post-election poll to the second is only a couple of months and already, the support for the progressive Conservatives dropped 16 points.  It fell another 11 in the next quarter, giving a cumulative drop of 27 points in basically less than six months.

No surprise, though given that Danny Williams first speech as Premier was the January 5 2004 debacle, followed immediately by the delivery of the budget to implement the speech's promise.

Once the budget was delivered and the world didn't end, there's a slight rebound.  The big jump in the last quarter coincides, not surprisingly with the October theatrics in which the Premier stormed out of a first ministers' meeting.  The trend is still upward, mind you, but notice that the climb out of the pit is not nearly as dramatic as the fall. 

Then in early 2005, there is the famous $2.0 billion transfer payment from Ottawa.  That's the one that has been hyped to death by any supporter of the current administration. Not surprisingly, the poll taken shortly afterward shows what was then a record level of party support.

Then look at what happens.

There's a drop of 20 percentage points in a single quarter. Quarterly increases of six percentage points, then 12 and 13 are followed by a single drop of 20. Not surprisingly, support for the other parties rises and falls in the same time, most noticeably for the Liberals. 

Notice though, that over the next three quarters, right up to early 2006 there is a steady climb in Progressive Conservative support and a corresponding decline in expressed Liberal support. The climb/decline is nowhere near as dramatic as the tumble in early 2005.  One year later, the Tories had not managed to climb back as high as they had been in that first quarter of 2005.

But go back and wonder what happened in early 2005?

In the absence of any major event, the sudden drop in early 2005 looks like a simple correction back to a norm.  The average PC party support over the previous quarters is 56.8%.  For the Liberals, the number is 28%.  In the second quarter of 2005, CRA polled 55% for the Tories and 31% for the Grits.

At the same time, though Roger Grimes resigned.  After that the support for the two major parties diverges. 

There are no dramatic increases or decreases, even allowing for the poll margin of error.  There's just a steady climb for the Tories until the numbers stabilise in the low to mid 70s. If you accept CRA numbers, that's basically where things stand today.

grimes2For all the effort, nothing has been able to secure the Progressive Conservative Party lead like Roger Grimes.  Take him out of the picture - and by extension generate the Liberal leadership issue - and voters make a clear choice.

Before that, while Grimes was still around, there was an evident willingness of respondents to support the Liberals and, perhaps more interestingly, to take their support away from the Progressive Conservatives quite dramatically.  They didn't give it back quite so enthusiastically.

So as you sip your hot beverage on a cold winter's evening, ponder this thought:  Roger Grimes is the secret to the Progressive Conservative electoral success.

The polls can't be wrong, can they?