07 June 2015

Q2 2015 Poll Speculation #nlpoli

Corporate Research Associates boss Don Mills has done a good job of teasing the results of his latest poll, due Monday.

"Significant" change in voter intentions, Mills tweeted on Friday and repeatedly over the weekend.

It's all fed a great deal of speculation.  Someone fed the self-styled Hydroqueen internal Liberal polling numbers and she has blogged them and tweeted about them repeatedly. Your humble e-scribbler jumped into another conversation based on the foggy early-morning memory and since that memory was so horribly wrong,  here's a review of the recent poll numbers based on more than memory.

So are those Hydroqueen numbers the sort of results CRA will release?

About how the predictions of further Liberal decline or of a Conservative rise?

Will CRA show any of that?

Probably not.

It's not that the internal Liberal polls are wrong, it's just that is always some variation between polls any way. On top of that, you've got to bear in mind that whether it was the NDP rise in the 2011 election or their surge in 2012,  CRA often misses things.  Their polling margins are notoriously wide.  That means that sometimes subtle shifts in opinion can go by them.

Just to give a refresher, here is the CRA polling, right up to last quarter.  The numbers show the results as a share of all responses, not as a share of  "decideds".  That makes it easier to compare CRA with other polls.

What you see in Q1 2015 are changes in Liberal, NDP, and Conservative numbers that are about half the CRA margin of error.  That's a really subtle shift for CRA and one that automatically makes you a little leery.

Look at the numbers for the quarters before that you see basically flat lines across the board.  There are a couple of odd blips, one for the Liberals upward and one for the Conservatives downward.  They look odd and there isn't any obvious reason why either would have happened.

What you can see from CRA are very broad trends.  Tories have been running downward for quite some time.   The NDP have dropped back and the Liberals have benefited, ultimately from all the shifts.  So overall, the CRA poll should show the ranking from top to bottom of Liberal, Conservative, and NDP.

Based on the more recent CRA polling, there'd be no shock of the Liberals were in the mid 30s. That's where they have been for most of the past year and a bit.  That would make the late 2014 jump an anomaly.

Well, it would be no shock for people following the numbers. For everyone else, though, that kind of drop would cause massive ripples among the chatterati who think that CRA is precisely accurate.  They'd be screaming about it.   And that screaming would energise the Conservatives and NDP.  It would also demoralise some of the Liberals.

Let's put a bit of leavener in all this, though.  Remember the margin of error.  Any given number in any CRA quarterly poll could be five points higher or five points lower than it is shown.  This is another reason why it is important to take the "decideds" torque out of CRA's numbers.

Look at the Tory and NDP numbers for the past year or so.  Tories are roughly 20 and the NDP have been roughly 10.  That gives you extreme spread one way where the Tories are at 25 and the NDP are at five. The extreme spread the other way would put the Tories and NDP in a tie at 15.  So yeah, if the CRA poll has been consistently off reality by that second spread,  it might not be surprising to see the Blue and the Orange - or is it Pineapple? - switch places in the next poll.

The Conservatives haven't exactly had a great winter.  Judy Manning.  A budget poorly sold.  Don Dunphy.

NDP doing relatively well with Earle McCurdy.  Maybe enjoying a bit of a bump from an upswing in federal fortunes.

Now look at the Grits and the Tories.  Upper end of the Tory range is 25.  Lower end of the Liberal range , without the past two quarters, would be... 30.  Five points apart.  The Liberals and Conservatives could have been within five points of each other for the past year.  CRA polling could have missed it consistently and yet the CRA poll would still be considered "correct" because the actual result is within the margin of error.

As a last little point,  let's go back to February and compare three different poll results.  All used slightly different methods and all were in the field around the same time.  Look at the variation in numbers that are all presented on the same basis.

The NDP averaged eight points.  Two polls had them at seven.  One had them at nine.  That was the smallest range of variation.

The Conservative results varied by six points, from a low of 20 to a high of 26.  CRA at 22 is only one point off the average of the three (23).

The Liberal results also cover a range of six points.  In that string of results, MQO nailed the average result bang on.

Based on that, where exactly would you guess the CRA results on Monday will be?

You are right.  Whatever you guessed,  you got it bang on.

That's the thing about speculation.  It's a game everyone can play.