Welcome to my blog. I learn so much from others who share with me their thinking, opinions, experiences, photos, music, pains, happiness, and so much more. In turn, I like sharing too and especially enjoy engaging and thoughtful dialogue. So don't be afraid to tell me what you think. If you want to know more about me and my work, I hope you will read a few of my posts and visit corestrategies.ca.  

I have the deepest respect, I really do, for some my colleagues in the polling and market research industry. I talk about some of these people on my core strategies website.

A few others, however, seem obsessed to the point of near madness, with who allegedly has the most accurate data. Increasingly, the measure by which many pollsters fixate as being "the best" seems to be who can boast, after-the-fact, as having the most accurate data by infinitesimal margins 19 times out of 20.

There is an implied competition, I guess, that he or she who can beat his or her chest the hardest and therefore the loudest, is, I further guess, the default winner of all future polling contracts between elections. Good for them.

It's a hell-of-game and a high-stakes gamble some of these players indulge in. And for the guy or gal consistently on the wrong side of a 5 or 10 point margin-of-error compared to actual election results, the industry mocking can be pretty brutal.

Don't get me wrong, bragging rights for being the most accurate has its tangible and intangible rewards, and I'd be lying if I said I don't give myself a pat-on-the back every time a Core Strategies poll is bang on. On that score, I completely understand what an Angus Reid, an Ipsos-Reid, or Léger-Léger...etc. are all about. But please understand those folks occupy and are motivated by a completely different space, in my opinion, than say a Decima, a Strategic Counsel, or even a Core Strategies.

Far too many focus groups later than I care to admit, I suppose my obsession is with insights into what people really think, why they think it and above all what I can help my clients achieve with the new-found knowledge. That is the sandbox I like to play in, as opposed to simply knowing for absolute one hundred percent certain that 84.0136% of the population loads the roll of toilet paper over versus under.

This is why, in recent months, even as I served as Green Party strategist, I've been thinking hard, really hard, about the underlying meaning and factors affecting political polls the past few years.

The volatility has been spectacular by comparison to what we saw when I first got into this business 20 years ago when I left Statistics Canada and headed for Decima as Allan Gregg's VP of Operations. Back then I watched and learned from the grand-daddy of our business, where if the numbers moved more than a few points in 6 months, out came the "holy-F" bomb.

By comparison, look at what we did in the London by-election. Twenty points behind and only 10 days to work with. Thanks to, among other things, a little strategic polling aimed at helping us really understand the "what" and the "why" of public opinion, suddenly it became possible through proper messaging to move the numbers by the almost 20% needed to win. The modeling we did near the end of the campaign suggested that, tops, another 3 or 4 days and Elizabeth May would have won the London by-election last November. Think of it, 20% in 10 days. Virtually unheard of prior to this.

The post analysis of how and why London happened the way it did is another story for another day, including the fact that yes, it was a by-election, and whacky things can happen in a by-election that are harder to duplicate - on purpose - in general election. Note, I said harder, not impossible.

But today, I want to come back to the main purpose of this blog entry. The volatility we see in political polling data one week to the next.

Once again, I will refer you to yet another gem by Chantal Hebert and invite you to read her May 2nd column, PM on defensive, Dion in hot seat. There simply isn't a better micro-analysis description of why numbers may or may not be shifting beneath us and as we speak. The analysis is bang-on and reflects perfectly the micro-dynamics at play. Decima's latest numbers on May 1st, Conservatives and Liberals Stalemated, are also proof-positive the numbers support the analysis and vice-versa.

But allow me, if I may, to introduce a larger macro-level factor I think toys with the numbers in a more fundamental manner. In my blog of April 18, I wrote "Not to worry, short of anymore really bad blunders, if the numbers get too strong in favour of a Harper Majority, the numbers will correct themselves soon enough."   It's been only two weeks since I wrote that, and already we are seeing the correction taking its effect.

The point I was trying to emphasize, was an Adam Smith like invisible-hand I think governs, or helps ensure, the numbers don't go too far one way or another. Canadians don't actually want a majority, either way, and I can prove it with two pieces of data. The first collected by Core Strategies this past January, and the second just collected April 26-30.

In January, we conducted a national poll of 1500 Canadians, and among other significant discoveries we learned that public opinion was evenly split in its desire for either a Majority or a Minority government. Here is what we asked, and found:

11. Thinking about the outcome of the next Federal election, would you like to {read and rotate} :

1. see another Minority government because it forces all the parties to “play nice” and to “reach compromises.”

                    Or would you prefer to

2. see a Majority government where one party can more-or-less govern for 4 or 5 years as it sees fit?

 

      

 

  Assuming you accept the notion Canadians are split between wanting a Majority vs. a Minority, this begged the question, well then, what do Canadians really want? Plagued by the volatility and seemingly whimsicalness of the numbers, I was convinced there is as much going on at a macro-level as there is the micro level affecting the numbers. Core Strategies set out to find out what and why.

Here are the results, and they are astonishing!

In my opinion the results turns Canadian politics on its ear, and with it political polling. It does not, yet, change the election process, but it sure changes the process for getting elected. The wisest of strategists "out there" will immediately figure this out. Those with their heads in the sand will remain so. Consider this...

Pollsters have been trained to ask the only question that ever seems to matter; how will you vote? Everything stems from there, even the 2nd choice vote. After that we have countless hours of interviewing and focus groups that attempt to answer the question why and what are the factors at play.

You're either a leader, or a wimp. Attack or take the high-road. Kyoto or pollute. Further the fight in Afghanistan and Iraq or retreat. Tax income trusts or invest. Tough on crime or human rights. Wear makeup or be a real man. Gay marriage or missionary position...etc.; All important attributes to become intimately familiar with, including how shifting perception, one way or the other, affects voter intention. Go ahead, knock yourself out.

As for me, about a month ago while taking a much needed vacation at our cottage near Mt. Tremblant, I thought of something else. A completely different tact.  A barbecued steak, a bottle of wine, and hours later staring at flames in the fireplace, this is what I concluded...

Stop focusing so much asking Canadians how they intend voting, and instead ask what they want the outcome to be. Because the difference between what Canadians want is vastly, vastly different from what they end up with. And so long as we have a disconnect between what Canadians want versus what Canadians get, we will have political instability and polling volatility.     

So following my stint as being the "most expensive person in the system...but worth every penny" I decided to invest my millions on a couple of omnibus questions through Ottawa based Opinion Search.

The question we asked was not 'how do you intend voting?' although we asked that too, but rather 'how many seats do you want each of the major parties to have when all is said and done.' To my knowledge, this question has never been asked before. At least not that I am aware of.

It's a question Core Strategies intends tracking. It's a question I believe tells us far more about Canadians, including what Canadians are thinking and want, other than simply asking "so who ya gonna vote for?"

 

 Survey says...

First, I needed a practical field question. Asking 1000 respondents to divvy-up 308 seats is a tall order. Accordingly, this is the question I used:

1. The Federal House of Commons is made up of 308 seats. To make the math easier, imagine the House of Commons is made up of only 100 seats. If it was solely up to you, what number of seats for each of the major political parties do you think would be ideal, between the: {read and rotate and then assist respondent in adding up to 100}

1. Conservatives
2. Liberals
3. NDP
4. Green
5. Bloc Quebecois

 

... this is what Canadians say they want

		Seats	%
	Conservative	97	32%
	Liberal	81	26%
	NDP	49	16%
	Green	42	14%
	Bloc	39	13%
	Total	308	100%

The results above are interesting enough on their own, and those looking to poo-poo the methodology will be quick to point out:

1. The resulting percentages are within the margin-of-error of the traditional vote question, so why bother?

2. Choosing seats per party isn't a choice Canadians get to make, so why bother?  

And there lies the difference between this pollster/strategist and others.

My interest is NOT in acquiring bragging rights for the smallest margin-of-error between how people say they will vote versus how they end-up doing so. We already know we are good at that. Indeed, for everything ranging from policy issues to campaign strategy, I am far, far more interested in how many seats say a Conservative voter wants Liberals to get and vice versa.

'What?' I can already hear the Conservative war-room asking. 'Are you trying to tell me there's some good-old Conservatives out there who didn't say they wanted all seats in the house to go to Conservatives?'   Yup, that's right Doug.

And so, here is Core Strategies gift to the political parties, media, and pundits; exactly how did the data fall out? How many seats for example do NDP voters want Liberals to have? How many Liberal voters want Liberal seats versus say Conservatives want Conservative seats, and what does this latter equation say about the strength and the commitment of the Liberal voter to voting Liberal?

 

		TOTAL			CON		LIB		NDP		GREEN		BQ
					seats	%	seats	%	seats	%	seats	%	seats	%
	CON	97	32%		171	56%	69	23%	53	17%	67	22%	59	19%
	LIB	81	26%		58	19%	125	41%	66	22%	64	21%	61	20%
	NDP	49	16%		32	11%	45	15%	97	31%	51	17%	39	13%
	GREEN	42	14%		24	8%	39	13%	51	17%	92	30%	37	12%
	BLOC	39	13%		23	8%	30	10%	41	13%	35	11%	114	37%
		308			308		308		308		308		308

 

After collecting and preparing this data, I've bounced it off a few colleagues. The reaction has been the same from nearly everyone. Wow! But why on earth would you give such precious data away? Good question. Maybe I'm feeling a little too generous after reading Don Tapscott's Wikinomics.

Or, maybe it's my view that any of the reputable pollsters and strategists can dream-up and ask such questions. The difference, as I see it, is that not everyone knows what the numbers really mean, not everybody understands why the numbers are what they are, and above all not everybody knows how to move the numbers like we did in London, or better yet, how to change the political structure to match what Canadians want, not what politicians want Canadians to want.

Like I said, that's the sandbox I play in.

May 4, 2007  

*** Press Release click here ***  May 4, 2007 -  Canadians want vastly different seat distribution in House of Commons, poll shows

Canadians want vastly different seat distribution in House of Commons, poll shows