Tuesday, February 20, 2007

The Conservative "Surge": Perception and Reality

I usually do not comment on polls. I find the media and the pundits so misread what most polls say that if I were to comment on all of their mistakes I would never get any of my own work done. However, the characterization of the latest Strategic Council poll was so outrageous I could not let it pass.

First the perception: There is not doubt that the poll seems to indicate a drop in Liberal support. That is helped along nicely by the overcharged rhetoric of CTV and the Globe and Mail. If I believed in such a thing I would say they they are biased against the Liberals, but in reality it is probably just them trying to sell some sizzle with a rather thin steak.

However, the poll and their analysis is not good news for the Liberals because it does create the perception of decay in support. No party can like being in that position when an election is imminent.

However, most Canadians do not live and breath politics so this poll will be quickly forgotten by them as they focus on other things they consider more important, although the perception the poll generates may linger.

That lingering perception has a nice silver lining, which I will go into in greater detail later.

The reality: This is a new poll that was conducted using a different methodology from previous polls and a different questionnaire from previous polls.

All polls are stand alone entities. They are a snapshot in time. It is possible to take a series of polls that use the same methodology and roughly discern a trend but to do so is fraught with pitfalls because they use different samples every time so they are not truly longitudinal surveys.

So when a pollster switches polling methodology they break the chain and they cannot state the new poll indicates a change in a 'trend' with any degree of validity.

Strategic Council demonstrated this themselves during the last election. At the beginning they were placing the "Party Support" question (the money question because that is the question the poll sponsor is really paying for) at the beginning of their poll questionnaire and the results showed a neck-in-neck race. Half way through the election they buried the money question in a new questionnaire and had it preceded by a bunch of questions regarding government accountability. Naturally, the results of that new poll resulted in better estimates for the Conservatives. So much so, they wound up in majority territory even though no other pollster put them there. Of course, they published a poll just before election day that was in line with other pollsters and I think the reason why is they put the money question back at the beginning, although I cannot prove it because I could never find that last questionnaire.

Both methods are statistically valid but needless to say that all things being equal you will get markedly different results for each poll and you cannot say with certainly why the results are different.

So no one can say the Conservatives have surged because the new poll is disconnected from previous polls. As well, changing the methodology of a poll will effect it results and you can never completely calculate what effect the change of methodology has on the results.

So while it is valid compare the results of this poll to the previous poll and state the Liberals are down a certain amount and the Conservatives are up a certain amount, you cannot say with absolute certainty that the change is truly the result in the change of support and not the change in methodology. And it is certainly a rather large mischaracterization to look at this poll and say the Conservatives have "surged".

Reality point number two: Despite the new methodology and questionnaire the difference between the two big parties in just 5 points. This is also despite the good results on the internals of the poll. Mr. Gregg, actually said something that was telling and I really wonder if he meant to say it. He stated, and I am paraphrasing, "Alot of Canadians respect Mr. Harper but they still do not seem to agree or like his policies". Interesting statement and very logical. I may like and respect the salesman but if I do not like what he is peddling I am still not buying.

I would also point out that the internals are not much different than they were last year when it was PM Martin facing Leader of the Opposition Harper. The sitting PM always has the better internals in between elections but it can change in a heartbeat once the writ is dropped. Just ask Mr. Martin about that.

Reality point number three: The Liberals have been hovering between 30-34% for almost a year. This is the first poll that puts them below that in a very long time. Considering what I stated above we cannot say with any certainty why the level of support has changed. The change in methodology could be a factor, although considering the magnitude of the drop it is reasonable to assume some of it is a result in the erosion of support. Why that support eroded is a whole new argument and this poll would not be very useful in such an argument.

Reality point number four: The undecided estimates are at 23%. Yikes.

Unintended consequences of creating the perception of the decay of Liberals support: This is the silver lining.

The pressure is now off of the Liberals and Mr. Dion and it is now placed squarely on the shoulders of Mr. Harper. This poll could have the effect of lowering expectations amongst the electorate for the Liberals and Mr. Dion. Therefore, if they can mount a reasonably competent election campaign they will exceed those expectations and grow their support.

That is what happened to Mr. Harper last year. No one expected much from him or the Conservatives. Yet he did run a competent, if uninspired, campaign and managed to eke out a victory.

Conversely, Mr. Harper will have to run a more than competent campaign to grow his support.

This does not mean that the Liberals should be complacent. For example they, should take the results of this poll to heart and present a united front on the issues. They no longer have the luxury to be divided over such esoteric issues as the extention of some of the provisions of the anti-terror laws. Talk about it, debate it, but once the leader makes a decision accept it and move on.

As well, the perception this poll could generate also means they will need to work hard to make sure they can run at least a competent campaign and maybe even run one that Canadians can find a little inspiring.


Blogger alaroz12 said...

Hello! You have a very nice blog! I'm here to share valuable info with you visit my blog,about Mozilla Firefox web browser.

February 20, 2007 6:32 PM  
Blogger Steve V said...


That is the best breakdown I have read, including my own.

"the poll and their analysis is not good news for the Liberals because it does create the perception of decay in support. No party can like being in that position when an election is imminent."

This is why polls matter, even if they are erroneous, or people don't put much stock in them. Like it or not, the MSM takes many of its cues from polls. It shapes the debate and frames people.

February 20, 2007 8:04 PM  
Blogger ottlib said...


It is not good to underestimate the impact of polls but it is not good to overestimate them either.

That is why I give a little piece of advice at the end of my post.

However, I would point out that polls are not an indicator of the future voting intentions of anybody. If that were the case Paul Martin would have won a majority government in 2004 as he was leading with over 40% of the vote just weeks before the election call.

February 20, 2007 8:34 PM  
Blogger Red Tory said...

Ottlib -- A very well-reasoned analysis of the poll. I’m not terribly concerned about the seemingly unfavourable results provided that it doesn’t become indicative of a downward trend. I hope that this does, as you suggest, take some of the pressure off Dion by lowering expectations, but it could also lead the media to a contrived “death spiral” interpretation where their attention is focused on internal grumblings and divisions within the party as various factions express their frustration at the party’s disarray and apparent lack of coherent vision.

February 20, 2007 9:14 PM  
Blogger s.b. said...

Yeah i would anser Stephen HArper to their question. It doesn't mean I would vote for him.

The only valid poll would be one that asks only one question with randomiozed order of parties.

"Who would you vote for if an election were held today."

That's it nothing else. Any other questions added or asked in any other way skew the results.

Strategic council at least has a representative name but their polls are just that strategic.

February 21, 2007 8:05 AM  

Post a Comment

<< Home