We’ve bashed the Corporate Research Associates polling around in these parts for quite some time.
Since since it’s the only game in town, however, it’s what we have to work with.
Here are some quickie observations on the most recent poll:
1. The margin of error for the most recent poll is plus or minus 4.9%. For November, 2008, the MoE was 2.8% and for the ones before that, it was 3.5%. Bear that in mind as you go through media coverage that talks about things being where they were.
2. The chart at left adjusts the CRA numbers to present them as a percentage of respondents rather than adjusting them as a percentage of decideds.
3. If we accept that the undecideds dropped seven percentage points from November, note that both the Liberals and the Progressive Conservatives picked up equally from that. The extra point came from the New Democrats.
4. The most accurate of all CRA polls - measured in terms of the margin of error - was the one in November. The MoE was plus or minus 2.8%.
5. There’s something slightly counter-intuitive about these numbers. At a time of growing economic uncertainty – including some sharp blows in this province – one wouldn’t expect party choice for the governing party to go up. Ditto for the satisfaction numbers which remain at astronomical levels. Look at the satisfaction numbers and it doesn’t get any less curious; you just wouldn’t expect 88% of people in the province to be mostly or completely satisfied with government giving all that’s gone on.
Doesn’t matter the party. Doesn’t matter the leader. When things are bad or going bad, there just isn’t as much generosity toward government. 10,000 ticked off nurses and their families should amount for something, shouldn’t they? How about a 1,000 paper workers? Apparently not.
6. Poll goosing. CRA was in the field beginning on February 9. They stopped polling on February 28th. Oddly enough, when asked about it by a talk show host on February 19, the premier knew CRA had started polling the week before.
He also offered the view that if he was going to goose the poll he would have started goosing the week before polling started.
Interesting.
The Equalization racket started in late January even though by all indications the provincial government had the numbers on Equalization and offsets from early November onwards. Anyone else find that news conference in the middle of the night thing odd, especially since government had already planned to scrum the finance minister the day after the federal budget came down?
Notice all the feisty, fighting talk running into that first week of February. Take that as sort of the background noise to the month. Nothing like a racket with Ottawa to get the juices flowing in polling season.
Then there’s the Lower Churchill project. The infeed to Soldier’s Pond – an integral part of the project – is shaved off for some inexplicable reason and submitted to its own environmental review process starting on January 30. On top of that there were two more major Lower Churchill announcements in February, one of which was merely to say the Crown corporation had solicited six expressions of interest; not received: asked for.
Government decided to try playing nice with the nurses. They did it very publicly in February a week into polling.
Let’s not forget the “historic” infrastructure announcement made – you guessed it – right smack in the middle of polling. That was followed by a few speeches to give the thing a few legs.
Overall, though, February was a very busy month for the provincial cabinet. Lots of good news and happy talk. Marinas, fur farms, air ambulances, all announced in February.
7. Rolling Stones Update: Can’t get no satisfaction? We’ve had some issues with CRA polling before and the tendency to generate results that leaving you scratching your head.
One of those would be “satisfaction”.
Let’s leave aside entirely the problem with the question - what exactly is the difference between “mostly” and completely”? – and look at the relationship between the satisfaction number on the one hand and the party support number on the other.
Without any prompting one might suspect that satisfaction goes with support. If you are mostly or completely satisfied with government performance, then you’d be inclined to support the government party.
Not so, according to CRA. In Nova Scotia, for example, people are very happy with Rodney’s government but they plan to vote for the Grits or Dippers.
We are not talking small numbers. The Liberal and NDP vote numbers here are running fairly steadily at a combined 64% of decideds for the past three quarters. The satisfaction numbers have been running in the 50s. So if you believe CRA, a majority of Nova Scotians like their Tory government but a larger number wants them out of office.
Huh?
It gets fruitier when you look at the Tories coming in second or – in the latest poll – third place among parties. People are happy with the job government is doing but they don’t want to vote for the government party.
Those odd numbers don’t just apply to Nova Scotia. Satisfaction with Roger Grimes was decently in the 50s pretty much right up until the end. Take a gander at CRA’s polling and there’s a good likelihood you’ll find lots of examples of this completely incomprehensible correlation. According to CRA, Nova Scotians like the job a government is doing but want to throw the bums out.
Doesn’t make sense.
-srbp-
7 comments:
I wouldn't worry too much about the Nova Scotia numbers, electoral history benefits the Tories over the NDP when voting day comes.
There are two reasons for this:
The NDP vote is notoriously inefficient, concentrated around metro Halifax, which leads to fewer seats than the vote tallies would normally suggest.
The NDP and NS Liberals have a symbiotic relationship, particularly on the mainland. When one of the parties falters (as the Grits have done since 1999), the other benefits. When they are both strong, they split the vote and elect Tories (sort of like Saskatchewan federally). The CRA poll indicates that the Grits and Dippers are duking it out for first place. This is good news for Rodney Macdonald.
@ herringchoker.
It wasn't really so much the point about whether the Dippers might form the government as much as the apparent inconsistency between the satisfaction and the party choice numbers.
I understood that. My point is that all polling numbers need to be viewed with an eye to past voting patterns.
That people are satisfied with Rodney Macdonald is good for the Tories. That Darrel Dexter is a bit of a plodder on the campaign trail also benefits the Tories as it makes the likelihood of mercurial campaign swings less likely. Of course, if the Liberals catch fire during the campaign, they have the potential to grow at the expense of the other two parties, but I think that unlikely (so far, at least).
Polling done outside of a campaign period is always problematic (because those polled are fairly uncommitted to the answers they give pollsters). The media give far too much play to poll numbers in terms of expectations, but this provides ample opportunity to write analysis pieces after the fact. If Macdonald does win the coming election he will have "defied the polls", which is a good reputation to have. But realistically, it won't be that hard a thing to pull off.
Ed;
It's hard to disagree with your analysis however on a purely political perspective one has to tip their hats to the ruling Cons.
Your assessment of polling around the time that pollsters are gathering info is not new as you suggest, however the similarities of radio stations doing the same thing around “sweeps” are no different. In the environment of political spin and getting the message out, it is the prerogative of government to release such announcements when it will receive the biggest bang. The announcement of fur farms and marinas are not news or revelations to the proponents and communities, only the release of this info on the gov site is news to the general public. No big deal in the scheme of things! Besides, who said this administration can not use the smoke and mirrors available to them!!
To have the Comm Directors and the thumbs on the 10th and 8th do something different since 2003 would be the big news. Maybe the Libs could be as creative during the same period and ratchet up their message to the public. The lack of strategy and political maneuvering during “February sweeps” from the Opposition is both laughable and sad!
Anthony Roy
@herringchoker:
Generally, I'd agree that the actual performance in an election may be different and poll results must be assessed along with other factors.
I haven't looked as closely at the CRA Nova Scotia numbers as I have at the NL ones going back decades so I can't really judge NS politics that way. Some tells me though that your comment about NDP vote efficiency is spot on.
@ Anthony:
Four years ago, you wouldn't have seen very many people who even knew the poll goosing occurred. When I first broached the subject systematically, there were plenty of skeptics.
Now it's generally accepted to be the case.
Announcing at the time of polling is ordinarily quite difficult. Normally you wouldn't know when the pollsters are in the field or, if you are doing your own stuff, you'd lay off anything unusual to ensure you get a pretty reliable reading.
In this instance, CRA polls every quarter with the provincial government as a client. As a result government knows exactly when they are in the field and when they are finished.
It's the only pollster in town who releases figures publicly so the value of the figures is in helping to shape public opinion. If you can help shape the polls you get the real intended impact.
Geoff Meeker has written on this. The numbers are used casually by reporters and others all the time. It helps keep alive an illusion but someone else put it even more accurately: it isn't that Danny is popular. It's that people think he's right because he's popular. The whole thing can become very circular in perpetuating the cycles and in another way it helps to minimise the political discussion and debate.
In a broader sense, as I've noted before, you also have to wonder about the resources that poll goosing eats up within the system. It takes not only the political staff and politicians but it stretches out into the bureaucracy. No one has an infinite amount of time so if poll goosing is a quarterly requirement, then every quarter there is real work which isn't being done.
Your point about the opposition is well taken. Bear in mind though that this sort of stuff is largely a function of resources. When you starve the other guys or resources and the other guys are already fewer in number than before, you create a situation in which it is much much harder for them to keep pace let alone try and counter-goose.
It's not really laughable but it is sad, from the standpoint that limiting resources to one of the other teams is obviously part of the overall strategy.
"...this sort of stuff is largely a function of resources."
Sorry to disagree but the resources the Opp office has is adequate to "blitz" the wires. While the resource capacity has been decimated the resources to make announcements, points, counter points and general noise is not.
Heaven forbid that the Party assist with the resources in order to get a systematic and coordinated approach to general PR during the "sweeps". A two week blitz is all they need. A one day dog and pony show highlighting the obstacles associated with a decimated budget will counter a marina and fur farm announcement.
Heaven forbid that the Libs employ a counter attack during the same time period. Shit...the humble escriber has only been referencing the goosing notion for the last few years. Besides, the same strategy was used by Tobin and Heidi...!!
It's time for the Opp to stop dwelling on the fact they got the shaft and engage in old fashion and inexpensive politics. A newser with the Young/MUN Libs to an appearance at a pancake dinner would ceratinly sway at least one CRA point their way.
Now that's earth shattering if I say so myself :)
Anthony Roy
No argument from me.
Post a Comment