Premier Alison Redford with her winning PC machine. Below: Opposition Leader Danielle Smith with her less successful Wildrose machine. Below Ms. Smith: Alberta Scan publisher Paul McLoughlin and pollster Janet Brown.
Not only are the Progressive Conservatives the most popular political party in Alberta, but Premier Alison Redford enjoys a commanding lead in public confidence among the province’s political leaders, the province’s latest poll says.
Premier Redford’s leadership has the approval of 62 per cent of Albertans – almost two thirds of the province’s voters – according to the survey by Trend Research of Edmonton, which was published yesterday.
This is the second poll in as many days that shows Ms. Redford’s PCs far ahead of all opposition parties in public esteem. But this one adds the information that the premier herself more popular than her own party. Both polls were in the field at the same time.
By comparison to Ms. Redford’s approval rating, Wildrose Leader Danielle Smith, Leader of the Opposition in the provincial Legislature, has the approval of only 42 per cent of Albertans, says the survey, which was conducted for well-known pollster Janet Brown and political newsletter publisher Paul McLoughlin.
That their leader has the approval of fewer than half of the province’s voters while the premier has the confidence of more than two thirds in a methodologically sound poll taken while the government party was in the midst of a swirling controversy over public health care executives’ expense accounts cannot be called good news for the market fundamentalist Wildrose Party.
The Trend poll – the results of which were published in Alberta Scan, Mr. McLoughlin’s subscription-only newsletter – also shows approval among Albertans province-wide for Alberta NDP Leader Brian Mason at 33 per cent, and for Alberta Liberal Leader Raj Sherman at 32 per cent. Alberta Scan is not available online.
On the relative popularity among committed voters of Alberta’s political parties, the Trend poll’s results were very similar to those of the survey by Environics Research Group that was published on Thursday and reported on this blog yesterday.
Like the new information about leader popularity, this seeming confirmation of Environics’ results, which could not be called particularly good news for the Wildrose Party or Ms. Smith, adds up to more grief for Alberta’s market fundamentalists and the province’s rural rump of social conservative zealots.
The Trend poll shows support among committed voters for Ms. Redford’s PCs at 49 per cent – an increase from 44 per cent in the April 23 provincial election. By comparison, the survey placed support for the Wildrose Party at 27 per cent, down seven points from 34 per cent on election day. Trend had the Alberta Liberals at 12 per cent and the NDP at 10 per cent province-wide. Nineteen per cent indicated they were undecided.
By comparison, Environics’ numbers showed the Tories at 43 per cent, Wildrose at 26, Liberals at 14, NDP at 13 and undecided at 13.
The trend poll was conducted between Aug. 10 and Aug. 20. Environics was in the field between Aug. 10 and Aug. 22. Because the dates are so similar, concluding that the Trend poll is more bad news for the Wildrose Party carries the same caveat as yesterday’s analysis of the Environics numbers.
To wit: Voters who may not have blamed Premier Redford for the expense account scandal at Capital Health, which happened long before Alberta Health Services was created let alone before her watch began, may still be upset at other uproars that have made news since. These include the closing of a planned police college and a viable nursing home for dementia patients in a Wildrose riding, plus a fiscal update that indicates the province will post a bigger deficit in 2012 that previously forecast.
Then again, they may not. Voters who are disinclined to vote for more progressive parties may nevertheless weigh Ms. Redford’s sins, real and imagined, against the radical market fundamentalism and extreme social conservatism of the Wildrose Party and decide to overlook a year or two more of deficit financing and a failure to pander to rural ridings that vote for the opposition.
The Trend poll of 900 Albertans is considered accurate with a margin of error plus or minus 3.3 per cent 19 times out of 20.
Some readers will naturally wonder, as did political analyst David Heyman in a comment on Facebook about yesterday’s post, if “polling in Alberta was rendered irrelevant after the last election, surely”?
Perhaps, but I tend to think that the spectacular failure of Alberta’s pre-election polling was driven by several factors, not all of which can be used to discredit the work of responsible pollsters like Ms. Brown and Environics.
First, of course, the electorate was volatile and voting decisions were affected by the famous bozo eruptions of some social conservative Wildrose candidates. So polls that may have been accurate on April 19 obviously weren’t accurate any more by April 23.
Second, however, was the addition of several poor quality polls to the mix – some of which may in fact have been deceptive push polls designed to produce results that would boost the Wildrose Party’s chances.
Finally, it is said here, the work of poll aggregators who lumped the results of several polls together created a misleading picture when some of the poorer quality polls were included.
Whatever happened, it illustrates that while polls remain valuable tools that are worth reporting, they need to be taken with the appropriate grain of salt.
This post also appears on Rabble.ca.
That appropriate grain of salt needs a wheelbarrow to be moved. The very methodology of polls is suspect, given that land line telephones are going away at an astonishingly rapid pace. CBC was saying they’ll be essentially gone in 5 years, which I’m not sure I believe. Sure, the pollsters say they can compensate using other sources of data, but it’s just more stirring of increasingly suspect data. It’s suspect because the pollsters don’t seem to grasp that more and more people either don’t respond to poll questions, or will lie. I encourage both these trends.
Right now they tell us they polled x many people, and it’s accurate blah blah blah. I’d also like to know how many phone numbers they dialed to get those x many people. And as you know, poll questioning is notorious for producing the results that the poll owner wants, or that they’ve been paid to produce. That why I think think polls are a waste of time to produce, to gather and analyze the data, to publish the results, and to carry out any action based on those results.
Who cares that one party or the other is doing well or poorly in the polls now? For better or worse we’ll have Redford running things for 3 or 4 more years, or until she deigns to hold an election because the timing is right.
So what’s new, 2/3 of voters should donate their brains to science to figure out how it can function on 2 living brain cells.
I think an important point is being ignored here. People are starting to play the polls games very well. We all know how politicians and pollsters interpret and use polls for many different reasons and one of them being brain washing. Many people are telling the polsters what they want to hear. People are not stupid and they are adapting to our exagerated reliance on polls. People are definitely adjusting to the control of public opinion in a very smart way. Soon these polls have no meaning whatsoever.
I disagree with the thrust of the comments here. What we have seen in recent years that while it is true that landline use is declining, and therefore it can be harder for traditional polls to garner the 1,125 or so responders they need to get that 2.6% margin of error 19 times out of 20 (which means a 95% confidence interval), the traditional methodology of polling remains more accurate than newer methods such as online panels. Reasons for this likely have little do with the mathematics of statistics and probability, and more to do with political psychology and behaviour. For example, landline telephones are rarer among young adults than middle-aged or older ones; however, voter turnouts among young adults are also lower, so what that means is that while young adults are under-represented in traditional polling, they are also less likely to vote, a fact which negates the impact of their omission from the outcome of the election the poll is trying to predict. On the other hand, online panels tend by their very nature to include the more politically engaged, who may not necessarily reflect the political viewpoints of the electorate at large; this is the most plausible explanation for the failure of the Alberta Party to show significantly in the April 23rd election.
The bottom line, then, is that while traditional polling methodology is imperfect, and growing more imperfect every year, it remains much less imperfect than newer polling techniques that have emerged.
The recent polls in the Quebec election were pretty much bang on. The bogus polls put out by Sun Media in the Alberta election have given the industry a bad name.
Not true. Online surveys and “robopolls” have consistently shown themselves to be roughly as accurate at predicting election results as traditional telephone interviews. For example, the Angus Reid online poll has been the closest to the actual results for the last two federal elections in a row. Or look at the Alberta election, where all the different poll types were about equally bad.
Also, all types of surveys will attempt to account for low responses among a particular demographic by weighing the people who *do* respond more heavily. So just because few young people answer a particular phone interview poll, it doesn’t mean young people will be underrepresented in the results.
The claim that polls can’t be trusted because people are deliberately giving false answers is just silly. No one but the most hardcore cynics & partisans do that – everyone else who doesn’t like to respond just hangs up.
Great news for Allison.
Now I wonder if she’s going use her new clout to release that 360-page report on Alberta’s electricity market.
The energy minister says he needs to review it first before releasing the information to the public.
My guess it may be undergoing some substantial revisioning where words like “gouged” (as in price gouging) are being substituted for phrasing like “market dynamics” (as in electricity consumers in Alberta have entered an exciting new era of market dynamism).
“Second, however, was the addition of several poor quality polls to the mix – some of which may in fact have been deceptive push polls designed to produce results that would boost the Wildrose Party’s chances.”
Hey David, since I keep hearing this particular conspiracy theory, let’s stop beating around the bush and identify just which polling companies you think were engaged in publishing fake results to help Wildrose win. Below are Wildrose support levels from the latest poll from each firm as of Saturday, April 21. Please identify which outlier(s) were attempting to deceive us.
Forum Research(*) – 41% (robocall)
Angus Reid – 41% (online)
Campaign Research – 41% (robocall)
Abacus Data – 41% (robocall)
ThinkHQ – 41% (online)
Leger Marketing – 42% (phone interview)
Return on Insight – 43% (phone interview)
(*) Note that Forum did one additional poll on Sunday, the day before the vote, which had Wildrose dropping to 38% and their lead over the PCs dropping from 9 points to 2 points. Of course, this is further evidence that the discrepancy between the polls and the actual results was due to a last-minute support switch by a volatile electorate, not bad polling or a pink flower conspiracy.