Voter turnout......

.....might well be at an alltime low this time around -- which usually bodes well for the emcumbent.

Something Wrong Here...

In 2005, the popular vote spread in favour of the Libs was 4% and they won 46 seats to 33 seats for the NDP... a 13 seat spread.

Yet, in Stewart's analysis of the ARS poll showing a wider 6% spread in favour of the Libs, his seat count is 43 Libs to 42 NDP... a one seat majority.

That's just not the way a 6% spread in favour of the Libs would end up on a seat count basis... it would be more like 50 Lib seats to 35 NDP at a minimum.

To Luke Skywalker

Hi Luke,

Thanks for the comments. I first started this analysis by looking at NDP vote/seat percentages and then Liberal/NDP spreads, but as they vary so much from election to election they do not lend themselves to accurate seat projections. But I noticed there is almost a perfect fit between Liberal vote share and seat share over the years.

The way to test how accurately this predicts elections outcomes is to use the regression equation available in my paper to reverse project seat shares - which it does +/- a few seats. I think it is the best method we currently have to estimate how this election is going, but I would be more than happy to look into any other suggestions you might have. Best wishes - Kennedy Stewart (kennedys@sfu.ca)

I Agree with Kennedy Stewart's Analysis

The NDP campaign to date has been lacklustre at best. They are not resonating with the voters.

The NDP are concentrating on very weak issues such as a few gaffes by Liberal candidates, and have said nothing about major issues such as the Convention Center overruns or the Liberal coverup of the Queen of the North sinking.

They have retreated from mentioning the B.C. Rail affair after the threat of a SLAPP suit by Patrick Kinsella's lawyer. Obviously the NDP hit a nerve with Kinsella and should be pursuing the issue further.

And what about the Cambie Street Merchants, who were shafted by Liberal Goverment decisions, while the Tsawassen power line whiners had their houses bought by the Liberal Government.

The best we can hope for is that the NDP doesn't get trounced and there is a leadership convention soon after the election.

Memo to Kennedy Stewart

I have been amazed in recent years by the ability of some parties to estimate the vote in target ridings, knowing where exactly to place "star" candidates for example. Do you have any insight into what kind of polling and projection methods parties use to accomplish this feat?

I am wondering to what degree very well organized parties are able to turn their targetted seats into self-fulfilling prophesies by the use of voter fraud. An American academic paper recently suggested that election night voter fraud in the US and Canada was much more widespread that most other experts would allow, and focussed its attention on districts with extremely small margins.

@Kennedy Stewart

And if the NDP are ahead in the polls or dead even?
Can you explain the last Angus reid poll that has the NDP with a 20 point lead on Vancouver island?

I make you a challenge Kennedy Stewart,have your students or yourself do your own poll.
Your study,your logic is flawed,the reason?...........
Mustel poll has been manipulated by a west Vancouver BC Liberal MLA that has business connections to them........

Ipsos reid polls,at least the polling in BC is skewed,the reason,Canwest Global has been pimping for Gordon Campbell and ever since 2005 when Canwest Global partnered up with Ipsos reid their polls(LOL,bunk)

Lets be perfectly clear.....Angus reid~s last 3 polls were......

NDP leading by 3 points--NDP leading by 5 points--BC Liberals leading by 6 points..

Those are real poll numbers,you should get out more often and talk to "real people" and you will find out real quick who is winning the election.
Campbell is running a polling ponzi scheme,Mustel/Ipsos will have some explaining to do on May 13th.

Here is another poll to consider.

http://www.robbinssceresearch.com/polls/poll_582.html

http://www.ipsos-na.com/news/pressrelease.cfm?id=2838

One merely has to look at all the Civic elections,the bi-elections,consider the bi-elections,......
Campbell had a province wide prime time NEWS/Press conference 15 minutes uninterupted just before the election,Campbell had spent 50 million dollars in a non-stop ad campaign from june through to the bi-election...

Campbell had 2 "STAR" cadidates running against 2 young rookie NDPers,,,,and,drum roll please.....The NDP rookies
Blasted Mcdiramid and put Griffiths into permanent retirement.

You want to predict?NDP-34 seats(now)-Where will they gain?

NDP will pick up 3 seats in Burnaby- 3 seats on Vancouver island--2 seats in Surrey,Fraserview-1 seat-south Delta Independent-NDP will pick up 1 in Maple ridge-..I expect the NDP to win half of the new ridings,Bennett will lose,NDP will win Prince George,and maybe Kamloops...

That puts the NDP up at 50 seats,that leaves room for potential losses of incumbents,but I don`t see the NDP losing any incumbent seats.

Finally,your forgetting who has been gored by Campbell?
Ferry users,Hydro user,Transit users,Seniors,Disabled,Children,The public sector,
Issues?-Carbon(bull)tax-ROR-Minimum wage-Fish farms-education-Bill 30--Massive pay raises-Autistic-Diabetic children-Poverty-Homelessness-CORRUPTION-CORRUPTION-Did I mention COPRRUPTION?

So MR. Stewart,Maybe you should leave your elitist little bubble,burn your thesis,and start over again.

MY prediction 50 NDP seats

Liberals 34 seats

1 independent

Cheers-Eyes Wide Open

To Rod Smelser

"...Do you have any insight into what kind of polling and projection methods parties use to accomplish this feat?"

"I am wondering to what degree very well organized parties are able to turn their targetted seats into self-fulfilling prophesies by the use of voter fraud."

Dear Rod,

Parties usually target ridings using past election results as it is too expensive to conduct reliable opinion polls on a constituency-by constituency-basis prior to an election. They usually only abandon constituencies in which they usually lose by more than 10-15 per cent of the vote. I don’t really see this as self-fulfilling prophecy as all parties are very reluctant to write-off any but the most difficult to win seats. If you think about it, as the Liberals won all but two seats in 2001 they might consider most seats in the province as winnable.

In terms of electoral fraud, aside from an occasional case involving very few ballots, this never happens in Canada. Rest assured Elections BC is held by most who study politics as one of the top election oversight organisations in the world. However, your point about double-checking results is valid. That is why we need exit polls in all provincial and federal elections. It always puzzles me why we don’t have exit polls in Canada as they are conducted on a regular basis in almost all countries including large, sparsely populated countries such as Russia and Afghanistan.

Best wishes - Kennedy

7th active model

Stewart Kennedy becomes the 7th practitioner to be included in our BC Seat Projection Avg.

Today, the Liberal Party lead is 61-24. Our long term momentum overlay projects a 64-21 victory by May 12th.

A tracking chart since the 2005 Election can be viewed at our website: http://www.Trendlines.ca/electbc.htm (along with the Federal riding projections).

sorry, that's Kennedy Stewart!

oops...

If the Liberals win the next election.....

....and by all accounts they will, watch for a complete disintegration of law and order in BC.

It is no surprise that Liberal MLA's think themselves above the law and with a rather dimwitted A.G. this will continue. The bottom-feeders of society, gangsters and drug dealers now think they are above the law as well and recent events show that this will happen.

The RCMP are now held in high odor and the public have been left in the lurch for a credible opposition as James and the rest of the NDP have catered to internal politics and not the public at large.

Campbell will win and he will rape this province and with absolutely no respect for the law, except the law of power, the public will begin to condone violent acts against the government.

This will insure the right wing-nuts to scare the average voter in supporting the Liberals or their replacements, leading to more and violent tactics by the disenfranchised.

By 2016 the province will be bankrupt and violence (shootings etc.) will be a daily occurrences as the drug lords will be the new 'law and order'.

This will be the Gordon Campbell legacy, one law for the rich, the powerful, and the ruling elites and other laws, mostly unaffordable, for the poor and working classes.

Turnout is key.

Elections B.C. is doing a good job with mail and t.v. campaigns informing voters of the election. This province also makes it easy to register and vote on election day; important since ridings have been altered.
Will this be enough to keep the turnout in the 60-63% of recent elections? The scant
coverage in larger media and low key efforts by all parties so far makes lower participation likely. The canucks success in the playoffs is a huge factor too.
A 5% drop in total voters would mean 1,000 fewer ballots in each riding. The big story of this election is which party has the motivated supporters.

Mr. Kennedy said....

"In terms of electoral fraud, aside from an occasional case involving very few ballots, this never happens in Canada..."

Which is fair enough.

However, especially given the methodological underpinnings of his data analysis, perhaps he would like to comment on the potential influence of skewed polling, an issue that was raised by another commenter above.

A specific case to consider involving the 2005 British Columbia election might be this one.

(which is a case that Mr. Paulsen also might wish to weigh in on given that he wrote the piece linked to above)

Thanks very much.

.

When all the guessing fails,

When all the guessing fails, we check the election poll booths, and then the experts can tell us they were correct no matter what the results.

I would hope that ndp supporters not pay too

much attention to mr stewart's poll.

I am finding the same pattern of inaccuracy as occurred last election is reemerging. Most pollsters had the ndp signifigantly behind in the 2005 election, yet the end result was nearly a statistical tie. Had about 5000 votes over 8 ridings gone differently we have had an ndp goverment. Had more ndp supporters, including many writers on the tyee, been more skeptical of polls, that alone could have insured an ndp victory. The basic function of polls as used today, is to discourage progressive voter turnout, that's why they're so heavily financed by rightwingers with deep pockets. If you doubt this then consider how many polls we have had in the last what? FOUR? federal elections "showing, that this time," oh yes, surely this time stephen harper is in majority government territory. These polls always seem to occur at suspicious points in the campaign, and never seem to predict upsets. If a poll discourages even 1% of an opposition party's supporters turning up at the polls on election day, they can in THAT way win elections.

I am struck by the many amazing similarities between this provincial election and the last:quasi-ndp supporters who spend half their time bashing james, media tricks, and equivocation around political debates (last year it was around the main televised debate, which it took 3 weeks for the mainstream media to admit james had won) and many pollsters who were sure that the ndp were at best going to pick up 20 seats.

But several other factors are beginning to work in James favor: 1.Voter registration is up 5%, and the ndp has the most committed, least likely to change it's mind support. 2.Van dongen, drunk driving, gay bashing, bad driving(laura mcdiarmid) have just cut at least 3 points from liberal support; van dongen has reminded people about campbell and hawaii, just when the liberals can least afford it, and the mainstream media has no choice but to stop harping so much on so-called ndp candidate weaknesses.3.The ndp are casting a much wider net this election, making signifigant criticisms about the economy and environment in mainstream media-tieleman's 24 hours last two columns. The liberal environmental green washing of a brown premier has been undermined by alexandra morton's ndp support. 4.The ndp has far more popular mlas -dix,simpson, Herbert, etc who are seen as more than just rubber stamps for the ndp, unlike the ever deteriorating reputation of bc liberal rubber stamp pimp mlas who are seen more and more as creatures of the premier.

So let's not let happen again what happened last time, and be misled by well meaning ndp supporters like kennedy, in thrall to the so-called magic of polls. Let's NOT let them do again what they did last time. And for james bashing ndp supporters, remember, you're voting for a whole party, with many excelent mlas, whereas bc liberal supporters are voting for king gordo and 84 rubber stamp, servile,would be mlas and cabinet ministers.

2005 Election v. 2009 Election...

The only two pollsters in BC with a decades long reliable track record are Mustel and Ipsos IMHO.

Mustel has always been "spot-on" in their final election polls compared to actual election results.

More importantly, the spread between the leading parties reveals the extent of a victory/defeat.

Mustel:

1. Feb. 13, 2005 - 6% spread in favour of the Libs;

http://esm.ubc.ca/BC05/polls.php

Feb.13, 2009 - 16% spread in favour of the Libs;

http://www.mustelgroup.com/pdf/20090213.pdf

2. April 11, 2005 - 8% spread in favour of the Libs;

http://esm.ubc.ca/BC05/polls.php

April 14, 2009 - 17% spread in favour of the Libs;

http://www.mustelgroup.com/pdf/20090414.pdf

Compared to this point in 2005, Mustel shows an additional ~10% spread in favour of the Libs.

3. May 9, 2005 [just before election], Mustel showed a 5% spread in favour of the Libs (election spread was 4%).

http://www.mustelgroup.com/accuracy.php

Bang-on.

One can also do the same exercise with the Ipsos polls both then and now.

Interestingly enough, Mustel did not show much change between party spreads during the 2005 election.

The party spread is the key component in terms of winning elections. The larger the spread, the more magnified the victory/loss will be for with either party in terms of seats under FPTP.

At this late stage in the game, even if the NDP becomes lucky and can reduce Mustel's spread from 17% to 10% by election day, the election outcome will still be disproportionate.

Bernard Schulmann, long-time political analyst who predicted the NDP's 2-seat 2001 election, has already concluded the outcome of a 10% spread in favour of the Libs:

Lib - 65 seats
NDP - 20 seats

http://bciconcoclast.blogspot.com/2009/03/playing-around-with-some-numbers-here.html

Caveat: Final election opinion polls will be determinate in final seat projections.

Forgot to Add...

Even Kennedy Stewart has the following seat projection under Mustel's current 17% spread:

Lib: 67 seats
NDP: 18 seats

http://www.kennedystewart.ca/BC%20Seat%20Projections%20-%202009.pdf

Falling Skies, too!

"....and by all accounts they will, watch for a complete disintegration of law and order in BC"

Yes, and the sky will fall, too, or continue to fall.

I would wager more like 55 Liberal, 35 NDP.

In response to Luke, Mustel has a pretty spot on record. They do the inside polls for both the NDP and Liberals, btw.

In your dreams, luke flystalker.

Mustel polls are a running joke, as they should be coming from a group, with as many insider ties to the bc liberals as the fraser institute, and the vancouver board of trade. In campbell's first term, mustel was notorious, for releasing feel-good bc liberal polls just BEFORE the next horrific cutback and dirty trick. Evi mustel considered running for the liberals in the 2005 election, but must of rethought it after considering the upcoming gravy train for years of best place on earth ads. I believe there's a current liberal cabinet minister with strong ties to mustel, as well.Mustel polls should therefore NEVER be used as election predictors, and it is beyond me why mustel even making polls does not constitute a massive conflict of interest. No, polls are for dogs, as one prime minister had it, dogs, and bc liberal hacks...and deluded progressives...show me the monitoring; there is none.

Frenchie...

As a partisan NDP hack, I understand where you are coming from.

Yet the Mustel facts seem to dictate otherwise:

1. Hospital Employees Union - Client for Mustel Poll;

http://www.heu.org/index.cfm?act=print&print_ID=5184

2. BC Federation of Labour - Client for Mustel Poll;

http://www.bcfed.ca/node/324

3. CUPE - Client for Mustel Poll:

http://cupe.ca/news/3685

4. BC Nurses Union - Client for Mustel Poll:

http://www.bcnu.org/whats_new_media/bulletins/2004/10-20.htm

5. BCTF - Client for Mustel Poll

https://bctf.ca/publications/NewsmagArticle.aspx?id=7786

I can go on and on... but you surely must now get the picture, n'est pas?

And the final word on Mustel from the left-wing Babble discussion forum:

http://www.rabble.ca/comment/980871/Unfortunately-Mustel%C2%A0

turn out

elections bc is bracing for a large turnout based on registration information so far...and when turn out increases, its usually to oust the governing party

Ipsos - NDP NEGATIVE Momentum...

BTW, on Friday, Ipsos released another public opinion poll and compared the Libs and NDP's momentum scores compared to their 2005 campaign findings (in parenthesis).

Lib: -22% (-19%)
NDP: -20% (+16%)

In other words the Libs momentum score has decreased by -3% compared to this point in the 2005 campaign, but the NDP’s momentum score has decreased by a whopping -36% compared to 2005.

http://www.ipsos-na.com/news/pressrelease.cfm?id=4358

And that certainly will have an impact upon voter turnout.

No matter how one slices and dices election 2009, it's gonna be a toughy for the NDP.

Mr flystalker:

I guess voters were just so impressed by mr van dongen and the other spot and stain ridden record of prospective bc liberal mlas, that they're now flocking to the bc neoliberals...perhaps somewhere in a galaxy far, far away but NOT in this election, flystalker...sure hope the premier keeps up the IDIOTIC fireside chat ads, I'm sure the eight year old demographic is incredibly impressed...

Frenchie...

Well, since ya love mud-slinging so much (unlike the BC electorate), ya forgot to mention the NDP's candidates... Ray Lam... and others:

http://www.cbc.ca/canada/bcvotes2009/story/2009/04/24/bc-election-candidate-conduct.html

Frankly, it appears that nobody even cares unless one is a die-hard party hack. ;)

As for van Dongen, CKNW had this 'unscientific" poll:

http://www.cknw.com/Polls/PopUpResults.aspx?QID=11569

Before you discount that "unscientific" poll, remember that CKNW conducted a poll on the previous day confirming that Carole James was superior to Campbell regarding the environment.

In any event, mainstream BC'ers are looking at the economy and who is most competent in government in that regard.

That's what it all comes down to.

Not partisan party hack mud-slinging. People get turned off by that stuff.