Nova Scotia: how did the polls and the projection do?

After the trauma of the elections in Alberta and British Columbia, it was almost more surprising to see last night's results in Nova Scotia generally align with what the polls were suggesting they would be.

Some pollsters did better than others. My detailed analysis of how the polls did can be found on The Globe and Mail's website. The article is available to Globe Unlimited subscribers only. If you like this kind of analysis and want to see more of it on The Globe and Mail as well as here on ThreeHundredEight.com, it needs your support. If you don't already have an online subscription to The Globe and Mail, please consider signing up as you might find my articles available to subscribers-only more often. You can find details on how to sign-up via the link above.

The projection did moderately well, particular in terms of the Liberals. Their vote haul of 45.5% fell just outside the likely ranges, which bottomed out at 45.8%. But there was a 21% chance that their vote would fall between the minimum and low projected ranges. Their total of 33 seats was projected exactly right.

At 26.9%, the New Democrats were just 0.8 points off from their projected vote totals. Falling between the average and high ranges, as they did, was considered a 68% chance probability. That they only won seven seats instead of the projected range of 12 to 15 was a surprise, but this was largely due to the performance of the Progressive Conservatives. Those seven seats they won did fall between the minimum and low range, though that was projected to be only a 5% likely occurrence.

The Tories ended up just outside the projected maximum level of support with 26.4% instead of 26.3%, and 11 seats instead of nine. That is unfortunate, to say the least. But this is partly due to the large percentage of the projection taken up by the final poll from Forum Research. More on why that was a problem later.

The Greens took 0.9% of the vote, just below the projected low range. Falling between the minimum and low level of support was considered a 35% chance, so nothing untoward there.

But one of the main reasons why the projection did miss on the seats for the Progressive Conservatives was that the polls did not do a particularly good job of gauging regional support, especially in Cape Breton. This is how it broke down last night:

The polling in Halifax and in the rest of the mainland was generally good, though no one got it bang-on. Cape Breton was a little more difficult. The full sample from Abacus, spanning the entire final week, was relatively close but the last set of numbers from them had the PCs doing much better, while the last set of numbers from Forum and CRA had them doing much worse. The samples were generally too small in Cape Breton, though, to get a good bead on the race.

The model did have some trouble translating the regional numbers into good seat totals. With the actual regional numbers plugged into the model, it would have projected a likely outcome of 22-33 seats for the Liberals (28 to be exact), 13-17 seats for the NDP (13), and 7-16 seats for the Progressive Conservatives (10). The Liberals ended up at the high end of that range and the PCs right in the middle of it, but again the New Democrats would have been over-estimated, falling in the minimum-to-low range. This is actually the first time in ThreeHundredEight's history of projecting a dozen elections that the model performed worse with the actual results plugged into it,

This demonstrates two things. The first is the limitations of a seat projection model in smaller provinces - on average, only about 8,100 votes were cast in any riding. That means that local factors can be especially important, and that is shown by the direction in which the model made errors: every which way. Normally, the model makes errors in the same direction as a party is over- or under-estimated. In Nova Scotia, however, that was not the case. For example, in the eight ridings the Liberals won where they were not projected to win, the NDP was projected to take six of them and the Tories two. In the seven ridings the PCs won where they were not projected to win, the Liberals were projected to win five of them and the NDP two.

Secondly, it highlights just how poorly the New Democrats did. They should have won more seats with the amount of votes they captured, particularly in Halifax. They ended up losing too many races that should have been winnable, with the Liberals benefiting.

In an election where just over half of those ridings that were not entirely new changed hands, the projection model was not particularly strong at the riding level. It made 18 errors, calling 33 of 51 ridings correctly for an accuracy rating of 65%. Taking into account the projected likely ranges, the likely winner was identified in 37 of 51 ridings, for an accuracy rating of 73%. That is not very good, but shows the hazards of small elections. The idea that the overall numbers are more important, and that riding-level errors balance each other out, is especially true in smaller elections.

The probability ratings for the riding projections did a decent job, however, with an average confidence of 72% in the ridings incorrectly called compared to 81% in the ridings correctly called. Half of the ridings incorrectly called had a confidence level of 70% or less.

One of the problems with this campaign was the emergence of the Forum Research poll on the eve of the election. Because the model weighs a poll by its median date, it rewards a firm like Forum that does its polling on a single day. This is actually not really the very best practice since it is better to poll over a few days to iron out the data and ensure that it isn't too dependent on the factors that can skew a poll when it is taken on a single day (i.e., are the people available to take a call on a Monday night different from those who can take a call on a Sunday night?). Coupled with Forum's large samples (due to the cost effectiveness of IVR polling), the final Forum poll took up almost three-quarters of the projection.

I don't think it is appropriate to reward this kind of poll, so going forward I will be weighing polls by the final date in the field, particularly during an election campaign. Doing so would have made CRA's poll weighted for Oct. 3 instead of Oct. 2 and Abacus's weighted for Oct. 6 instead of Oct. 5. At the very least, this would have brought the Progressive Conservative result into the maximum range and undoubtedly produced a better projection.

It is gratifying, however, that the polls did a good job in Nova Scotia. That makes the coverage this site provided throughout the campaign useful, since it was an accurate reflection of the ups and downs of the last four weeks. The lessons learned from this campaign will be digested and the new data added to the model's calibrations, and on we go to the next vote, slightly more confident than we were yesterday that we can trust what the polls are telling us.

Final Nova Scotia projection: Liberal victory

Stephen McNeil's Liberals should win the provincial election in Nova Scotia tonight, relegating Darrell Dexter's New Democrats to Official Opposition status. Jamie Baillie's Progressive Conservatives should remain as the third party in Nova Scotia's Legislative Assembly.

ThreeHundredEight's final projection for the Nova Scotia provincial election heavily favours the Liberals to win a majority government, though the slimmest of minorities is also envisioned as a possibility. No scenario puts the NDP ahead of the Liberals in the seat count, though it is possible that the New Democrats could have a very bad night and find themselves in third place behind Baillie's Tories. But the expectation is for the Liberals and New Democrats to swap places in the assembly, and for their 2009 seat counts to be virtually reversed.

The likely outcome

Final projection (click to magnify)

The Liberals are projected to win between 30 and 38 seats in tonight's vote, putting them well above the majority threshold of 26 seats. The model gives them 33 seats, which would represent their best electoral performance in the province since 1993. The are at an estimated 47.7% support in the polls, with the likely vote range stretching from between 45.8% and 52.5% of the vote.

Finishing second should be the New Democrats, with a projected likely range of between 12 and 15 seats. The model awards them 12 seats exactly, their worst result since 1999. They are expected to take between 26.1% and 28.7% of the vote.

Ranges

Where the Progressive Conservatives will finish is an interesting question. The projection model has had them as low as 15.9% at the mid-point of the campaign, but they appear to have experienced an uptick in the last few days. They are projected to win between four and seven seats, with six being considered the most likely outcome. They should take between 21.4% and 24.7% of the vote, and averaged 23.3% in the polls. If this occurs, this will be the lowest number of seats they have won since the Second World War and their lowest share of the vote since Confederation.

The Greens did not nominate a full slate of candidates, and so are not projected to do very well. They are expected to take between 1.2% and 1.7% of the vote and win no seats. Independents are expected to take about 0.4% of the vote.

How it could all go wrong

As I have discussed on several occasions, the fundamentals in the polling being conducted in Nova Scotia are extraordinarily strong for the Liberals. In addition to a lead wider than anything seen in the late stages of the campaigns in British Columbia and Alberta, McNeil has been out-polling Dexter by a significant degree at a personal level and the Liberals generally lead on the issues that matter.

In fact, some of the lessons of B.C. seem to have been learned. The polls that were done by the Corporate Research Associates and, especially, Abacus Data looked into the race in great detail and went beyond the simple voting intentions question. Because of that, we were able to see that the Liberal lead was not only wide but deep - something that neither the B.C. New Democrats nor the Alberta Wildrose could say before they were disappointed on election night.

Nevertheless, the model is prepared for a surprise tonight and observers of the election should be as well. The model assumes a degree of error as bad as anything we have seen in the last few elections, making for some wide ranges.

Based on those worst and best case scenarios, the Liberals could theoretically take anything between 43.9% and 64.4% of the vote and win between 25 and 43 seats. The New Democrats could win between 25.1% and 34.5% of the vote and between four and 23 seats, while the Progressive Conservatives could win between 19.5% and 26.3% of the vote and between two and nine seats.

As you can see, the Liberals and New Democrats do not overlap either in votes or seats. This eliminates an NDP victory as a theoretical possibility, though it is always possible that the polls in Nova Scotia will do even worse than they did in B.C. and Alberta. If that is the case, woe be upon them.

The NDP and PCs do overlap, however, and it is possible that the Tories could finish second in both seats and votes. The final Sun News Network/Abacus Data poll suggested that could be the case. Vote efficiency is certainly in the NDP's favour, however, and they should be able to finish second.

Regional breakdowns

Because I wasn't sure what kind of detail would be available during the campaign - and until Abacus started polling, that detail was rather pitiful - I did not break down the projection by region. But the model was and is making regional projections based on the CRA data that was available (estimating it when it wasn't) and incorporating the Abacus and Forum Research numbers when they emerged.

Riding projections

Halifax is the major prize of the campaign, with the Halifax Regional Municipality containing almost two out of every five seats in the province. The polls did suggest that this was one of the closest races in the province, but the Liberals appear to have moved ahead in the final days. They started the campaign at over 50% in the HRM and ended it closer to 45%, but for a time the gap between the Liberals and New Democrats in the capital was a mere two or three points.

The model estimates the Liberals will take between 44% and 50% of the vote in Halifax, netting them between 12 and 14 seats. The NDP should win between 30% and 33% of the vote and between four and seven seats, while the Progressive Conservatives should take between 20% and 23% and one or two seats.

The rest of the mainland region was dominated by the Liberals in most polls, with the party over 50% throughout the campaign and the two other parties splitting the rest between them. How the Liberal steamroller will perform here, however, remains to be seen. Will NDP and PC incumbents be able to hold?

The Liberals should take between 45% and 52% of the vote here and between 11 and 17 seats, making it the region with the widest range for the party. The Tories are expected to finish second with between 24% and 28% of the vote and three to five seats, while the New Democrats should take between 23% and 26% of the vote and between five and seven seats.

Cape Breton is, to me, the biggest mystery of the election. The polls from CRA and Forum have both shown very weak numbers for the Tories with support in the mid-to-low teens, while Abacus has given them about 40% support. That makes a huge difference for the four PC incumbents on the island.

The weight of the evidence, primarily due to Forum's large sample on Cape Breton, points to the party finishing third with between 17% and 19% of the vote and zero seats. I have to admit that I believe the model will be wrong here and that some of the incumbents will survive. But with between 49% and 56% of the vote in Cape Breton, the model gives the Liberals seven of the island's eight seats. The NDP is expected to take between 27% and 30% of the vote and win one seat.

Polling consistency

The polls have been very steady in Nova Scotia, showing hardly any movement since the campaign began. Campaigns always matter, but sometimes they serve to just confirm what people were already thinking.

In the last week of the campaign, the Liberals were assessed to be at between 46% and 48% in almost every poll. That is a remarkable degree of consistency considering the generally smaller samples. Where the Tories and New Democrats stand, however, is a different matter.

CRA was suggesting that the PCs were polling very badly at under 20% until their final poll of the campaign was broken down. The last three days of their polling pointed to PC gains. Abacus showed the same thing, with the party polling at 22% to 25% in their first missives but 28% in their final one. Forum pegged the Tories at 23%, splitting the difference.

And that difference is what puts the NDP in a solid second or a precarious third. CRA put the New Democrats at 31% in their final poll, while Forum put them at 26% and Abacus at 24%. Unless the polls are wildly wrong, the winner of the polling sweepstakes will be the polling firm that put the Tories and NDP in the right spot. Generally, though, you have to favour the government - my gut tells me CRA will be closest to the mark.

How the leaders have fared

The last time anyone other than McNeil was chosen by Nova Scotians as the best person to be premier in a poll was in February 2012, and his numbers only got stronger during the campaign. CRA gave him between 34% and 38% on the best premier question in their final polls, up 11 to 20 points over Dexter. He barely polled above Baillie, something that was consistent in Abacus's numbers as well. In their polling, though, McNeil was lower (just above 30%) and the 'don't knows' were higher.

Abacus gave us a bit more information as it asked Nova Scotians to rate their opinions of the leaders. This was also good for McNeil, who managed 51% to 54% positive scores in all of Abacus's polling during the final part of the campaign. His negative rating was increasing, however, from 22% to 27% in the final days, but that still gave him a remarkable +25 score by the end. Dexter's positive score was stuck at around 37% with a negative rating around 47%, while Baillie's numbers improved slightly in the final days. He finished with a 41% positive to 33% negative score.

The projection model considers both Baillie's and Dexter's re-election chances to be good but by no means assured. Dexter is projected to take between 41% and 47% of the vote in Cole Harbour-Portland Valley, against 36% to 42% for his Liberal challenger. Baillie is slated to take between 44% and 50% in Cumberland South, against 41% to 47% for the Liberal candidate. Both calls are made with 59% confidence. I suspect, though, that they will probably be re-elected with larger margins than the model suggests (big swings can be difficult to assess accurately).

McNeil's re-election in Annapolis is made with 100% confidence, though the steam is shooting out of the model's seams on this one as it gives him between 81% and 93% of the vote. It was already a good riding for the party, and with the Liberals making such huge gains in the region their support is probably being inflated here.

A big night for Nova Scotians and pollsters

Nova Scotians generally give their governments a second kick at the can, the last time a government being booted out after one term being in the 19th century. But we saw in New Brunswick that re-election streaks can be broken. The election of an NDP government in Atlantic Canada was a historic first in 2009, so the defeat of a first term government in Nova Scotia can be a historic second in 2013.

That is, if the polls are right. Everything points to the polls being right. The Liberal lead is among every demographic, among likely voters, among undecided voters. The polls should not be wrong. If they are, the industry will be dealt a blow of tremendous proportions. Both CRA and Abacus were doing their polling the old-fashioned way: daily tracking with live-caller interviews. Both CRA and Abacus investigated how likely respondents were to vote, and Abacus probed their motivations to a great degree. Forum swooped in with a larger sample, more superficial poll, but simply confirmed the findings of their two colleagues.

Nova Scotia is a small province, but a lot of eyes will be on the results tonight. Canadian New Democrats hope that Dexter will pull off a surprise, and at least win a moral victory. Canadian Liberals hope that McNeil's election will mark the first salvo of a Liberal return to prominence. And Canadian pollsters hope that a strong call will be made and the industry's reputation will be partly rehabilitated. It should be an interesting night.

Final polls from Forum and Abacus show Liberals on track for big win

The final polls of the Nova Scotia campaign were released today by Abacus Data and Forum Research, the two showing the Liberals enjoying a lead of between 18 and 22 points. There seems to be little doubt that the Liberals are on track for victory, but who will finish second in the vote count is up for debate.

We have not heard from Forum during this campaign, but the firm often puts out last minute polls. And unlike their election eve polls in Alberta and Quebec, this one aligns with what the other polling firms are showing and does not suggest we are in for a surprise.

Forum gives the Liberals 48% support, compared to 26% for the New Democrats, 23% for the Progressive Conservatives, and 3% for the Greens. It is virtually identical to most of the numbers we have seen from CRA and Abacus Data during this campaign.

The PC's strength in the north end, however, is different from what Abacus has found while the Tories' weakness in Cape Breton is similar to CRA's findings and in contrast to Abacus's. But with the PCs having four incumbents in Cape Breton, it seems strange that the party would be doing so badly there. And those 5% of Nova Scotians hoping to vote Green on Cape Breton will be disappointed - the party does not have a single candidate there!

But overall, Forum suggests the Liberals should win handily. They lead among all age groups, income brackets, and education levels, along with among both men and women. And the firm has been doing some daily tracking since October 4, according to its report, showing remarkable consistency for all parties.

The full details of the Forum poll, which called 28,868 households in order to build its sample (a response rate of 3.3%), can be found on the polling firm's website. The PDF to the report is here.

The final Sun News Network/Abacus Data poll of the campaign shows a change in order of the parties, but no statistically significant shifts. And Forum and Abacus are within each other's margin of error.

Abacus has the Liberals still way ahead with 46% of the vote, compared to 28% for the Progressive Conservatives and 24% for the New Democrats.

Tracking polling by Abacus and the Corporate Research Associates throughout the campaign have had the Tories in third, so to have them in second place is quite a change of fortunes. But the margin of error being what it is, the Tories's edge over the NDP is not statistically significant.

Among likely voters, the Liberals led with 46% to 27% for the PCs and 26% for the NDP, suggesting that the advantage the Tories had in the overall vote over the New Democrats may diminish when people actually vote.

The Tories continue to poll well in Cape Breton, while a step backwards in Halifax by the New Democrats is the main reason for the order change.

The polling firm contacted 7,326 households to build its sample, a response rate of 8.2%. Sadly, this is typical of polling today.

Some of the findings of the poll that I found concerning earlier today have partly gone away. While the Liberals are now at a bigger disadvantage in Cape Breton, their 12-point lead in Halifax should deliver them enough seats to form a majority government. On who would make the best premier, McNeil is seven points up on Dexter in the capital. And for those who pegged the economy as the most important issue, the Liberals now have a three-point lead over the NDP (though they also now have a three-point deficit on the issue of healthcare). When everyone was asked about the economy, though, the Liberals increased their lead over the NDP - they are now ahead by 32% to 18%.

In addition, the Liberals' fundamentals remain strong: a lead among all demographics and very strong personal numbers for McNeil, particularly when compared to Dexter.

Now that the polls are in it is up to those Nova Scotians who did not yet cast their ballots to do so. This is an important election for them, of course, but it is also an important election for pollsters. They need a good call. There is enough agreement between the polls from the Corporate Research Associates, Abacus Data, and Forum Research that they will all sink or swim together. But there are also enough differences that some of them will come out of tomorrow night's vote looking better than the others.

Close races in Halifax and Cape Breton

Over the weekend, the Sun News Network/Abacus Data poll continued to roll, showing that the race remains stable province-wide. But going through the entrails of the poll we see that things are not so clear cut for Stephen McNeil's Liberals.

The projection continues to inch downwards for the Liberals, but the margin remains incredibly wide. The Liberals lead with a projected 48.1% of the vote, or a likely range of between 46.3% and 52.9%. The New Democrats are second with 29%, or between 29% and 31.9%. And the Progressive Conservatives are back over 20% with 21%, or a likely range of between 19.3% and 22.3% of the vote.

In terms of seats, the Liberals are still projected to win a majority government with between 29 and 36 seats, or 34 more specifically. Darrell Dexter's NDP is projected to win between 12 and 16 seats, or 12 more precisely, while the PCs of Jamie Baillie should win between four and seven seats.

The extremes open up some more interesting possibilities, however. With an extreme range of between 21 and 41 seats, the Liberals could theoretically win a landslide majority or an enlarged Official Opposition. This is a change for the party, which had for some time been over the majority threshold in even their worst case scenario. The New Democrats could still conceivably win a minority government, or as few seats as to put them in a tie with the Tories.

The projection will be updated either late tonight or on Tuesday morning to incorporate the last set of numbers from Abacus and any other polls that emerge on this election eve.

The latest poll from Abacus released yesterday shows very little change in voting intentions as the last days of the campaign slip by. Compared to their first independent sample, conducted between September 30 and October 1, no party has moved more than two points provincially. That is well within the margin of error.

The latest numbers have the Liberals at 47%, the NDP at 28%, and the PCs at 24%. That is a seemingly insurmountable lead. It gets wider among what Abacus estimates to be likely voters: 48% for the Liberals, 27% for the NDP, and 24% for the PCs.

But recent experience rewards caution, and there are a number of things in this poll that point to the potential, slim as it is, for a surprise tomorrow night.

Why the Liberals should win

The poll numbers are very strong for the Liberals. A 19-point lead is not something that is easily overcome. At this point of the Alberta campaign, the gap between Wildrose and the PCs was between seven and 10 points. In B.C., the gap between the NDP and the Liberals was between five and 14 points. With polling up to around three days before the vote, I was projecting a roughly five-point margin in Alberta and an eight-point gap in B.C. That is a far cry from 19 points.

The fundamentals also work in the Liberals' favour. They are ahead among likely voters and lead in all age groups, including the high-turnout oldest demographic. McNeil is out-polling Dexter on who would make the best premier by 12 points, or 18 points among those with an opinion. A majority of Nova Scotians, or 51%, have a positive opinion of the Liberal leader, with only 27% having a negative opinion. That positive score increases to 58% among likely voters. By contrast, Dexter has a positive rating of only 37% to 46% negative, increasing to a 39% to 48% split among likely voters. Only 28% of Nova Scotians think the NDP deserves re-election. And on who would be best able to handle the economy, the Liberals beat the NDP by 31% to 21%.

Put together, it makes it extremely unlikely that the Liberals would be able to blow a 19-point lead, either due to turnout, a late swing, or polling error.

And why they might not

But rather than look back at the polls with the benefit of hindsight, let's take a look now at what might prove problematic on election night. There is always the possibility of massive polling error and/or a late swing (even 22% of likely voters are undecided), but let's put that random occurrence aside and see what the Abacus poll itself tells us.

The Liberals could have difficulty winning enough seats to form a majority government. Abacus finds the race to be very close in Halifax and Cape Breton. The Liberal lead is only six points in the capital, with the margin of error for the regional sample at almost seven points. About half of Haligonians (49%) have a positive opinion of Dexter, just three points less than those who have a positive opinion of McNeil in the HRM. If the polls miss the mark even a little, or if turnout benefits the NDP in the city, the Liberals could end up losing a lot of close races.

And then there is Cape Breton. There are only eight ridings in this region, but if things go badly in Halifax than the Liberals will need to win in Cape Breton. The latest poll puts them only one point up on the Progressive Conservatives, who have an incumbency advantage in half of those eight ridings. If there is a late swing in both Halifax and Cape Breton, the Liberals might not be able to run the table enough in the rest of the mainland to form a government.

There are also some hints in Abacus's polling that the Liberals might not be as structurally strong as it appears on first glance. While the Liberals do lead when respondents were asked directly about the economy, the results were different among those who named jobs and the economy as their top electoral issue.

Among the 36% of respondents who chose that as their top issue, the NDP had a three-point lead over the Liberals on who was best able to handle it. On healthcare, a top issue for 29% of Nova Scotians, the Liberal lead was of only four points.

What this suggests is that while Nova Scotians generally see the Liberals as better economic managers, among those to whom this is most important the NDP has an edge.

But overall, the voting intentions and leadership numbers are so slanted towards the Liberals and Stephen McNeil that it seems that these small chinks in the armour can not be enough to overturn the expected result. We will have to see what the final polls of the campaign say before coming to a definitive conclusion, however.

Gap narrowing in Nova Scotia election, but Liberals still in control

Polls are starting to show a narrowing gap in Nova Scotia's provincial election, theoretically putting a Liberal minority in play. But Stephen McNeil still enjoys a very comfortable lead with only a few days to go before the vote.

The projection now gives the Liberals less than 50% support, as they have slipped to 48.9%. The New Democrats are approaching 30% once again, and currently stand at 29.6%. But the Progressive Conservatives have fallen just under the 20% mark to 19.4%. The likely ranges puts the Liberals at between 47% and 53.8%, the New Democrats at between 29.6% and 32.6%, and the PCs at between 17.8% and 20.5% of the vote.

It no longer seems likely that the New Democrats could finish third in either the popular vote or the seat totals, and the extremes of the seat projection does theoretically put the NDP in a position to win 25 seats to just 24 for the Liberals. This assumes a B.C. or Alberta degree of error in the polls, but to such an extent that the odds of it occurring are about 1 in 50 (if the election were held today). By comparison, at the end of the race in British Columbia I was giving the B.C. Liberals about a 1 in 5 chance of winning the most seats. But, if things go the other way, the Liberals could theoretically win 43 seats, with just six going to the New Democrats.

Overall, the model gives the Liberals between 32 and 37 seats, the NDP between 11 and 16, and the PCs either three or four. The precise projection is currently 36 seats for the Liberals, 11 for the NDP, and four for the Progressive Conservatives.

The Chronicle Herald/Corporate Research Associates poll released this morning is quite interesting, as it shows that some real movement is potentially taking place in these final days of the campaign. The overall sample, spanning a week from September 27 to October 3, shows stable numbers: 52% for the Liberals, 29% for the NDP, and just 17% for the PCs. But when looking at just the last three days of polling, the Liberals are reduced to 47%, the NDP is increased to 31%, and the PCs to 20%. That is much more in line with the polling being done by Abacus Data (as it should be since Abacus has only been polling in the last few days).

Of course, some caution needs to be exercised with the newer sample. Being a sample of just 270 decided and leaning voters, the margin of error is about six points. While that is large enough to exclude the possibility of the three parties being in any other order (it isn't as simple as just, for example, reducing the NDP by six points and increasing the PCs by six), it doesn't exclude the possibility of say, the Liberals being well over 50% and the NDP being well under 30%.

CRA's leadership numbers remain more or less unchanged, with McNeil at 36%, Darrell Dexter at 20%, and Jamie Baillie at 14% over the seven-day sample.

The Sun News Network/Abacus Data poll released last evening is only a partially new poll, as it simply adds another one day and 200 interviews to the poll that was released on Wednesday. With the newer data, the poll looks a lot like the latest, three-day sample from CRA. This erases the discrepancy the polls were showing before.

Abacus's likely voter model boosts the Liberals to 50%, against 28% for the NDP and 22% for the Tories.

Note the regional results, particularly in Halifax. With only a seven-point gap between the Liberals and NDP, a lot of seats should still be in play in the city.

It also worth noting that Abacus shows similarly strong leadership numbers for McNeil, most importantly his net +30 positive/negative rating. This is in sharp contrast to Baillie's +6 rating and Dexter's -7.

Also, the Liberals are beating the NDP on the economy by a margin of 31% to 18% and the party is doing a much better job of retaining their 2009 support. Fully 92% of Nova Scotians who said they voted Liberal in 2009 are planning to do so again, while only 49% of NDP voters from 2009 are sticking with Dexter.

The CH/CRA poll is the last of the campaign, so Abacus will take us the rest of the way. They will have new numbers out tonight from last night's polling, and will also be in the field on Saturday and Sunday. It is comforting that, just as CRA checks out and Abacus gets rolling, the numbers are converging. We will see where Abacus puts them through to the eve of the election.

NS Liberals remain on track for big win

The election in Nova Scotia will be held in only a few days' time and the polling has heated up. The rolling Chronicle Herald/Corporate Research Associates poll will conclude with a final report on the weekend, while the new Sun News Network/Abacus Data poll will run right up until the end of the campaign. With two surveys now in the field, we can hone in a little closer on what is going on.

Since the last major update last week, things have hardly budged. In the interim, however, the Liberals were flirting with an absolute landslide victory. Now, they just seem poised for a very comfortable majority.

Click to magnify

The projection model currently estimates Liberal support to fall between 48.9% and 56%, with the polls pointing more directly at 50.9%. That would translate into 35 seats for Stephen McNeil's Liberals, or a likely range of between 32 and 39 seats. The threshold for a majority government is 26 seats, and only an absolute worst case scenario currently points to the possibility of a Liberal minority.

Darrel Dexter and the New Democrats are projected to be at 26.4%, with their likely range stretching to 29% of the vote. In terms of seats, they would likely win between 12 and 15 if an election were held today.

The Progressive Conservatives have lifted themselves back out of the teens and now stand at 20.8%, or between 19.2% and 22.1% of the vote. This should give them between three and five seats, though if things go badly for the NDP there is an outside chance that Jamie Baillie could head up the Official Opposition.

Polling in the campaign so far has been remarkably consistent, but there has been some movement primarily between the Liberals and Tories. The injection of the Abacus polling throws in a little uncertainty as well, as the firm has the Liberals lower and the PCs higher than CRA's estimates.

Let's first take a look at the CH/CRA poll which, though stretching back further than the SNN/Abacus survey, does contain some newer information. But to put it in proper context, let's compare the latest tracking numbers to the last couple weeks.

As you can see, the difference between CRA's latest numbers and the last completely independent sample is marginal. The only shifts, a gain of one point for the Liberals and a drop of two for the NDP, is well within the margin of error. But when also looking at the survey from early on in the campaign, a trend is a bit clearer. For the New Democrats, it is absolute stability: they have wobbled between 27% and 29% in CRA's three independent samples. The PCs have dropped significantly from 23% to 17% (a statistically significant decrease). The Liberals have been consistently increasing, from 48% to 53% and now 54%. The overall gain of six points during the campaign is within the margin of error, but only just.

The new Abacus poll tells a similar story, but the details are slightly different.

The poll gives the Liberals 48% support compared to 26% for the NDP and 25% for the Progressive Conservatives. In other words, it agrees with CRA that the Liberals hold a virtually insurmountable lead. They also agree that the NDP is in the mid-to-high 20s.

But the big difference is, of course, the PCs at 25%. That is important for the party, as it is exactly where the Tories stood on election night in 2009. On the other hand, CRA is pointing to an all-time worst result. We will have to see if the final set of numbers from CRA shows an uptick that puts it more in line with Abacus's polling.

It also has to be recalled that CRA is incorporating data in the field running back to September 26 in their latest poll, while Abacus has data just from September 30. So comparing the surveys is like comparing one bag of apples to another bag of apples that includes a few rotten ones.

Abacus's extremely detailed report includes plenty of interesting information, including regional breakdowns. They show that the Liberals are leading throughout Nova Scotia, but that their advantage is less pronounced in Halifax.

The details in the poll make it clear just how much of an advantage the Liberals have. They lead in all age groups, including those older voters most likely to head out to the polls, are seen as the likely winners by half of the population, and lead the NDP on important issues like jobs and the economy and health care. McNeil beat out Dexter and Baillie by a margin of 30% to 19% and 15% on the Best Premier question - a margin similar, though somewhat tighter, than CRA's - and he has a much better favourability rating. Fully 53% of Nova Scotians polled said they had a positive opinion of him, compared to 37% for Baillie and 35% for Dexter. Perhaps more importantly, he also had the lowest negative rating: 22% to 35% for Baillie and 45% for Dexter. Those are murderous numbers.

Abacus is also calculating the voting intentions of who they consider likely voters. Among likely voters, Liberal support increases to 51% while the NDP stays at 26% and the PCs fall to 23%. Normally this is the kind of data I would prefer to use, but the samples are already very small and Abacus hasn't included regional breakdowns of likely support. So I will be sticking to the general-population numbers, but we should keep these calculations in mind. They point to the Liberals being able to out-perform their polls, while the PCs could under-shoot them.

Both polls still show a large proportion of undecideds. CRA reports 15% undecided and 16% no response. That has been dropping, though. The proportion of undecideds was around 22% at the beginning of the campaign. Abacus showed a much larger 35% undecided (and an additional 5% no response), but that is not too dissimilar from the total 31% undecided/no response that CRA is showing in their latest poll.

Yes, yes, British Columbia and Alberta have taught us all we'll ever need to know about prematurely calling a race. But both of those elections showed much closer gaps than the 22 and 27 point gaps that these two polls are showing. So we should be safe. Famous last words.

NS Liberals move solidly into majority territory

The Corporate Research Associates daily tracking poll for The Chronicle Herald shows little movement as the campaign in Nova Scotia progresses. And the only small movement it does show is away from the incumbent government and the third place Tories and towards the Liberals of Stephen McNeil.

I wrote about the state of the race and the election projection in this week's article for The Globe and Mail, and I invite you to check it out.

Since the last major update on this site, the projection has swung rather significantly in the Liberals' favour. The party is now projected to take 50.3% of the vote, up 4.8 points from where the Liberals stood with polling up to Sept. 18. That gives them a 23-point lead over the New Democrats, and even their likely low range still puts them 21 points up on the NDP. Stretched to its extreme, the Liberals would still win by 10 points even if the polls fall on their face as they have in recent elections.

Click to magnify

The New Democrats have slipped 1.1 points to 27.3%, while the Progressive Conservatives have dropped 3.1 points to 20.1%. They seem to have bit hit hardest by the gains McNeil has made during the campaign so far.

The Liberals are projected to win between 30 and 38 seats with these numbers, up from between 25 and 34 from Sept. 18. That puts them comfortably in majority territory. Their absolute minimum at this stage is 25 seats - not enough to win a majority, but enough to win the election.

Darrell Dexter's New Democrats have dropped to between 13 and 16 seats from between 14 and 21, and their maximum range (23 seats) now puts them below the Liberals. CRA polling does not support the potential for an NDP upset if the election were held today.

The PCs have dropped from between 3 and 7 seats to between 3 and 6, but more importantly their most likely result has dropped by three seats to only four. The Liberals picked those up, along with another seat from the NDP.

The Greens have dropped 0.9 points to 1.3%, and are expected to take between 0.9% and 1.3% of the vote. This is after taking into account their slate of only 16 candidates out of a possible 51. Because they have not managed a full slate, the likelihood of the polls over-estimating the Greens is amplified significantly, even if the tracking polls give them just 2% of the vote.

The most recent set of numbers from CRA, running from Sept. 18-24, is almost an entirely new sample from the analysis of last week (the polls overlap only for Sept. 18). They show no statistically significant movement over the last week, suggesting that the campaign is relatively stable. But the trends are pointing in the direction of the Liberals: they have picked up three points to hit 51% support, the highest they have been in CRA polling for as long as I can recall. The New Democrats have slipped one point to 27%, while the Progressive Conservatives have dropped three points to 20%.

On leadership, McNeil gets 37% (+2) on who would make the best premier, followed by Dexter at 20% and Jamie Baillie of the Tories at 15%. That is a drop of four points, just flirting with the margin of error. The campaign has not been going very well for Baillie, though the reviews of last night's debate were relatively positive for the PC leader. Nevertheless, McNeil's leadership edge is rather crushing. The recent election misses in B.C. and Alberta, for example, showed nothing like this.

Each dot in the chart below represents a result from a poll that was in the field on that particular day. As CRA's polling spans a week, we cannot know for certain when movement is taking place. But this sort of tracking gives an idea.

As you can see, the campaign so far has been pretty steady. The Liberals started the campaign at 48% before dropping to 47% and then upticking to 49% and now 51% support. Meanwhile, the NDP has wobbled between 27% and 29%, while the PCs have been pretty steadily sinking, from 23% to 21% and now 20%.

The consensus opinion on the debate last night was that all did pretty well, but Dexter and Baillie performed better than McNeil. But it doesn't seem like these were the kinds of performances that should single-handily change the game. We won't start seeing the effect of the debate until the weekend, when CRA's sample starts leaning post-debate. It will be interesting to see if things shift, because at this stage of the race the Liberals are going to win in a landslide.

Liberals gain in opening round of Nova Scotia election

The first polling numbers for the provincial election campaign in Nova Scotia were released yesterday. The poll by the Corporate Research Associates for The Chronicle Herald shows that Stephen McNeil's Liberals have extended their lead over the governing New Democrats as the campaign kicks-off. The projection accordingly puts the Liberals on track to win a majority government.

The Liberals have picked up 4.8 points in the vote projection since the last update incorporating polling data up to Aug. 31, and now lead with 45.5% support. The New Democrats have dropped 2.4 points to 28.4%, while the Progressive Conservatives are down 1.6 points to 23.2% support.

The likely vote ranges give the Liberals between 43.7% and 50% support, compared to 28.4% to 31.2% for the New Democrats and between 21.3% and 24.6% for the Tories. As even the maximum and minimum ranges of the Liberals and NDP do not overlap, this suggests that the Liberals should be able to win the popular vote unless a historic degree of polling error occurs.

From a flimsy minority, the Liberals have moved into a comfortable majority. The projection model gives the Liberals 30 seats with these numbers, a gain of eight since Aug. 31. The NDP has fallen six seats to 14 and the PCs have dropped two seats to seven. But a majority government is not completely in the bag for McNeil: the likely range still puts him as low as 25 seats, one short of a majority. His upper likely range is a far larger majority of 34 seats. The New Democrats would be likely to win between 14 and 21 seats with these levels of support, while the Tories could expect to win between three and seven seats.

The maximum range for the New Democrats, however, does put them ahead of the Liberals' minimum range. This means that if the polls completely miss the call and the gap narrows to some four points, the NDP could pull more seats out of the equation.

But the Liberals remain firmly in control of the situation. The CRA poll gave the Liberals 48% support among leaning and decided voters in Nova Scotia, a gain of seven points from the firm's last poll of Aug. 8-31 (which excluded leaners). That is a statistically significant increase in support, so it would appear that the campaign has gotten off to a good start for McNeil.

The New Democrats were down three points to 28% and the Progressive Conservatives were down two points to 23%, both shifts within the margin of error. The Greens were down one point to 2%. Of the entire sample of eligible voters, 22% were undecided.

The margin of error for the entire sample is +/- 4.0%, 19 times out of 20, and is +/- 4.9% for the sample of decided and leaning voters (406 of them).

McNeil picked up five points on who would make the best premier with 35%, putting him well ahead of Darrell Dexter's 20% and Jamie Baillie's 19%. This will be an interesting number to keep an eye on. In British Columbia, for example, Christy Clark was making bigger gains on this question than her party was in the final weeks of the campaign. Support for John Percy of the Greens for premier was not mentioned in the Herald report on the poll.

The Chronicle Herald is promising regular, daily updates on the polls in Nova Scotia. It will be interesting to see how this is handled as we have not seen daily tracking in any provincial campaign since New Brunswick's 2010 election. I encourage the Herald to be as copious with the data as possible. Their first report was missing a few bits of information that the Election Act requires, namely the field dates of the poll itself, but hopefully that will be rectified in the future.

Election laws in most provinces are actually quite strict on these sorts of things: for example, Nova Scotia requires a full report (including weighting schemes) be made available upon request (owners of the data can charge 25 cents per page!) at some point during the campaign. Similar laws exist elsewhere in Canada, but the laws are rarely followed to the letter. In Nova Scotia, the law requires someone like me, who is reporting the poll within 24 hours of its publication, to include all the same information that the poll sponsor is required to report! (As always, I will do my best to follow the laws of the province I am writing about.)

For the purposes of transparency and disclosure, it would be ideal if The Chronicle Herald released full data reports of their polls after a reasonable delay. After all, once the headlines are written that data is not worth very much anymore and of interest just to us wonks.

In any case, it seems that the race remains the Liberals' to lose. Support for the three parties has been generally wobbling back and forth for a long time now, and it is certainly not good news for the government that the gap is shrinking as the campaign begins. The opposite happened in British Columbia. We will see whether this race will perk up a little in the coming weeks.

NS NDP gains as Liberals move into minority territory

The Corporate Research Associates released their latest numbers for the provincial scene in Nova Scotia, showing that Stephen McNeil's Liberals still hold a significant lead over Darrell Dexter's New Democrats. But the numbers might not work very well for McNeil, as the Nova Scotia projection now puts him in a position to win only a minority government.

The vote projection is almost entirely based on the latest CRA poll, the only real difference coming from their May survey (taking up a whopping 0.4% of the projection) and the model's estimate of support for other parties and independents being 0.7% (rather than the 0% of CRA). So, the vote projection has moved exactly like CRA's latest poll: a four-point drop for the Liberals and a five-point gain for the NDP.

Taking into account past polling error, that puts the likely band of support for the Liberals at between 39.1% and 44.8% of the vote, compared to 30.8% to 33.8% for the New Democrats. The Progressive Conservatives remain in third with between 22.8% and 26.3% of the vote. (As a reminder, and for first time readers, the results highlighted in the 'boxes' in the projection chart are considered the most likely results, based on how the polls have been wrong in recent elections.)

Click to magnify

The seat projection has moved more dramatically. The Liberals have slid six seats in the projection to 22, putting them four seats short of a majority government. The NDP gained seven seats to reach 20, while the PCs were down one seat to nine.

Nevertheless, the Liberals are still in the better position as they have only dropped three seats in their likely result. They had been projected to take between 24 and 33 seats with CRA's May numbers, and they have since slid to between 21 and 30 seats. That puts them well in range of a majority government, and with the NDP's most likely range moving to between 20 and 22 seats Dexter's odds of victory are slim.

That is, if the polls are off by only an average amount. If the polls miss the call as wildly as they have in some recent campaigns, then an NDP majority is certainly possible (it was not considered possible when they trailed by almost 20 points - polls are never that wrong). Their maximum seat haul is 33, assuming enormous error in the polls. If it goes the other way, however, they could be reduced to 11 seats and perhaps displaced as the Official Opposition by the PCs.

The Liberal vote is simply not very efficient, and with CRA's polling showing the two parties neck-and-neck in the Halifax region the race is very much undecided. Particularly considering that 41% of respondents to CRA's poll were actually undecided.

That is a huge number, but nothing unusual for CRA. It has actually dropped two points from their last poll. Whether it is something that really needs to be of concern, however, is another matter. As Don Mills explained in an interview with me, CRA does not include 'leaning' voters in their polls outside of campaigns. That has the potential to inflate the size of the undecided vote considerably.

CRA found the Liberals to have fallen to 41% while the NDP had risen to 31%. That five-point gain was just outside the margin of error for decided voters, whereas the Liberals' four-point drop was just inside of it. The PCs slipped one point to 25%, while the Greens were unchanged at 3% support.

For the New Democrats and Liberals, this puts them back to where they were between August 2012 and March 2013. The PCs, meanwhile, have been stuck at around this level of support for years.

McNeil topped the leadership table with 30% saying he would make the best premier, compared to 19% for both Dexter and Jamie Baillie of the Tories. That is generally unchanged from May. But McNeil is down from the 35% he had on this question a year ago, while Dexter's numbers are getting worse compared to his party. He had 26% support for premier in November 2012, for example, at a time when his party was at 29% support among decided voters.

But in addition to the significant gain, the NDP has another small bit of silver lining in this poll. Government satisfaction rose to 42%, the highest it has been since November 2012, while dissatisfaction fell by four points to 45% - the lowest it has been since February 2012, when the NDP was polling at 44% and had a 17-point lead over the Liberals. Perhaps this suggests that those undecided voters may be satisfied enough with the government, but not yet sure they want to vote for it again.

Will this be enough for Dexter to call the election? Nova Scotians still expect the call any day now. Having put his party within spitting distance of the Liberals, when before they did not have a chance at all, is a minimally positive development. The 'momentum' may falter if Dexter waits until CRA's next quarterly report to pull the plug. And as we have seen time and again, campaigns still matter. Dexter has every reason to hope he can pull off what Alison Redford, Christy Clark, and Jean Charest (almost) did, in that they were all equally on the road to defeat.

The new projection model for the upcoming Nova Scotia election

The next provincial election likely to occur will be in Nova Scotia. Premier Darrell Dexter needs to call the vote by the spring of 2014, but all indications are that he will drop the writ some time in the late summer or early fall. Accordingly, ThreeHundredEight is today launching the Nova Scotia projection.

Unfortunately, there isn't a lot of polling data to go on just yet. The most recent numbers we have for the province date to the end of May, when Corporate Research Associates was last in the field for their quarterly polling. They should be out with the numbers from their next quarterly poll in early September.

The projection model has been given a few tweaks, as is always the case after each election - and particularly after such a humbling election as the one in British Columbia. One thing that has not been tweaked, however, is the seat projection model or the way that polls are weighed. The seat projection model has, time and again, shown itself to be a useful tool if the numbers plugged into it are accurate. That means the challenge is not to estimate the number of seats each party will win, but rather what proportion of the vote they will get.

A detailed description of how the polls are weighed and seats are projected can be found here. I will not go over that again, but rather explain some of the changes in the new model and the philosophy behind them.

First, a quick guide to what the new main projection graphic is showing (click to magnify):

At first glance, you will note that I have dropped forecasts for the election date from the chart. This is because I have decided to take a different approach to projecting electoral outcomes.

The model I used in the provincial elections of Alberta, Quebec, and British Columbia were based on a simple premise: what can the polls tell us about how they will be wrong? The inherent problem with that approach is obvious.

I was using the sample size of all the polls being taken into account by the model to estimate a rough margin of error for the projection. That gave me my high and low ranges. I then also used the degree of volatility from poll to poll to estimate the potential for future volatility in the days remaining between the last set of polls and the election day.

But if the polls are off, it does not make much sense to use those polls to guess at how they will be off. The data they are recording is not going to give a hint at the potential error if the foundations upon which they are built are faulty.

Instead, the approach for the Nova Scotia election (and, if it works well, for future elections) is to use the degree of error polls have shown in the past to assess the probable range of outcomes.

I was already using this sort of adjustment in the last few elections. This was applied directly to the poll average, using the average over- or under-estimation that the polls had done in past elections to guess at how the polls might over- or under-estimate party support in future elections (in B.C., this was done for the Greens and Conservatives and worked very well). But rather than use this information to make a best guess, I will instead be using it to give a likely range of outcomes.

This is calculated based on a party's position in the legislature at dissolution: the governing party, the Official Opposition, a third party with multiple seats, a third party with a single seat, and parties without a seat in the legislature. The electoral outcome for each of these parties in recent elections is then compared to the polling average.

All cases in which a party in a particular position in the legislature was under-estimated in the polls is then used to calculate the average "High". For example, the average under-estimation (when polls under-estimated a party's support) in recent elections for the governing party has been by a factor of 0.97. That means that the weighted polling average (which does make an independent estimate of "Other" support) is adjusted by a factor of 0.97. The same is done for cases of over-estimation to calculate the "Low", and these numbers are then used to project the number of seats that can be won at these high and low numbers.

The minimum and maximum projections ("Min." and "Max." on the chart) are simply the worst cases of over- and under-estimation that has occurred in recent elections. That means the Alberta Progressive Conservatives in 2012 (governing party, under-estimated), the Newfoundland and Labrador Liberals in 2011 (Official Opposition, under-estimated), Wildrose in 2012 (other party with multiple seats, over-estimated), etc. In other words, it means that if the election outcome falls outside of these minimum and maximum ranges, the polls have missed the target by an unprecedented amount. That this is a possibility should not be entirely discounted.

This hopefully gives readers a full understanding of the potential range of outcomes that are possible, based on past polling performance and what the data is showing. My role is not to make bets, but to try to figure out what the polls are saying and what they aren't saying, and to give people in idea of what to expect. But to narrow it down a little, I have also included boxes showing the range of most likely outcomes for each party. This means that each box represents the degree of polling error that has occurred in a majority of recent elections for a party in a similar legislative position. The chart below spells this out for the Nova Scotia vote:

As the governing party, there is 68% chance that the electoral outcome for the New Democrats will fall within the average-to-high range. This is the range that is highlighted in the main projection chart. If we want to extend that further, we can say there is a 79% chance it will fall between the low and high marks, or there is an 84% chance that it will fall within the average to maximum range. This suggests that we should expect the polls to under-estimate NDP support - though that is not necessarily what is going to happen.

For the Liberals as the Official Opposition, the range is not so tight. The most likely individual outcome is for the result to fall within the average-to-high range (42%), but it is more likely that it will fall outside of that range (the remaining 58%). To find the smallest range that incorporates the most likely outcome, we have to stretch that to the low-to-high range. There is a 63% chance that the outcome will fall within that range.

For the PCs, a third party with multiple seats, a slight over-estimation is the most likely individual outcome but we have to stretch the range also from low-to-high to get to a 67% chance. For the Greens, an over-estimation is almost certain: there is a 60% chance the outcome will fall within the low-to-average projection, and a 95% chance that it will between the minimum-to-average.

This is the extent of the probabilities that will be calculated for the province-wide projection. I was not pleased with the picture the probabilities painted in the B.C. election: the NDP was given a 98.3% chance of winning the popular vote, and a 83.3% chance of winning the most seats. Technically, that doesn't mean the probability forecast was wrong. Realistically, the polling data may not have supported a 1.7% chance of the B.C. Liberals prevailing. The seat projection probabilities were probably closer to the mark (expecting this sort of "B.C. Surprise" in four out of every five elections is not entirely unrealistic), but I've decided to drop this calculation for the time-being until I have the chance to go over in more detail what these calculations were based upon. I think the seat probabilities are on the right track (they were based, after all, on 308's track record), but the vote probabilities may have in the end been based on inappropriate data.

Probability calculations for individual seat calls, however, will remain a feature of the new model. They performed very well in the B.C. election, the first time they were calculated:

In the ridings where the confidence of a correct call was between 50% and 59%, 60% were called correctly. When the confidence was between 60% and 69%, the accuracy rate was 71%, and so on, as the chart above shows. The results of the B.C. election have been incorporated into the probability calculations, which has had the effect of narrowly boosting them upwards (yes, one of the ridings is called with 100% confidence!).

Unless more detailed polling data is made public, the model will not be presenting projections at the regional level. The model is designed for regional-level polling, however, splitting the province up into Cape Breton, the Halifax Regional Municipality, and the Rest of the Mainland. Corporate Research Associates divides the province like this, but I don't know yet whether this information will be available throughout the campaign. If this information is not available, regional support will be 'estimated' from province-wide polls in order for the data to be plugged into the model. I don't suspect a problem related to this, however. I got a peek at CRA's regional results for their last poll, and they generally mirrored what a proportional swing would calculate from province-wide results. It seems that, at this stage at least, support for the parties in Nova Scotia are increasing and decreasing proportionately at the regional level.

I am hopeful that these new measures will provide a good result in Nova Scotia. But it is a small province with which to launch a new methodology. The populations of Toronto or Montreal, not to mention a few other cities, are larger than that of Nova Scotia (accordingly, I am hoping to cover their mayoral elections). Why should readers who don't live in the province be interested?

Well - the province does have interesting elections. There are three competitive parties and a lot of ridings where all three parties have a chance of winning. The NDP government there is only in its first term and is the first to be elected in Atlantic Canada. A rebuke may not mean anything in particular for the federal New Democrats, but it certainly does not bode well. The Liberals have been buoyed throughout the region, perhaps due to the federal party's new appeal, and have not been in power in Nova Scotia since 1999. For the PCs, they need to get out of the trough that delivered them their worst performance in provincial history in 2009.

If that isn't enough, provincial results in Nova Scotia do track with those of the federal parties in the province generally well - if not always explicitly, at least relatively (they seldom go in different directions).

So, perhaps the provincial election result in Nova Scotia will tell us something about how the federal parties will do in the province in 2015. Perhaps not, however. The numbers generally tracked pretty well until recently.

I don't suspect a huge amount of polls for the race in Nova Scotia, unless one of the firms feel they have something to prove. But hopefully there will be enough to keep things interesting.

Provincial Liberals up throughout Atlantic Canada

Earlier this week, the Corporate Research Associates released their Atlantic Quarterly polls for all four provinces on the East Coast. The results show gains by the provincial Liberal parties throughout the region, with mixed results for the New Democrats and Progressive Conservatives. The polls also show that, with the exception of Prince Edward Island, incumbents are in trouble in every province.

Before we get to the numbers, a few notes on methodology. CRA uses traditional live-callers to put together their polls. They tend to have very large  proportions of undecideds, ranging in these four provinces between 35% and 55%. This means the sample sizes for decided voters tend to be smaller with appropriately larger margins of error.

Also, CRA releases no crosstabs or demographic data, but clients subscribe to the Atlantic Quarterly and so they receive more data than the unwashed masses. It would be preferable if CRA released all of their data relating to political questions, as they are picked up by media throughout the region.

Nova Scotia is the province next slated to go to the polls (either this fall or before June 2014), and the New Democrats are in trouble there. They have dropped six points since CRA's last poll from November, a statistically significant decrease that has put them in a tie with the Progressive Conservatives at 26%. The Liberals were up to 45%, though their gain was just within the margin of error for their results. These are recent record highs and lows for both the Liberals and NDP.

There was a big increase in undecideds, however, from 48% to 55%. That makes the election, whenever it is called, very important. Many voters are still on the fence. Stephen McNeil was the favourite choice for premier by 31%, followed by Darrell Dexter and Jamie Baillie at 18%. Satisfaction with the government stood at 40% to 49% dissatisfaction.

New Brunswick will hold its next election in the fall of 2014, and here too the incumbent government is in trouble. The Tories are down to 29%, while the Liberals were up to 41% under new leader Brian Gallant. The New Democrats were up to 27%. These are recent record highs (Liberals and NDP) and lows (PCs) for all three parties, though none of the shifts are outside the margins of error. Fully 45% of respondents were undecided, unchanged from February-March.

Gallant topped the premier question with 31%, followed by David Alward at 20% and Dominic Cardy of the NDP at 17%. Satisfaction stood at 38%, while dissatisfaction was up to 51%. Satisfaction with the government has been relatively more steady, however, than the PCs' support levels.

Prince Edward Island isn't heading to the polls until 2015. But Robert Ghiz is quite comfortably ahead, with 52% support to 26% for the Progressive Conservatives and 21% for the New Democrats. Here again, the shifts aren't statistically significant, and 41% were undecided.

Ghiz was at 38% for the best person to be premier, followed by Mike Redmond at 15% and Steven Myers at 11%. Satisfaction with the government was at 45%, with dissatisfaction at 51%.

Finally, and most interestingly, is Newfoundland and Labrador. They are also only holding their next election in 2015, but the political landscape is changing. The New Democrats retained a narrow lead with 37%, but the Liberals and Progressive Conservatives swapped places. The Liberals picked up 14 points since February-March and placed a close second with 36%, with the Tories down 11 points to 27%. Worth noting, however, is that the number of undecideds increased by nine points to 35%.

The NDP is still at an all-time high (though slightly down from February-March's numbers), while the Liberals are polling better than they have in quite a while. For the Progressive Conservatives, 27% is a far cry from the 70% (and more) support they were polling at before the 2011 election.

Premier Kathy Dunderdale ranked third among the leaders with 21% choosing her as the best person to be premier, a drop of 11 points. Dwight Ball of the Liberals (interim leader but a candidate to take over) was up seven points to 24%, while Lorraine Michael remained well ahead with 36%.

Satisfaction with the government dropped 13 points to only 32%, while dissatisfaction was up 11 points to 63%. These are astounding numbers: 26% were 'completely dissatisfied', while a year ago that was only at 9%.

In terms of seats, the incumbent governments would be defeated in New Brunswick, Nova Scotia, and Newfoundland and Labrador, while the Liberals in Prince Edward Island are safe.

Seat projection based on CRA's polling

In New Brunswick, Gallant's Liberals would win 34 seats and form a majority government, with Alward's Tories pushed to the opposition benches with 13 seats. The New Democrats would win eight.

McNeil's Liberals would win a majority government of 32 seats, with Dexter's NDP winning 11 and the Tories the remaining eight.

The Liberals would sweep Prince Edward Island, winning all 27 seats in the province.

In Newfoundland and Labrador, the Progressive Conservatives would be reduced to only seven seats but hold the balance of power in a minority legislature. The New Democrats would win the plurality of seats with 21, taking 13 of the 14 seats in the St. John's Region. The Liberals would win 20 seats, 15 of them in Newfoundland west of the Avalon and Burin peninsulas. How this arrangement would work out is anyone's guess.

Though the samples outside of Newfoundland and Labrador were two small to decisively determine that the Liberals were making real gains, the trends are quite clear. Aside from a small one-point uptick in PEI, the Liberals gained six points or more throughout the region. They are under-going a bit of renewal, with a leadership race underway in Newfoundland and Labrador and Gallant having been named the leader of the NB Liberals last year. Plus, there is the Justin Trudeau factor. It can't be ignored, considering how well the federal party is polling in Atlantic Canada.

The Progressive Conservatives had less clear results, with a big drop in Newfoundland and Labrador, a modest one in New Brunswick, a modest gain in Nova Scotia, and a larger one in Prince Edward Island. The trends are a bit clearer for the New Democrats, who were down everywhere but New Brunswick, where they only gained a point. This is, however, much less tied to the federal scene as the provincial Tories are polling better than their federal counterparts. But just like at the federal level, the region appears to be in major flux.

Approval ratings for 30 premiers and leaders

Angus-Reid's quarterly polls on premier approval ratings are always interesting, but this week they made it even better by releasing the approval ratings for all major party leaders in nine provinces (poor PEI was excluded, as even with a 7,000+ national sample the number of respondents would be tiny on the island).

It isn't the first time that these numbers have been released, as Angus-Reid was kind enough to provide me with them back in September. But hopefully they will continue to do this every quarter, as it provides a great deal of information.

Let's rank the 30 premiers, opposition leaders, and major party leaders in the nine provinces by approval rating.

Three of the top five have not changed since September: Brad Wall, Lorraine Michael, and Adrian Dix. Only Wall, Michael, and Danielle Smith have a better than 50% approval rating, while all others have under 50%.

Quite a few more leaders have a disapproval rating of 50% or more: David Alward, Tim Hudak, Pauline Marois, Dexter, Alison Redford, Amir Khadir, Christy Clark, and Kathy Dunderdale.

Note that the leaders of Québec Solidaire and the Quebec Liberals are not correct in this poll. That is understandable for the Liberals, who chose Philippe Couillard as their new leader (replacing interim leader Jean-Marc Fournier) in the midst of this polling. Khadir, however, should have been replaced by Françoise David. Khadir is no longer the male spokesperson for the party, and in effect the female spokesperson (David) is the de facto leader. I suspect she would have had a much higher approval rating than Khadir, and might have cracked the top ten.

It is hard to find any partisan advantage in these numbers. Conservatives like Wall, Smith, and Brian Pallister are ranked highly, while Clark, Dunderdale, and John Cummins are at the bottom. Stephen McNeil, Brian Gallant, and Dwight Ball score well for the Liberals, while John Gerrard and Raj Sherman do not (and you might include Clark with them as well). The NDP has three of the top five in Michael, Andrea Horwath, and Dix, but also Dexter in the bottom tier.

There does seem to be an anti-incumbent bias in the numbers, as nine of the top ten are third party or opposition leaders, while five of the nine premiers are ranked 20th or lower. There doesn't seem to be a gender bias, though: Redford, Dunderdale and Clark are at the bottom of the list but Michael, Smith, and Horwath are at the top.

Who has made the biggest gains since the September poll? If we look at it only in terms of ranking, we see that Ball made the biggest gain, jumping 18 spots to 10th overall. Gallant is ranked 17 spots higher than interim leader Victor Boudreau was, while Dominic Cardy gained ten spots to move into the top half of the list. Atlantic Canadian leaders were on the move.

The leader that had the largest fall was Redford, who was ranked second in September but is now ranked 25th. Her approval rating has just plummeted in the last few months. Dunderdale also fell, by 17 spots, to 28th, while Marois and Cam Broten (compared to John Nilson) each fell eight spots to 20th and 21st overall. Wall, meanwhile, held on to top spot while Cummins and Victor Lau were unchanged as the bottom two on the list.

But is an approval rating the best way to measure these leaders? Broten, for example, was just recently named leader of the Saskatchewan NDP and his approval rating of 32% is actually quite decent, considering his disapproval rating is only 19% (49% have yet to make up their minds).

The chart to the left ranks leaders by their net approval rating (approval minus disapproval) and it changes things quite a bit. Edit: An earlier version of this post had incorrect net ratings for Wall and Sterk.

Broten makes the biggest move, increasing by 15 spots to sixth overall. Sterk also jumps 13 spots, as she has a large "not sure" total as well.

The bottom four are all premiers, meanwhile, with Dexter, Redford, Clark, and Dunderdale at the bottom of the list. Marois is also in the bottom tier. Wall remains the only premier in the top tier, but now becomes the only premier in the top half, too. Wall also has the distinction of being the only premier ranked more highly than all of his main opponents.

Most other premiers are ranked below two opponents: Dunderdale, Clark, Redford, Dexter, Marois, and Alward are all third or fourth in their province. Selinger is ranked just a little more highly than Gerrard, while Kathleen Wynne is ranked much more highly than Hudak (though she trails Horwath).

Approval rating is a pretty important indictor of ability to win. If an election were held today, based only on voting intentions polls, the approval ratings leaders (not net ratings) in every province except Ontario and Quebec would win (and we can't even say that with certainty about Quebec, as Couillard might have ranked more highly than François Legault).

These rankings aren't anything to sneeze at - if the elections in B.C. and Nova Scotia go as expected, premiers will take up three of the top six spots in the approval ratings rankings in a few months. And if these ratings continue, one has to wonder whether some of the other leaders at the bottom of the list (Redford, Dunderdale, Cummins, maybe Hudak) will be on their way out in the next year or two. Gerrard is already leaving and is just waiting for his replacement, while Khadir shouldn't have been included at all. That accounts for seven of the bottom 10 spots in the approval ratings ranking, and eight in the net ratings ranking.

Are approval ratings the kiss of death? Possibly - three of the four names that have disappeared from the list since September were ranked 17th or lower.

Honeymoon in Ontario, divorce in Nova Scotia?

In Ontario, the premier is just a few weeks on the job and is currently enjoying a bump in the polls. In Nova Scotia, the premier is at the tail end of his first mandate and could be a few weeks away from being out of a job.

In my weekly article for The Globe and Mail, I take a look at what the polling shows when new premiers are named between elections. You can read the article here. The short version is that new premiers usually boost their party's numbers before falling back to earth, and that the odds of re-election are as good as a coin flip. Heads or tails, Premier Wynne?

In the latest of my twice-weekly articles for The Huffington Post Canada, I take a look at the latest survey out of Nova Scotia by the Corporate Research Associates. Darrell Dexter is expected to call the next election in the province soon, as he must call it before the summer of 2014. With the announcement that he will present a balanced budget in April, everyone in Nova Scotia seems to believe that will be the starting bell for the next campaign.

I delve into the numbers in my Huffington Post article, so I invite you to head over there to read the piece. I won't go over the results a second time here, but we can take a look at what kind of legislature these results would produce.

Nova Scotia will be using a new electoral map in the next campaign, with the number of ridings reduced from 52 to 51. Electoral geography can be very important in deciding results, particularly in a province like Nova Scotia where all three parties are competitive.

Seat projection for CRA poll

And despite leading by seven points, Stephen McNeil's Liberals might be unable to secure a majority. The projection, which uses the new boundaries, gives his party 24 seats to 20 for the New Democrats and seven for the Progressive Conservatives.

The Liberals need 26 seats for a majority, so it is certainly possible that the projection model could be off by at least two seats to the benefit of the McNeil. But it is just as likely to be off in the other direction, handing a few extra seats to the Tories or even the New Democrats.

Two factors could play an important role in deciding the outcome. The New Democrats have the advantage of incumbency, and a few of their MLAs may be hard to defeat. That puts a few extra seats in the NDP's column. On the other hand, the Liberals would have the advantage of being the front runner, and their vote distribution may turn out to be more advantageous in close races where voters realize that the Liberal candidate could sit in cabinet.

But people should beware before scoffing at this sort of counter-intuitive seat projection. In the run-up to the Manitoba election in 2011, this site was consistently projecting an NDP majority even when they were polling behind the Tories. News reports on the eve of the vote said that the election was too close to call - and it was, at least in terms of the popular vote. But this site projected that even with only a few percentage points separating the PCs from the NDP, Greg Selinger would be able to win a larger majority than the one he had when the campaign began. In the end, the projection model called 56 of the 57 ridings in the province correctly. In fact, the model under-estimated the size of the NDP's majority by one seat.

Could the same sort of thing happen in Nova Scotia, with the polls showing an easy Liberal victory but the seat result being far closer? The race in and around Halifax could be what decides the election: if the Liberals can grow their support disproportionately in the capital, they can knock off a few New Democrats and secure a majority. If the NDP holds firm in the city, they may be able to keep the Liberals to a minority. And if the races closes by a few more points, they could even hold on to government.