Latest national poll median date: October 20
Projections reflect recent polling graciously made publicly available by pollsters and media organizations. I am not a pollster, and derive no income from this blog.

Friday, October 4, 2019

Polling-Related Adjustments 2019, Round 4

This post reviews polls of smaller areas than the six standard polling regions (BC, AB, SK/MB, ON, QC, Atlantic), and polls providing breakdowns of smaller areas. Resulting adjustments to the model are indicated in bold. You can view earlier rounds of polling-related adjustments here (round 1)here (round 2) and here (round 3). (I may still update the Round 3 post for polls ending in September.)

Update Oct. 6: Mainstreet's Rosemont--La Petite-Patrie and Orléans riding polls added. No adjustment.

Update Oct. 9: Mainstreet's Saint-Hyacinthe--Acton riding poll added. Adjustments made.

Update Oct. 11: Mainstreet's Montmagny--L'Islet--Kamouraska--Rivière-du-Loup riding poll added. Adjustments made.

Update Oct. 12: Mainstreet's Beauce riding poll added. Adjustments updated.

Update Oct. 12: Léger's 10/7-8 QC breakdown and Campaign Research's 10/8-10 ON and MB/SK breakdowns added. Regional adjustments updated.

Update Oct. 15: Mainstreet's Sherbrooke riding poll added. Adjustments updated.

Update Oct. 16: Mainstreet's Trois-Rivières riding poll added. Adjustments made.

Update Oct. 19: The following are added:
- Mainstreet's 10/9 Jonquière poll. Adjustments updated.
- Insights West's 10/13-16 BC poll. Adjustments updated.
- Campaign Research's 10/16-18 MB/SK breakdown. Adjustments updated.

Update Oct. 19: The following are added:
- Nanos' 10/4-13 QC breakdown, Forum's 10/11 QC poll and Léger's 10/13-15 QC poll. Adjustments updated.
- Mainstreet's Québec, Vancouver Granville and Laurier--Sainte-Marie riding polls. Adjustments made/updated.

Update Oct. 20: Léger's 10/17-18 QC breakdown added. Adjustments updated.

Update Oct. 20: Nanos' 10/7-13 ON breakdown and Léger's 10/10-14 GTA poll added. No change to adjustments.

Update Oct. 20: Campaign Research's 10/16-20 ON breakdown and IRG's 10/15-17 battleground region breakdown added. Adjustments updated.

I. Regional Polls

Atlantic Canada
Innovative Research's 10/15-17 poll
IRG shows the Liberals performing better than expected in Atlantic cities at the expense of the NDP, and the reverse pattern holding in rural areas. The following adjustments are added and overlaid on top of the provincial adjustments made previously:
- 10 urban ridings: NDP -2.5, LIB +2.5
- 22 rural ridings: NDP +1, LIB -1

Campaign Research's 9/30-10/2 poll
This poll broadly confirms previous adjustments. No change.

Léger's 10/7-8 poll
This poll shows that the recent upheaval in QC was far from uniform, with the BQ rising far more outside Montreal CMA than inside, and the Tory vote holding up better in Quebec City than elsewhere. Major changes (in bold) to sub-regional adjustments were made:
Quebec City CMA: CON -4, NDP +1, LIB +3, BQ -1
Montreal CMA: CON -2.5 (-1 in far suburbs); NDP +3 (Island), +2.5 (near suburbs); LIB +3.5 (Island), +3 (near suburbs), +1 (far suburbs), GRN +0.5 (Island of Montreal only), BQ -3.5 (-0.5 in far suburbs)
Rest of Quebec North: CON +3.5, NDP -3, LIB -4, BQ +3.5
Rest of Quebec South: CON +4, NDP -2.5, LIB -4.5 (-3.5 in Eastern Townships), GRN -0.5, BQ +3.5
(Salaberry--Suroît is mostly outside of Montreal CMA but shares some characteristics with it. Since the Montreal CMA and Rest of QC S adjustments go in opposite directions, I have left this riding unadjusted. Moreover, due to the large difference between Quebec City and Rest of QC North adjustments, I am computing the adjustment in Beauport--Côte-de-Beaupré--Île-d'Orléans--Charlevoix and Portneuf--Jacques-Cartier according to their rough population distribution: 55% and 58% respectively in Quebec City CMA, 45% and 42% outside.)

Nanos' 10/4-13 breakdown
Forum's 10/11 poll
Léger's 10/13-15 poll
Since Nanos' poll had a very coarse regional breakdown, and since Forum's regions were not clearly defined, they had very little influence on the adjustments. Instead, almost all of the changes are based on the Léger poll, which gave a fine regional breakdown thanks to its sample size of over 3,000. We learn that the Liberal vote is less concentrated in Montreal than thought, perhaps because that's where voters are most likely to defect to the NDP due to Singh's debate performance. Eastern QC is an area where the Bloc has seemingly risen less, and where the Liberals are holding up better. The adjustments are (I'm not bothering to bold the ones changing since they're almost all changing, except for the Greens):
- Island of Montreal: CON -1.5, NDP +3, LIB +1, GRN +0.5, BQ -2.5
- Laval and near South Shore suburbs: CON -1.5, NDP +2.5, LIB +1, GRN +0.5, BQ -2
- Far South Shore suburbs: CON -1, LIB +1, BQ -0.5
- North Shore suburbs: CON +2.5, NDP -3, LIB +1
- Quebec CMA: CON -5, NDP +2, LIB +1.5, BQ -0.5
- Rest of South of St. Lawrence down to and including Rivière-du-Loup and Mauricie: NDP -1, LIB -2, GRN -0.5, BQ +3
- East of Rivière-du-Loup, Saguenay--Lac-Saint-Jean, Côte-Nord: CON +4, NDP -3, LIB +1.5, GRN -0.5, BQ -1.5
- Lanaudière and west (outside Montreal CMA): CON +3, NDP -4, LIB -3, BQ +4

I have canceled the riding adjustment for Montmagny--L'Islet--Kamouraska--Rivière-du-Loup. I have not applied these changes to Beauce, Shefford and Compton--Stanstead, who have old riding adjustments. The above note concerning Salaberry--Suroît, Beauport--Côte-de-Beaupré--Île-d'Orléans--Charlevoix and Portneuf--Jacques-Cartier applies.

Léger's 10/17-18 poll
The final Léger poll shows a different support pattern across the different regions of QC. The adjustments above are changed as follows (these are overlaid on top of those above):
- Montreal CMA (except far suburbs): CON +0.5, NDP -1, LIB +1.5, BQ -1
- Quebec CMA: GRN -0.5, BQ +0.5
- Rest of QC: CON -0.5, NDP +1, LIB -1.5, BQ +1
(Same comments as above apply.)

Campaign Research's 9/30-10/2 poll
There are some minor changes, in bold, to the adjustments, which are now:
7 central Toronto ridings: NDP -4, LIB +3, GRN +1
Halton/Peel: CON -1.5, NDP +2.5, GRN -1
York and Durham inside Toronto CMA: CON -1.5, NDP +2.5, GRN -1
Durham outside Toronto CMA: CON -1, NDP +2, GRN -1
Simcoe: CON -2, NDP +1.5, LIB -0.5, GRN +1
Hamilton/Niagara: CON -1.5, NDP +0.5, LIB +1, GRN 0
Southwest: CON +0.5, NDP -1.75, LIB +0.75, GRN +0.5
East outside Ottawa: CON -1, NDP +0.5, LIB -1.5, GRN +2
City of Ottawa: CON +2, NDP +1.5, LIB -3.5
North: CON +5, NDP -4, LIB 0, GRN -1

Campaign Research's 10/8-10 poll
Here are the tweaks to the adjustments:
7 central Toronto ridings: NDP -5, LIB +4, GRN +1
Halton/Peel: CON -1, NDP +3, LIB -1, GRN -1
York and Durham inside Toronto CMA: CON -1.5, NDP +2.5, GRN -1
Durham outside Toronto CMA: CON -1.5, NDP +2.5, GRN -1
Simcoe: CON -2, NDP +1.5, LIB -0.5, GRN +1
Hamilton/Niagara: CON -2, NDP +1, LIB +1, GRN 0
Southwest: NDP -1.5, LIB +1, GRN +0.5 [CON adjustment removed]
East outside Ottawa: CON +0.5, LIB -2.5, GRN +2 [NDP adjustment removed]
City of Ottawa: CON +1.5, NDP +2, LIB -3.5
North: CON +5, NDP -5, LIB +1, GRN -1

Nanos' 10/7-13 breakdown
Léger's 10/10-14 GTA (actually Toronto CMA) poll
These polls together had no effect on the Ontario adjustments. Both were fairly neutral in the City of Toronto. Léger's poll was very negative for the Liberals in the 905 (relative to uniform swing, using Léger's recent ON results as baseline), but Nanos' poll was positive for them there, so overall, it's consistent with the current slightly negative LIB adjustment. Other deviations from uniform swing were in line with current adjustments.

Campaign Research's 10/16-20 poll
Here are the tweaks to the adjustments:
7 central Toronto ridings: NDP -5, LIB +4, GRN +1
Halton/Peel: CON -0.5, NDP +2.5, LIB -0.5, GRN -1.5
York/Durham: CON -1.5, NDP +2.5, GRN -1
Simcoe: CON -2, NDP +1.5, LIB -0.5, GRN +1
Hamilton/Niagara: CON -1.25, NDP 0, LIB +0.75, GRN +0.5
Southwest: CON -1, NDP -1.5, LIB +1.5, GRN +1
East outside Ottawa: CON +0.75, NDP +0.5, LIB -2.75, GRN +1.5
City of Ottawa: CON +2, NDP +3, LIB -4.5, GRN -0.5
North: CON +5, NDP -4.5, LIB +1, GRN -1.5

Campaign Research's 9/30-10/2 poll
Campaign Research's 10/8-10 poll
This These polls suggests that the Liberals are holding up better than expected in SK, and less so in MB. The latter is consistent with the recent Probe Research MB poll. However, earlier Campaign Research polls showed the opposite, so I will hold off for more evidence before making an additional adjustment here. New adjustments: LIB -0.9 in MB, +1.05 in SK. Opposite adjustments are made for the NDP; added to existing adjustments, the totals are: NDP +2.4 in Winnipeg, +0.9 in MB, -2.05 in SK.

As for the adjustments made previously (lower CON, higher NDP in Winnipeg; opposite in SK), this poll confirms them, and if anything suggests they could be larger. Thus, they are upheld, and may be increased going forward. these polls, taken together, confirm them.

Campaign Research's 10/16-18 poll
This breakdown suggests that increases to the pro-CON/anti-NDP adjustment in SK and pro-NDP/anti-CON adjustment in MB are warranted. The total adjustments are now:
- Winnipeg: CON -2.4, NDP +3.3, LIB -0.9
- Rest of MB: CON -0.9, NDP +1.8, LIB -0.9
- SK: CON +2.05, NDP -3.1, LIB +1.05

British Columbia
Insights West's 10/13-16 poll
This poll suggests that maintaining or increasing previous adjustments (Rest of BC: CON+, NDP-; Vancouver Island: LIB+, GRN-; opposite in Metro Vancouver) are warranted. Furthermore, the Tories appear to be doing a bit better than expected on Vancouver Island, and the NDP a bit worse. The new adjustments are:
- Metro Vancouver: CON -1.5, NDP +2, LIB -1.5, GRN +1
- Vancouver Island: CON +1.5, NDP -1.5, LIB +3, GRN -3
- Rest of BC: CON +2.5, NDP -3.5, LIB +1
Note: Fraser Valley ridings and West Vancouver--Sunshine Coast--Sea-to-Sky Country are not adjusted, as these ridings (per majority of population) are not part of Metro Vancouver, but they are not rural ridings and may be included in Metro Vancouver by Insights West. The riding adjustment in Vancouver Granville supersedes the Metro Vancouver adjustment.

Innovative Research's 10/15-17 poll
IRG largely confirms that previously made adjustments should be larger, except in the rest of BC, where it appears that the Tories are not overperforming quite as much.
- Metro Vancouver: CON -1.5, NDP +2, LIB -2, GRN +1.5
- Vancouver Island: CON +2.5, NDP -2, LIB +4, GRN -4.5
- Rest of BC: CON +1.5, NDP -3, LIB +1.5

II. Riding Polls (by Mainstreet unless otherwise noted)

Note: A riding poll will result in a model adjustment ONLY IF it shows a substantial deviation from regional swing and/or it is in a riding with unusual factors. Most riding polls will therefore be ignored: they generally did not help much in 2015.

Rosemont--La Petite-Patrie, ? (published 10/6)
This poll's results are almost uncannily similar to the model projection! No adjustment.

Orléans, ? (published 10/6)
The difference between the model projection and these polls results is well within the poll's margin of error for all parties. No adjustment.

Saint-Hyacinthe--Acton, 10/2
According to this poll, contrary to model expectations, the Tories are not competitive in this riding. Adjustments: CON -6, NDP +1, LIB +3, GRN -1, BQ +3.

Montmagny--L'Islet--Kamouraska--Rivière-du-Loup, 10/5
This poll has the Tory incumbent lower, and both the Bloc and Liberal candidates a bit higher than expected. Adjustments: CON -4, LIB +2, BQ +2 (CON -4, LIB +4 after 10/7-8 Léger update, cancelled after 10/13-15 Léger update).

Beauce, 10/9
This poll still shows a wafer-thin Conservative lead, despite the Tory drop in QC. On the other hand, Bernier has a slight advantage with older voters, which helps him. I have reduced the PPC adjustment from +30 to +28.5, while the CON adjustment goes from -30 to -29 (the missing 0.5% due to the age issue). Note, however, that Bernier's strong performance in the Oct. 10 debate may give him a boost.

Sherbrooke, 10/?
This poll is worse for the Bloc and better for the NDP than expected, probably because the previous Mainstreet poll of this riding did not mention candidate names, which likely hurt the NDP incumbent. The new adjustments have both the Liberals and NDP pulling from the Bloc and Tories (rather than the Liberals pulling from the NDP and Tories): CON -2, NDP +3, LIB +4, BQ -5 (NDP +2, LIB +3, BQ -5 after 10/13-15 Léger update).

Trois-Rivières, 10/9
This poll suggests that the star Tory candidate, former mayor of Trois-Rivières, is pulling a personal vote that could see him overcome Scheer's unpopularity in QC. Adjustments: CON +5, NDP +1, LIB +1, BQ -7 (CON +7, BQ -7 after 10/13-15 Léger update).

Jonquière, 10/9
This poll confirms the earlier adjustment of the NDP incumbent taking from the BQ. It also suggests additional adjustments where the NDP also takes from the Liberals and Greens. Thus, the adjustments are: NDP +6.5, LIB -2, GRN -2, BQ -2.5 (NDP +9.5, LIB -5, GRN -2, BQ -2.5 after 10/13-15 Léger update). The BQ remains well ahead, however.

Québec, 10?
This poll deviates significantly from the projection, which had Duclos falling behind the BQ challenger due to the BQ surge. However, the poll suggests that Duclos is holding his vote. Adjustments: NDP -4, LIB +7, GRN -1, BQ -2.

Vancouver Granville, 10/15
This poll has almost exactly the same gap between Wilson-Raybould and the Liberal challenger as the projection. However, both are higher than expected, as is the Tory candidate. Unfortunately, the NDP and Green numbers were not published, so I split the negative change for the NDP and Greens proportionally to their vote shares. The adjustments (which are in lieu of the Metro Vancouver adjustment) become: CON -6, NDP -15, LIB -8, GRN -3, Wilson Raybould +32.

Laurier--Sainte-Marie, 10/16
This poll shows Liberal star candidate Guilbeault slightly behind the Bloc, as the model expected. However, both frontrunners are doing better than expected, at the expense of the NDP and Greens. Adjustments: NDP -3.5, LIB +2.5, GRN -2, BQ +3. (This is in addition to the baseline adjustments for Duceppe's absence and Guilbeault's presence added earlier and to the Montreal adjustments.)

No comments: