Apart from general information on opinion polls, this page provides tables of polling results for you to keep track of changes in voters' preferences as the campaign unfolds. While the focus of most media attention is often on the national levels of support for the various parties, in the end the regional and local levels of support are much more indicative of how parties may fair on election day. For a more complete picture of the state of the campaign, compare the parties' standings in separate tables for:

SES Research provides an interesting insight into Canadians' second choice of whom to vote for, drawn from polling conducted between June 14 and 22.

Look back to the evolution of public support for the parties during the 2000 election campaign.

Reflections on Polling

Jeff Gray, of the Globe & Mail, wrote up an excellent primer on technical issues in opinion surveys during the 200 election, and it is still pertinent for the current election. The Library of Parliament also has an interesting background paper on Public Opinion Polling in Canada. 

Matthew Mendelsohn and Jason Brent of Queen's University provide a useful guide to Understanding Polling and Margin of Error - this is essential reading for an insight into just what significance we can attach to particular survey results. One thing to keep in mind is that most polls are published with the national margin of error reported; but the margin of error will be much higher for any provincial or regional figures that are also reported.

The Laurier Institute has an interesting collection of material on opinion polls and electoral support for political parties.

The fact that individuals say they are likely to vote for a certain party does not necessarily mean that they actually favour that party's policy positions.  A June 2004 poll by Ekos found that the Conservative Party may have had the largest group of potential voters at the time, but respondents actually favoured polices found in the Liberal and NDP platforms. This phenomenon poses a problem when parties are voted in as a negative reaction to the incumbent governing party - the voters may really be teaching the old rascals a lesson rather than endorsing the new party's policies.

When reading opinion poll results pay close attention to how large the sample size and "undecided" figures are.  A large sample makes for more accurate results, particularly at the regional level.  A small undecided figure means that the polling company has probed with follow up questions to finding out who a person is most likely to vote for, and included them in the pool of decided voters.  But these leaners, as they're known, are expressing a preference, not a decision.  They can also be quite volatile and either change their party preference or even decide not to vote at all. Also be cautious in reading too much into polls that were conducted with only one day of interviewing.  Generally speaking, several days of interviews are thought to be preferable in order to ensure more chance of capturing a wider range of society in the poll's sample.

Daily Polls...

CPAC and SES Research have combined to beat the big news networks to the punch, by offering daily opinion poll results of the national standings of the main political parties. 200 people are interviewed every day, and every three day's worth of sampling are combined for 600 in the current day's results. This is a good way to keep track of the evolving pulse at a national level. Unfortunately, the sample size is too small to rely on for regional-level standings. Note that the final interviews conducted June 22-24 had double sampling for a total N of 1200 in the last poll.

Polls vs. Election Results

Election night on June 28th brought a very different result from the what the media and academic predictions had led everyone to believe.  One explanation lies in the way in which most polling companies rolled "leaners" in together with truly "decided" voters. Many of those leaners obviously had second thoughts in the last few days of the campaign and decided not to vote Conservative.  An indication that this could happen is seen in the last Compas poll that revealed that more respondents believed that a Liberal-led government (either majority or minority) would be better for the country than a Conservative-led government by a margin of 41% to 28%. It is interesting that the level of support for a Conservative-led government was 5 percentage points lower than those who said they were thinking of voting Conservative (33%).In Decmeber 2004 Ekos Research issued its own analysis of the apparent disparity, arguing that last-minute voter shifts occured that polling companies were prohibited from publishing in the last two days.

Link to Poll	Polling Data
Election Day	Share of national votes	12.4	29.6	4.3	36.7	15.7
Ekos	June 21-24 n=5254 error= +/- 1.4% Undecided: n.a.%	11	32	5	33	19
SES	June 22-24 n=1200 error= +/- 2.8% Undecided: 19%	12	30	4	34	20
Compas	June 22-23 n=810 error= +/- 3.5% Undecided: 23%	13	33	4	34	15
Ipsos	June 21-23 n=2000 error= +/- 2.2% Undecided: 8%	12	31	6	32	17
SES	June 21-23 n=1000 error= +/- 3.9% Undecided: 20%	12	30	3	34	21
Environics	June 17-22 n=1444 error= +/- 2.5% Undecided: n.a.%	11	33	5	33	18
SES	June 19-21 n=600 error= +/- 4.1% Undecided: 21%	10	31	4	34	21
Ipsos	June 18-20 n=1000 error= +/- 3.1% Undecided: 9%	13	28	6	34	16
SES	June 18-20 n=600 error= +/- 4.1% Undecided: 19%	12	33	5	33	18
SES	June 15-17 n=600 error= +/- 4.1% Undecided: 20%	10	29	5	34	22
Ekos	June 16-17 n=895 error= +/- 3.3% Undecided: n.a.%	14	31	4	29	21
Ipsos	June 16-17 n=1000 error= +/- 3.1% Undecided: 9 %	12	32	7	29	16
Compas	June 16 n=600 error= +/- 4.0% Undecided: n.a.%	11	34	-	35	17
Léger	June 15-16 n=895 error= +/- 3.3% Undecided: 17%	12	35	4	34	16
Pollara	June 15 n=1267 error= +/- 2.0% Undecided: n.a.%	12	36	-	31	16
SES	June 13-15 n=600 error= +/- 4.1% Undecided: 18%	12	34	3	32	19
Ipsos	June 11-14 n=1000 error= +/- 3.1% Undecided: 9%	12	32	6	31	17
SES	June 11-13 n=600 error= +/- 4.1% Undecided: 15%	11	34	5	33	18
Ekos	June 7-9 n=2117 error= +/- 2.1% Undecided: 12%	12	34	5	30	19
Ipsos	June 4-8 n=2000 error= +/- 2.2% Undecided: 13%	11	31	7	32	16
Léger	June 2-7 n=1500 error= +/- 2.5% Undecided: 14%	9	30	-	33	19
Ipsos	June 1-3 n=1001 error= +/- 3.1% Undecided: 11%	11	31	6	32	17
SES	June 1-3 n=600 error= +/- 4.1% Undecided: 17%	9	34	3	37	17
Léger	May 27-31 n=3107 error= +/- 1.8% Undecided: 14%	12	30	5	35	17
Ipsos	May 28-30 n=1000 error= +/- 3.1% Undecided: 12%	11	30	6	34	16
Ekos	May 25-27 n=1306 error= +/- 2.8% Undecided: n.a%	11	30	-	38	18
Compas	May 14-19   n=1000* error= +/- 3.1% Undecided: n.a.%	11	31	-	39	17
Ipsos	May 11-17   n=2000 error= +/- 2.2% Undecided: 3%	12	26	5	39	15
Ipsos	May 7-13   n=2002 error= +/- 2.2% Undecided: 6%	11	24	5	40	15
Compas	April 30 - May 2 n=600 error= +/- 4.1% Undecided: 13%	12	27	-	39	20

Notes:

*The core national representative sample was approximately 1000 with over-samples in BC and Alberta, for a total sample of 1579.

Links to Polling Companies

• Compas
• Corporate Research Associates
• Decmia
• Ekos
• Environics
• Ipsos Canada
• Léger Marketing
• Mustel
• Pollara
• SES Research
• TNS Canadian Facts
• Queen's University Public Opinion Surveys Archive

Predictions...

Predicting the number of seats a party will win is as much witchcraft as number crunching and intelligence gathering on local events.  As election day draws near, it is perhaps the most fun aspect and wildly unreliable aspects of the spectator sport that elections provide.  My own predictions on June 27th were:

In the end, however, the actual election results showed that voters ended up giving the Liberals significantly more seats that the Conservatives. Other published predictions were:

Note* Prof. Barry Kay of the Laurier Institute for the Study of Public Opinion and Policy gives detailed estimates of party seat shares based on current opinion poll data are provided; he includes Chuck Cadman as a win under the Conservative Party even though he is running as a independent. During the 2000 elections, Barry Kay also provided projections (pdf format) and you can check their accuracy against the actual results.

Check out Milton Chan's Riding By Riding Information Exchange, which relies on correspondents with local knowledge to predict each riding's likely outcome. A similar service run by David Savitt for the 1997 election predicted 89% of the winners.

For a more complete picture of the state of the campaign, compare the parties' standings in publicly released polls conducted by 10 different research companies. Follow the developments in separate tables for:

SES research also provides a concise report on changes in regional support for the period May 30-June 24.

Link to Poll	Polling Data
SES	June 22-24 n=192 error= +/- 7.2% (province) Undecided: 19%	38	6	28	28
Ipsos	June 21-23 n=800 error= +/- 3.5% (province) Undecided: 4%	38	8	25	26
Ipsos	June 21-23 n=400 error= +/- 5.0% (region) Undecided: 4%	37 Lower Mainland	9 Lower Mainland	28 Lower Mainland	24 Lower Mainland
Ipsos	June 21-23 n=154 error= +/- 9.0% (region) Undecided: 4%	32 Island/Coast	11 Island/Coast	26 Island/Coast	30 Island/Coast
Ipsos	June 21-23 n=233 error= +/- 6.6% (region) Undecided: 3%	46 Interior/North	6 Interior/North	20 Interior/North	28 Interior/North
Ipsos	June 18-20 n=1000 error= +/- 3.1% (national) Undecided: 9%	36	8	32	18
Mustel	June 10-17 n=721 error= +/- 3.6% (province) Undecided: 10%	36	7	26	28
Mustel	June 10-17 n=721 error= +/- 3.6% (province) Undecided: 10%	18 Vancouver	6 Vancouver	37 Vancouver	39 Vancouver
Ipsos	June 16-17 n=1000 error= +/- 3.1% (national) Undecided: 9 %	34	7	32	21
SES	June 12-16 n=135 error= +/- 8.7% (province) Undecided: 16%	37	12	27	23
Ipsos	June 11-14 n=1000 error= +/- 3.1% (national) Undecided: 9%	37	12	27	23
Ekos	June 7-9 n=2117 error= +/- 2.1% (national) Undecided: 12%	41	5	29	26
Ipsos	June 4-8 n=2000 error= +/- 2.2% (national) Undecided: 12%	34	13	30	20
Ipsos	June 1-8 n=1066 error= +/- 3.0% (province) Undecided: 11%	33	10	27	27
Léger	June 2-7 n=1500 error= +/- 2.5% (national) Undecided: 14%	31	12 "Other Parties"	31	26
Ipsos	June 1-3 n=1001 error= +/- 3.1% (national) Undecided: 11%	31	13	32	23
Léger	May 27-31 n=3107 error= +/- 1.8% (national) Undecided: 14%	32	9	29	28
Ipsos	May 28-30 n=1000 error= +/- 3.1% (national) Undecided: 12%	39	-	34	14
SES	May 26-30 n=118 error= +/- 9.2% (provincial) Undecided: 21%	37	5	31	26
Ekos	May 25-27 n=1306 error= +/- 2.8% (national) Undecided: n.a%	37	-	29	30
Compas	May 14-19   n=1000*  (province) error= +/- 3.1% Undecided: n.a.%	40	-	32	26
Compas	Vancouver/FraserValley May 14-19 n=207** error= +/- 7.0% (regional) Undecided: n.a.%	45	-	31	22
Ipsos	May 11-17   n=2000 error= +/- 2.2%  (national) Undecided: 3%	31	5	33	27
Ipsos	May 7-13   n=2002 error= +/- 2.2% Undecided: 6%	23	10	37	32
Compas	April 30-May2   n=600 error= +/- 4.1% Undecided: 13%	31	-	33	32

Link to Poll	Polling Data
SES	June 22-24 n=145 error= +/- 8.3% (province) Undecided: 10%	62	2	23	13
Ipsos	June 21-23 n=2000 error= +/- 2.2% (national) Undecided: 8%	59	6	20	11
Ipsos	June 18-20 n=1000 error= +/- 3.1% (national) Undecided: 9%	45	11	25	17
SES	June 12-16 n=88 error= +/- 10.6% (province) Undecided: 11%	51	5	32	14
Ekos	June 7-9 n=2117 error= +/- 2.1% (national) Undecided: 12%	57	4	25	14
Léger	June 2-7 n=1500 error= +/- 2.5% (national) Undecided: 14%	52	3 "Other Parties"	30	15
Ipsos	June 1-3 n=1001 error= +/- 3.1% (national) Undecided: 11%	50	10	27	15
Léger	May 27-31 n=3107 error= +/- 1.8% (national) Undecided: 14%	54	4	28	12
Ipsos	May 28-30 n=1000 error= +/- 3.1% (national) Undecided: 12%	49	-	22	14
SES	May 26-30 n=88 error= +/- 10.7%(province) Undecided: 13%	51	2	33	14
Ekos	May 25-27 n=1306 error= +/- 2.8% (national) Undecided: n.a%	55	-	31	12
Compas	May 14-19   n=1000* error= +/- 3.1% Undecided: n.a.%	57	-	27	14
Ipsos	May 11-17   n=2000 error= +/- 2.2%  (national) Undecided: 3%	50	5	31	10
Ipsos	May 7-13   n=2002 error= +/- 2.2% Undecided: 6%	51	4	29	10
Compas	April 30-May2   n=600 error= +/- 4.1% Undecided: 13%	40 AL, SK,MB	-	33 AL, SK,MB	22 AL, SK,MB

Link to Poll	Polling Data
SES	June 22-24 n=137 error= +/- 8.5% (province) Undecided: 14%	34	2	39	24
Ipsos	June 21-23 n=2000 error= +/- 2.2% (national) Undecided: 8%	36	5	32	25
Ipsos	June 18-20 n=1000 error= +/- 3.1% (national) Undecided: 9%	37	7	33	17
SES	June 12-16 n=89 error= +/- 10.6% (province) Undecided: 13%	32	6	32	30
Ekos	June 7-9 n=2117 error= +/- 2.1% (national) Undecided: 12%	36	3	32	28
Léger	June 2-7 n=1500 error= +/- 2.5% (national) Undecided: 14%	38	3 "Other Parties"	35	21
Ipsos	June 1-3 n=1001 error= +/- 3.1% (national) Undecided: 11%	45	-	38	13
Léger	May 27-31 n=3107 error= +/- 1.8% (national) Undecided: 14%	34	2	35	26
SES	May 26-30 n=77 error= +/- 11.4% (regional) Undecided: 24%	42	4	35	20
Ekos	May 25-27 n=1306 error= +/- 2.8% (national) Undecided: n.a%	38	-	36	23
Compas	May 14-19   n=1000* error= +/- 3.1% Undecided: n.a.%	27	-	35	32
Ipsos	May 11-17   n=2000 error= +/- 2.2%  (national) Undecided: 3%	33	-	38	10
Ipsos	May 7-13   n=2002 error= +/- 2.2% Undecided: 6%	31	-	37	24
Compas	April 30-May2   n=600 error= +/- 4.1% Undecided: 13%	40 AL, SK,MB	-	33 AL, SK,MB	22 AL, SK,MB

Link to Poll	Polling Data
SES	June 22-24 n=393 error= +/- 5.0% (province) Undecided: 18%	32	4	39	25
Ekos	June 21-24 n=5254 error= +/- 1.4% (national) Undecided: n.a.%	35	5	38	21
Ipsos	June 21-23 n=2000 error= +/- 2.2% (national) Undecided: 8%	34	6	38	20
Environics	June 17-22 n=1444 error= +/- 2.5% (national) Undecided: n.a.%	35	-	40	19
Ipsos	June 18-20 n=330 error= +/- 5.5% (province) Undecided: 7%	30	6	42	20
Ipsos	June 16-17 n=1000 error= +/- 3.1% (national) Undecided: 9%	38	6	34	18
SES	June 12-16 n=259 error= +/- 6.2% (province) Undecided: 15%	40	2	36	23
Ipsos	June 11-14 n=1000 error= +/- 3.1% (national) Undecided: 9%	35	4	34	22
Ekos	June 7-9 n=2117 error= +/- 2.1% (national) Undecided: 12%	38	6	34	21
Ipsos	June 4-8 n=2000 error= +/- 2.2% (national) Undecided: 12%	35	5	40	20
Léger	June 2-7 n=1500 error= +/- 2.5% (national) Undecided: 14%	36	-	35	24
Ipsos	June 1-3 n=1001 error= +/- .31%  (national) Undecided: 11%	35	-	32	23
Léger	May 27-31 n=800 error= +/- 3.5% (provincial) Undecided: 14%	37	5	39	18
SES	May 26-30 n=254 error= +/- 6.3% (province) Undecided: 16%	43	3	33	22
Ekos	May 25-27 n=1306 error= +/- 2.8% (national) Undecided: n.a%	35	-	40	22
Compas	May 14-19   n=1000* error= +/- 3.1% Undecided: n.a.%	39	-	42	18
Ipsos	May 11-17   n=2000 error= +/- 2.2%  (national) Undecided: 3%	27	-	49	17
Ipsos	May 7-13   n=2002 error= +/- 2.2% Undecided: 6%	26	-	49	16
Compas	April 30-May2   n=600 error= +/- 4.1% Undecided: 13%	40 AL, SK,MB	-	33 AL, SK,MB	22 AL, SK,MB

Link to Poll	Polling Data
Léger	June 23-25 n=1003 error= +/- 3.8% (province) Undecided: 18%	48	11	30	5
SES	June 22-24 n=286 error= +/- 5.9% (province) Undecided: 29%	51	11	28	7
Ipsos	June 21-23 n=2000 error= +/- 2.2% (national) Undecided: 8%	48	9	29	5
Environics	June 17-22 n=1444 error= +/- 2.5% (national) Undecided: n.a.%	50	10	25	10
Ipsos	June 18-20 n=210 error= +/- 6.9% (province) Undecided: 5%	53	8	23	9
SES	June 12-16 n=173 error= +/- 7.6% (province) Undecided: 27%	52	9	30	5
Léger	June 15-16 n=1000 error= +/- 3.1% (province) Undecided: 10%	48	13	30	6
Ipsos	June 11-14 n=1000 error= +/- 3.1% (national) Undecided: 9%	48	13	22	9
Ipsos	June 11-14 n=1000 error= +/- 3.1% (national) Undecided: 9%	48	11	26	7
Ekos	June 7-9 n=2117 error= +/- 2.1% (national) Undecided: 12%	54	13	22	8
CROP	June 3-9 n=600 error= +/- 4% (province) Undecided: 12%	44	14	32	8
Léger	June 2-7 n=1500 error= +/- 2.5% (national) Undecided: 14%	46	13	31	8
Ipsos	June 1-3   n=1001 (national) error= +/- 3.1% Undecided: 11%	45	n.a	28	n.a
Léger	May 27-31 n=3107 error= +/- 1.8% (national) Undecided: 14%	48	8	33	9
SES	May 26-30 n=178 error= +/- 7.5%  (province) Undecided: 28%	43	15	30	10
Ekos	May 25-27 n=1306 error= +/- 2.8%  (national) Undecided: n.a%	45	11	36	6
CROP	May 25-27 n=845 error= +/- 2.8% (provincial) Undecided: n.a%	45	11	36	6
Compas	May 14-19   n=1000* error= +/- 3.1%  (national) Undecided: n.a.%	43	7	40	9
Ipsos	May 11-17   n=2000 (national) error= +/- 2.2% Undecided: 3%	48	10	28	6
Léger	May 12-16   n=1005 (provincial) error= +/- 3.1% Undecided: 6%	44	10	35	7
Ipsos	May 7-13   n=2002 (national) error= +/- 2.2% Undecided: 3%	46	10	31	10
Compas	April 30 - May 2 n=600 (national) error= +/- 4.1% Undecided: 13%	30	5	37	7

Link to Poll	Polling Data
SES	June 22-24 n=135 error= +/- 8.6% (region) Undecided: 15%	24	4	45	25
Ipsos	June 21-23 n=2000 error= +/- 2.2% (national) Undecided: 8%	23	7	44	26
Ipsos	June 18-20 n=1000 error= +/- 3.1% (national) Undecided: 9%	41	2	37	15
SES	June 12-16 n=86 error= +/- 10.8% (region) Undecided: 12%	41	1	34	24
Ekos	June 7-9 n=2117 error= +/- 2.1% (national) Undecided: 12%	28	4	43	25
Léger	June 2-7 n=1500 error= +/- 2.5% (national) Undecided: 14%	29	6 "Other Parties"	42	22
Ipsos	June 1-3 n=1001 error= +/- .31%  (national) Undecided: 11%	31	-	46	21
CRA	May 14-31 n=1200 error= +/- 2.8% (province) Undecided: 18%	24 Nova Scotia	-	40 Nova Scotia	27 Nova Scotia
CRA	May 14-31 n=3403 error= +/- 1.7% (regional) Undecided: 22%	27	-	45	19
Léger	May 27-31 n=3107 error= +/- 1.8% (national) Undecided: 14%	30	6	41	21
SES	May 26-30 n=74 error= +/- 12% (regional) Undecided: 25%	26	3	45	27
Ekos	May 25-27 n=1306 error= +/- 2.8% (national) Undecided: n.a%	21	-	61	18
Compas	May 14-19   n=1000* error= +/- 3.1% Undecided: n.a.%	26	-	50	20
Ipsos	May 11-17   n=2000 error= +/- 2.2%  (national) Undecided: 3%	29	-	49	20
Ipsos	May 7-13   n=2002 error= +/- 2.2% Undecided: 6%	28	-	48	20
Compas	April 30-May2   n=600 error= +/- 4.1% Undecided: 13%	36	-	41	21

Notes:

*The core national representative sample was approximately 1000 with over-samples in BC and Alberta, for a total sample of 1579.

return to top