Talk:Opinion polling in the Canadian federal election, 2006

From Wikipedia, the free encyclopedia

Contents 1 Seat predictions 2 Unified graph? 3 Straw poll on seat projections 3.1 Voting 3.1.1 Comments 4 Seat Predictions Using Bad Data 5 Liberals and Conservatives in a dead heat 6 Charts on the left hand side 7 Last update time for polls 8 switch columns 9 Ekos one-day polling 10 Remove Seat Projections? 11 Greens 12 Relevance? 13 Combined Polls Graph 14 may 19th Compass Poll 15 Please use shorter URLs without commas for this article. 16 Majority or Minority? 17 Polling Company Bias - Strategic Counsel vs SES 18 Manipulation of the Order of Polls? 19 Recent Strategic Counsel Oddity

[edit] Seat predictions

This probably should not be included, I actually run a website with seat predictions and using the H&K predictor provides very unreliable results due to it simply assuming that all factors are equal to the last election. For instance, the predictions posted on this page always shown an independent (Chuck Cadman) winning which is obviously impossible as Mr. Cadman has passed away. It also shows one instance in which there are two independents: Mr. Cadman and one of either Jim Pankiw or Grant Devine both of which ran as independents in 2004 but are not running this time. - Jord 22:19, 12 December 2005 (UTC)

• I agree ... seems pretty hokum to me. Nfitz 23:06, 12 December 2005 (UTC)

• I absolutely say keep it. Dispite its flaws, the seat projection gives a much more relaistic view of party standings than the opinions polls. For example, it reflects the fcat that 40% of the popular vote generaly translates to a majority win. Just put some warnings at the top of the section. -Arctic.gnome 19:05, 13 December 2005 (UTC)
  • Well if you want to keep it, it would probably make more sense to use seat projections with a proper analysis, perhaps that on Democratic Space or even mine (I am trying to be modest). What he have got here is wholly unencyclopedic. - Jord 19:54, 13 December 2005 (UTC)
    • I somewhat agree with substituting the current predictor with another more accurate one (if such a thing exists), but I would support including any one (present company included) that is more accurate and has consensus support here. In any event, a qualifier must be added. E Pluribus Anthony 20:04, 13 December 2005 (UTC)
      • Well my preference would be to include none at all because they are rife with error, at least those done by myself and democratic space due allow for recognition of change between 04 and 06 in different races. They give slightly different results because of a slightly different methodology but in both cases they are far more accurate than just plugging raw data into the predictor. I am sure that rant went from both sides and back again so let me simpify. The results of plugging numbers into the H&K predictor should not be included under any circumstances. I could live with the more well thought projections being included but if it was put to a vote, I would oppose it. This is coming from one of the people that do those predictions. They are interesting but are not scientific, done by amateurs and do not belong on Wikipedia. The best solution would be to include a note saying that popular vote is not necessarily indicative of seats to be won, several interesting projections of what these results would mean can be found at and include some external links - Jord 20:16, 13 December 2005 (UTC)
        • I agree that it is not perfect - no seat predictor is. If someone finds one that is definately better (and can explain why), then we should definately use that instead of H&P. The main reason why I think this section should be included is that most people don't care about voting percentages - they care about how many seats a party gets. While the H&P won't be 100% accurate, it does indicate if a party probably gain/loose seats (or stay roughly the same), and whether or not the winning party will have a majority. There is also the point that in Canada (like the UK), having more votes than another party does not always result in more seats. Specifially, if the Liberals and Conservatives tie on votes, the Conservatives get mroe seats. Thus voting percentages alone do not tell the true story. Tompw 23:16, 13 December 2005 (UTC)

[edit] Unified graph?

Hi! Would there be any value in having a single unified graph depicting poll trends? While general trends can be surmised from the many current graphs and polls/firms, a single one would allow for a ready comparison throughout (and I don't think an erroneous one, given the multiple polls/firms). In case of multiple polls on any given day, I'd use the poll figures with the largest sample size. Thoughts? E Pluribus Anthony 10:20, 13 December 2005 (UTC)

Sounds a good plan to me... though I would say with multiple polls on a given day, why not use the average percentages, possibly weighted by number of people polled. Also, don't ask whether to do it - be bold and do it! Tompw 23:19, 13 December 2005 (UTC)

Great! A unified graph perhaps weighted by polling number should be fine (and hopefully not border on original research); my initial suggestion was for simplicity ... as the discussion regarding predictors may reveal, simplicity is not always preferred. :) Give me a few days. As well, I didn't want to undertake what could be a time-consuming venture if there wasn't support for it or if it might be erroneous or not helpful. To paraphrase (regarding typefaces and editing): I'm fine with being bold, but I also don't want to be a 'spastic' colon ... :) Thanks again! E Pluribus Anthony 03:12, 14 December 2005 (UTC)

[edit] Straw poll on seat projections

There is some debate as to whether or not seat projections should be included on this page. Currently, a Wikipedian plugs the numbers from a given poll into the Hill & Knowlton election predictor which simply applies the national margin of change in the given poll to each riding. For example: the Liberals received 36.7% of the vote in 2004, if a poll shows them at 38% then it will increase every Liberal candidate's share of the vote by 1.3%.

The fundamental flaws with such a model are that this does not account for the many ridings in which there was a predominant issue in the last campaign that does not exist in this one (or vice versa) or the presence of a strong candidate in the last campaign that does not exist in this one (or vice versa).

An additional flaw is that, beyond the 4 main parties, all other votes are bunched together as "Other". If the Green Party were to get a surge in the polls, the results might show a bunch of "Others" getting elected but, in fact, what may be projected, even on this flawed system, is that the combined vote of the Greens and other minor parties and independents in a riding was greater than another party but no one of those other parties in fact won.

Finally, there is a serious problem in its projection of three particular races: Surrey North, Souris—Moose Mountain and Saskatoon—Humboldt -- all had high profile independents, none of which are reoffering. It is not uncommon for a projection with a high Green vote to show these independents winning because their vote totals + those of other parties are all counted as "Other".

Several websites, particularly democraticSPACE.com and jord.ca, do similar projections to plugging the data into the H&K projector but do the swings on a regional basis and take into account the different scenarios of the individual races.

The above seat projections are both flawed in different ways. Raw data from the Hill & Knowlton projector ignores regional and local factors. The two websites are prone to human error. Neither of them are scientific. Nonetheless, some Wikipedians argue that the projections should be included because popular vote in a poll is not a good indicator of what the result will be, especially in a close race. For instance, in the 1979 election Joe Clark won, and nearly got a majority government, despite losing the popular vote by over 4%.

[edit] Voting

The object of this straw poll is to try to ascertain where the consensus leans in terms of what should be included.

• Option A is using numbers obtained from the H&K predictor
• Option B is using numbers from either or both of the websites (or other sites doing similar work)
• Option C is to not include seat projections unless they are provided by the pollster (some pollsters will do this as the campaign progresses if we assume they'll operate in the same fashion as previous campaigns

Please list your first then second choice in the table.

User	1st	2nd
Jord	C	B
arctic gnome	B	A
E Pluribus Anthony	B *	C
tompw	B *	A
nfitz	C	B *

[edit] Comments

• * only if consensus can agree on which, with proviso; if not ... E Pluribus Anthony 03:17, 14 December 2005 (UTC)

Same here Tompw 15:31, 14 December 2005 (UTC)

It looks like everyone who has commented here before has voted, and the most popular option seems to be to use the numbers from one of those other sites. So can anyone recomend which to use? -Arctic.gnome 03:47, 21 December 2005 (UTC)

Personaly, I like democraticspace.com, but the other one is okay. Alternatively we could use something like electionprediction.org, which only reports "safe" seats, and calls the rest "too close", but I think that kind of defeats the point of seat projections. -Arctic.gnome 04:06, 21 December 2005 (UTC)

Hi! I, too, prefer democraticspace.com (or/then jord.ca), then (distantly) the one currently in place. One of the 'features' of Canadian electoral politics is regionalism, or moreso in recent times, and predictions better embracing that may be less erroneous than currently. Anything else defeats the purpose, in which case we should only provide links to all of them (methinks). E Pluribus Anthony 11:18, 21 December 2005 (UTC)

Indeed, both of the above sites use regional data as the rightly should, a national results of 36-29 (as seen in the last election) has proved to give a completely different result in terms of seats on several occassions when polls have had the same numbers as a result of the regional numbers changing radically but the national average of such changes staying the same. Electionprediction.org does not use polls at all but is based on the guts of its readers and, in the end, its editors. It is even less scientific than the seat projections, though conceivably more accurate. I say that as a person who works on electionprediction.org as well ;) Jord 16:56, 21 December 2005 (UTC)

• So despite the fact that there is wide consensus to drop the unreliable seat projections, they are still there. Does anyone want to go ahead and put up the democraticspace ones? - Jord 21:16, 6 January 2006 (UTC)

[edit] Seat Predictions Using Bad Data

I just deleted the December 9, Strategic Counsel predictions. There was no December 9 result! Also, I've fixed some of the numbers in the polling table for other Strategic Counsel results, so if people have used the results in the table for the seat predicitons, they will have to fix them! Not sure where people are getting these bad Strategic Counsel results from ... but I wish people would put in links when they add them ... all the results are on the Strategic Counsel website. Nfitz 18:10, 14 December 2005 (UTC)

I changed the January 1st Strategic Counsel results to December 31st ... as there were no polls on January 1st. But there's still an Ipsos-Reid result dated December 31st - however there is no December 31st poll reported in the table above! Not sure if this is legit or not ... Nfitz 00:23, 5 January 2006 (UTC)

[edit] Liberals and Conservatives in a dead heat

I changed the intro to the topic to reflect that the Liberal Party of Canada and the Conservative Party of Canada are in a dead heat in the Opinion polling. SFrank85 02:13, 3 January 2006 (UTC)

I just updated the intro to reflect the fact the the conservative party now appears to have the lead

I edited slighly. The Conservatives appear to have a lead, but not a substantial lead. They are still statistically tied accoring to some polling! Though I think it's fair to say they they are in the lead. Nfitz 20:28, 10 January 2006 (UTC)

[edit] Charts on the left hand side

I removed the charts on the left as they havn't been updated since the start of the election. If someone would care to update them and put them back that would be cool.

Sorry about that. I will update them tonight! -- Earl Andrew - talk 04:42, 10 January 2006 (UTC)

[edit] Last update time for polls

Do we really need this - if someone wants to know, they can always check the history. And personally, I've been forgetting to update this for weeks. Though I did movfe it to the Polls section, so it's actually there to update! Nfitz 20:31, 10 January 2006 (UTC)

I removed this, as no one commented, and it is normally out of date! Nfitz 13:56, 17 January 2006 (UTC)

I agree: the time is unnecessary; besides, Wikipedia isn't a news service.

Relatedly, I think initial (single) indications/wikifying of the year are necessary, so a visitor doesn't potentially have to question in which year one particular poll amongst many took place in ... as I did when glancing through the plethora of entries over three years. It's also not excessive.

As well (once the election is fait accompli), I'm going go through and nix redundant wikified dates in the tables both in this article and the election article. E Pluribus Anthony | talk | 14:08, 17 January 2006 (UTC)\\

The year would be useful once this is all over. Perhaps in the interim, just at the last 2005 poll and first 2006 poll (and the 2004 election) would be useful, as it saves two edits every time you add a new poll (adding the year to the new one, and deleting it from the previous). If others want to add, great, but ... Nfitz 20:08, 17 January 2006 (UTC)

Agreed: those adding new poll entries should be cognisant, though, of adding the year to the current entry and nixing it from the previous one. E Pluribus Anthony | talk | 20:35, 17 January 2006 (UTC)

[edit] switch columns

Somebody should switch the Cons and Lib columns, now that Cons is leading. That order would amke more sense.--Sonjaaa 04:26, 11 January 2006 (UTC)

The Liberals have still led in more polls. They should probably be arranged this way up until the election as it reflects the results of the previous election. (The BQ comes last because they're a regional party with no influence outside of Quebec.) - Cuivienen 05:05, 11 January 2006 (UTC)

The Conservatives are now clearly leading, but even then, I do agree that the results should be based on the previous election. Snickerdo 05:29, 11 January 2006 (UTC)

It isn't worth switching every time the polls predict a new winner. The positions reflect how the parties did is the last election. -Arctic.gnome 05:24, 11 January 2006 (UTC)

The lead has switched a few times since the polling started. Seems a lot of work, to change the columns around each time the lead changes. If anyone really things it should be based on the final standing, than the last election standing, they could switch it around after the election. Nfitz 15:07, 11 January 2006 (UTC)

I'm new here, but I agree they should be left alone. The order that the parties appear should be picked at the outset and left there. To order it based on standing creates unnecessary work (when ranking shifts) and unnecessary argument (what about statistical ties, or when multiple polls have different rankings?). - Otter Escaping North 19:27, 13 January, 2006 (UTC)

As the incumbent party, regardless of polling figures, the Liberal values should be in the first column until and unless the governing party changes. E Pluribus Anthony | talk | 20:44, 17 January 2006 (UTC)

[edit] Ekos one-day polling

Ekos is now doing one-day polling that they themselves note "we do not consider a single night of tracking to be scientifically credible. These nightly releases are intended only to respond to curiosity about day-to-day movements." I'm pondering if we want to add these or not ... Nfitz 15:16, 11 January 2006 (UTC)

• I'm starting to lean towards adding (perhaps with a note) - they are showing something, and they are talking to more people per day that SES. Anyone else got any thoughts? Nfitz 19:02, 11 January 2006 (UTC)

If something as unscientific as the seat projections deserve to be on the page, I certainly think one day polling should qualify. - Otter Escaping North 19:30, 11 January 2006 (UTC)

[edit] Remove Seat Projections?

Should we be looking at removing the seat projections? I don't know why all the ones we are showing come from a single generator ... shouldn't we be just putting those in reported in the media, rather than generating our own Nfitz 19:44, 13 January 2006 (UTC)

Nfitz, see above - it was agreed that they should be removed and replaced with ones from an external source but no one has taken the initiative to do so, if you are up to, please do!! - Jord 19:56, 13 January 2006 (UTC)

Ah, forgot about that! I was gone for Christmas when it was all concluded. Maybe on Sunday .... any takers before then? Nfitz 23:53, 13 January 2006 (UTC)

I removed them and replaced them with all of the final projections I could find on the web (and all of those which Andrew Coyne could find as well) and have archived what was there at Opinion polling in the Canadian federal election, 2006/Old projections in case someone wants to un do it. - Jord 17:13, 25 January 2006 (UTC)

[edit] Greens

The Green party logo looks kind of funny. Why is it grey and not yellow? --JGGardiner 02:09, 12 January 2006 (UTC)

That is the actual shade of green they use. Snickerdo 18:03, 13 January 2006 (UTC)

Oh thanks. I've never seen the grey version. I know the yellow one, which is their offical logo registered with Elections Canada and I've seen it in green also but not in grey. Or whatever that colour is. Hard to tell on such a small picture. --JGGardiner 20:10, 14 January 2006 (UTC)

They used that logo last election, although they may have switched back to the yellow one this election. -- Earl Andrew - talk 01:00, 15 January 2006 (UTC)

[edit] Relevance?

Will this article still be considered relevant enough to be on Wikipedia after the election? I'm not saying it should be done either way, I just don't commonly see articles with such a specific nature in wikipedia - usually it would be "opinion polling in canadian elections" - and it would include provincial, federal, and (or so its looking) senatorial elections.--68.73.55.59 19:01, 15 January 2006 (UTC)

• Sure. It will be of historical relevance to see how the polls changed during the campaign. -Nichlemn 10:31, 16 January 2006 (UTC).

[edit] Combined Polls Graph

I have created a combined polls graph if anyone is interested in posting it... not sure I can do it myself.

it's an excel plot with a coloured "band" showing encompassing the range of poll results.

Excellent graph! I was going to put one up looong ago, but alas ... :) One query: is it possible to extend the graph further back to provide a better indication of long-term trends? E Pluribus Anthony | talk | 23:14, 15 January 2006 (UTC)

Thanks for the encouragement =) I'll give the extansion a go.

I am usually opposed to such graphs, but the way you made it makes it look good. I guess there will be no need to update my graphs? -- Earl Andrew - talk 02:50, 16 January 2006 (UTC)

I would say it was worth updating the the graphs by pollster, because it's interestign to compare results between them. (For example SES on average give Liberals 3% more than Statregic Council). That said, it's less time critical, now we have an excellent unified graph. Tompw 11:08, 16 January 2006 (UTC)

It would probably make sense (IMO) to update all of the graphs just prior to election day (a snapshot, if you will), and then again after the election (to provide a complete picture). I don't think there's a need – unless someone's compelled to! – to update them with the numerous polls being conducted on a daily basis before the election. It may also be prudent to do so beforehand if they indicate a change in support ... and at this point, this isn't wholly apparent (yet?). But whatever works! :) E Pluribus Anthony | talk | 11:17, 16 January 2006 (UTC)

Very good!

Just a note, there will be no polling on the weekend, as it is against the law (48 hours) -- Earl Andrew - talk 02:35, 17 January 2006 (UTC)

I updated the graph, as well as tried to make it more accurate and round the corners =) I'll update it again once or twice more before the election - galneweinhaw (how do I add my handle here like you guys??)--galneweinhaw 17:11, 18 January 2006 (UTC) =D

To add your handle, click the formatting button that looks like a signature above the text window when you edit. Alternatively, put four tildas: ~~~~ thirty-seven 07:18, 18 January 2006 (UTC)

[edit] may 19th Compass Poll

The Liberal and Conservative Values look like they could be reversed? Will try to find a the actual poll. -galneweinhaw

Doing a Google search I came across this. The fourth page has numbers that are on the article right now, not sure if this is the same poll. Qutezuce 03:50, 16 January 2006 (UTC)

Ah, great info! I added the link, and fixed the date. Very small sample ... and perhaps why no one is using Compass these days! :-) Nfitz 20:53, 16 January 2006 (UTC)

Looking at the other questions in the poll, it may also have been a push poll, or perhaps just a very short term reaction to the events. - galneweinhaw

[edit] Please use shorter URLs without commas for this article.

The current URL for this article: http://en.wikipedia.org/wiki/Opinion_polling_in_the_Canadian_federal_election,_2006

is difficult to pass on. URLs with commas are not often made correctly clickable in forums, and in email. I am talking about those that make all URLs clickable.

Also, in forums the URL is too wide for many people using 15-inch notebook PCs. Especially those people who use a larger text size to read more easily. URLs without dashes will not wrap to 2 lines in many forums. So the URL stays on one line, and that forces the forum message text to extend too far to the right.

Most forums have one or more side columns. So that side column width eats into the forum message space width. People then have to scroll each and every line of message text in that forum thread. It effectively is a thread killer. The moderators often don't have the time to put the URL into an HTML link where one does not see the URL.

So many people stop passing on the URL for this article. --Timeshifter 13:30, 16 January 2006 (UTC)

You can always use the alternate name: http://en.wikipedia.org/wiki/2006_Canadian_federal_election_opinion_polls it's still long, but is under 80 characters and has no commas, so should survive well as a link. Looks fine on my 15" laptop, only takes up one-third the page width in Outlook. Nfitz 20:58, 16 January 2006 (UTC)

[edit] Majority or Minority?

I'm not quite sure how this works - but if the conservatives win 154 seats, then do they hold a minority or a majority? Mainly, I'm asking who's going to be the next speaker. If the conservatives win, do they nominate one of their own to be the speaker, or would it remain peter milliken's? If it remains peter milliken's, they have a majority, if not, then 154 gives them a minority.--68.73.53.71 16:42, 16 January 2006 (UTC)

A new one is chosen with each parliment. Now, the Conservatives could try to get Peter Milliken to retake the position, but if it really did come out to 154 Conservatives, then he'd probably reject and force the Conservatives to promote one of their own. -- VederJuda 17:01, 16 January 2006 (UTC)

With 308 seats in the half, they don't have a majority. So by definition, I'd say they would have a minority. Here's an interesting case. In the 1921 election the Liberals won 118 of the 235 seats, which is 50.2% of the seats. However it is referred to as a minority government. There's not that much detail listed, but I'm guessing that after the speaker was elected, that the Liberals only had 117 of the 234 votes (given that by tradition the Speaker doesn't necessarily vote with the government). In the 2006 election, they would need 155 seats to give them a majority of the 154 of 307 voting seats. Nfitz 21:14, 16 January 2006 (UTC)

No that's not the case, it became a minority at one point due to by-elections but then returned to a majority due to subsequent by-elections. - Jord 21:51, 16 January 2006 (UTC)

Someone in the know should tidy up those articles ... Nfitz 22:41, 16 January 2006 (UTC)

After the 2003 Assembly election in Wales, the Labour party faced the same problem of having half the seats. The final result was Lab 30 Plaid 12 Cons 9 LD 6 Oth 1. The Assembly settled this by reelecting the previous speaker, a Plaid Cymru member. This gave Labour a working majority of one seat, which was later lost when a Labour member was expelled. See Peter Law and Welsh Assembly election, 2003 for more details. In the 2000 U.S. Senate election, The Republicans and the Democrats were tied with 50 seats apiece. Bill Clinton was still President when the Senate first met. That gave the Democrats a working majority of 50 seats + the Vice President (Al Gore), who is the Senate's Presiding Officer. George W. Bush took office later in Jan. 2001. This gave the Republicans a working majority of 50 seats + the Vice President (Dick Cheney). This majority lasted until June 2001, when a Republican Senator became an Independent, giving the Democrats a working majority of 51 (50 Dems + 1 Ind). This proves that things like what Harper could face have happened before. - Thanks, Hoshie | 08:30, 19 January 2006 (UTC)

[edit] Polling Company Bias - Strategic Counsel vs SES

Looking at polls from Dec 1 to Jan 16 (when SES and Strategic Counsel polled on the same date), SES showed an average of 3% lead for the Liberals, whereas Strategic Counsel shows an average lead of 0.3% for the Conservatives. What is the source of this systemic bias? -AaronKreider-

-The Strategic council is the biased poll - SES was almost exactly right last election. The Strategic council is headed up by Allen Gregg as its lead pollster. He was an insider in the Mulroney conservative government, and worked on Kim Campbells re election campaign. He was actually the one who came up with the lets attack Jean Chretiens face add. --24.222.65.32 23:36, 20 January 2006 (UTC)

As a result of compiling the data in a spreadsheet, I've noticed this as well. For the days since 1st December that both Strategic Counsel and SES have issued polls, Strategic Counsel have, on average, given the Liberals 3.3% less than SES; the Torys 0.1% more; NDP 0.9% more; the Bloc 1.3% more and "other" 1% more. So, there is a definate bias going on with the Liberals... either SES are pro Liberal, or Strategic Counsel are anti Liberal. Tompw 00:14, 21 January 2006 (UTC)

Original research notwithstanding, regardless of perceived bias (and I don't deny that it may exist), polls should be rendered/presented and ordered neutrally – chronologically and, when that is unworkable, alphabetically. Any interpretations should be cited, verifiable, etc. before ending up on any Wp article. E Pluribus Anthony | talk | 00:20, 21 January 2006 (UTC)

Once again the SES poll is almost perfectly accurate.

[edit] Manipulation of the Order of Polls?

I noticed that an IP has been rearranging polls of the same date to have a specific result on top. It's no a big deal, just annoying me, mostly cus it keeps screwing up the notes. --galneweinhaw 06:23, 19 January 2006 (UTC)

It is getting annoying. Particularily as they are doing it without changing the notes. We've had edits in the last 24-hours by 209.217.84.140 and 206.191.56.151 - neither of who have any other Wikipedia edits except on this page! Both seem to traceback to magma.ca Nfitz 17:47, 19 January 2006 (UTC)

To maintain impartiality, I'd arrange them according to date (as per prior discussion) and on the same date by release time (if available). If release times are unavailable, they should be arranged alphabetically by firm. E Pluribus Anthony | talk | 17:49, 19 January 2006 (UTC)

Over the course of a day, that's what tends to happen. New ones at the top. In any given day, Ekos goes first think in the morning, SES comes out about 2 pm, and SC comes out around 10-11 pm. The others, it's hard to tell ... sometimes we don't hear about it until days later. Nfitz 17:53, 19 January 2006 (UTC)

I know; I only reiterate this for those who may try to use Wp for impolitic or do so contrary to what should be a neutral (chronological) ordering scheme. :) E Pluribus Anthony | talk | 17:56, 19 January 2006 (UTC)

I've reordered the last few days by the general release time. Hmm, seems there's been some kind of push to put the SES results at the top of the page. Both IPs seems to be the same individual, as they also keep rounding the latest SES result to 0 decimal places, despite SES uncharacteristically doing them as 1 place. Oh well, they'll be happy soon, as the daily SES release will be out shortly! Nfitz 18:11, 19 January 2006 (UTC)

[edit] Recent Strategic Counsel Oddity

I added a note yesterday about the recent Strategic Counsel polls that both CTV and the Globe and Mail have been playing down in their news. Feel free to remove it whenever appropriate.--galneweinhaw 20:42, 19 January 2006 (UTC)

I didn't get the feel that Globe and Mail were playing it down in the newspaper this morning. I didn't watch CTV; what were they saying? The numbers do seem quite different that what's coming out of SES and Ekos; mostly in Quebec and the West, by the looks of it. Nfitz 21:30, 19 January 2006 (UTC)

For example, on CTV's online polling thingy (yesterday anyway) here: http://www.ctv.ca/mini/election2006/static/interactives/poll_trkr/index.html their statement for the 18 Jan poll is pretty funny when you look at their graph =) On the news they have been mentioning it only briefly (where you'd think these numbers would be the leading story). They didn't even discuss the results in an interview with their pollster! --galneweinhaw 23:03, 19 January 2006 (UTC)

As for the G&M, When you have numbers showing one party at the highest they've ever been and the contender at the lowest they've ever been, That should be front page stuff. I find it quite odd, especially since they've had similar result 2 days in a row. I guess we'll see tonight. - --galneweinhaw 23:05, 19 January 2006 (UTC)

Looks like the results are more in line with the other pollsters now ... you know how they say the results are good within a few percent, 19 times out of 20 ... well I think those were our 1 out of 20 results ... I guess that days's polling has fallen off the poll now. I'll remove that note, as I see it's served it's purpose.Nfitz 13:46, 20 January 2006 (UTC)