Yeah. Fundraising means little to nothing. When a party out funds the other parties it just means they have richer donors. Big surprise that these richer donors would be funding the Cons.
It never says on Nanos polls who is paying for them. Why would Nanos go to the expense to conduct and publish a poll for free? If a polling company just puts a poll "out there" with no explanation of its funding source, we have to assume the funding source does not want to be made public. Therefore we can only deduce there is some kind of hidden agenda. It looks like they just took the last election results, cooked the numbers a bit, and put it out.
Why should anyone believe this (or any other) poll?
"any type of capitalism is inherently unstable and unsustainable, prone to boom and bust, and trending asymptotically to absolute inequality and absolute authoritarianism."
- Isn't it true of any government/economic system? It's not just important to "spread the wealth around," but it's also important to spread the power around. Both governments and corporations are controlled by people, and some of them will be "bad apples", regardless of the system of government. I don't want either the corporations nor the government getting too much power. Let them balance each other out.
I disagree. People in general do not control corporations, at least not the significant corporations. Those are controlled by the wealthy alone. Government is chosen by the people. We haven't been choosing very well but we have the power to change that. The Government belongs to the people.
Your problem is you think people consider polls as precise measures. Most don't. They are indicators not predictors. They generally confirm our general notion of what is going on based on news and our perception of the people around us. For example, if Trudeau sent the military into BC to force the pipeline through his numbers would go down. If a poll showed his numbers increasing anywhere but Alberta I'd call it BS.
I expect Trudeau's numbers to be down a bit. If people were judging Trudeau alone his numbers are down a lot, but if he is being judged in comparison to Scheer and Singh, I expect his numbers to be quite stable, maybe down a bit in comparison to Scheer but nothing dramatic. I also judge polls against themselves. Angus Reid is obviously Conservative. I don't know if Nanos leans Liberal or not because generally they do seem to have reasonable numbers and they are accurate at election time. It's true that they could be accurate the day before the election and be fudging the numbers in between but there is no indication that is happening. That doesn't mean they are spot on. There is always a margin of error. Polls indicate trends. They aren't precise measures.
Are you arguing that polls are 100% meaningless? The equivalent of throwing dice in the air?
Why should anyone believe this (or any other) poll?
Because Nanos Research (formerly known as SES Research) has been the most accurate pollster since the 2004 election? They are also the only Canadian pollster to still utilize live telephone calls. All others have swithced to much cheaper online polls or automated telephone polls, some of which were discredited in past elections.
I disagree. People in general do not control corporations, at least not the significant corporations. Those are controlled by the wealthy alone. Government is chosen by the people. We haven't been choosing very well but we have the power to change that. The Government belongs to the people.
The government is indeed elected by the people, unlike corporation leadership. But if a government controls all of the media, they can literally get away with murder and still win elections.
On a day-to-day basis, both are controlled by very small groups of people, who may be out of touch and are prone to corruption. If they are not yet wealthy, they can become wealthy. It's just a bad idea to give too much power to one tiny group of people, whether they are a mega-corporation or a mega-government.
Why should anyone believe this (or any other) poll?
Because Nanos Research (formerly known as SES Research) has been the most accurate pollster since the 2004 election? They are also the only Canadian pollster to still utilize live telephone calls. All others have swithced to much cheaper online polls or automated telephone polls, some of which were discredited in past elections.
You did not answer my question about who is paying Nanos' bills. Typical Canadian response. Ignore a question by using an irrelevant non-answer.
Nanos will not disclose who their paymasters are, so I am suspicious of their motives. Unless they clearly state who is paying for each poll, I must conclude there is an ulterior motive, no matter what the results.
Of course Nanos discloses their paymasters. All you have to do is search their website and read their reports. http://www.nanos.co/our-insight/
Many of their polls are specifically ordered by CTV. Nik Nanos regularly appears on CTV, so I would assume their weekly polls are also a part of the CTV package deal.
If you don't trust anyone, you will never be disappointed. Now you are pushing Nanos. Why are you doing that? What is your hidden agenda? What is CTV's agenda? To get the Communist Party elected?
All the pollsters claim their polls are the best A few years ago we did an analysis of the polling firms here Many had political connections At that time Angus Reid and Ipsos Reid were considered the most accurate I think based on the average of their polls in the period prior to the election But my hunch is that it changes and one pollster is close in one election but not so close in the next one
Right now for example in the Ontario election Ipsos’s most Recent poll says PCs then NDP then Libs and more recently Darrell Bricker of Ipsos says there has not been any change Ipsos also polls for the Ontario Government But who knows for sure Probably no one
Use public information to assess pollster's accuracy, not their own self-evaluations. Both Ipsos and Angus Reid used to conduct live telephone polling, but they no longer use that methodology. They both switched to online polling.
Regular people are much more likely to use a cellphone in the 21st century, rather than doing an online survey. Online polling and IVR polling has produced wildly inaccurate results in several provincial and municipal elections.
Imagine that. They are operating in the 21st century
Online polling has no "margin of error" because it is unscientific. There is no way to randomize. The polling companies get people to join their "forum" to answer survey questions of all kinds, for example, what laundry detergent you use. People join and answer the surveys because you can win prizes or air miles points. It isn't modern it's cheap.
If you want to make a statistically significant poll, you have to select your sample carefully. Thus if 55% watch TV, then 55% of your poll sample should watch TV. Your sample should match age, urban and rural, gender, identity, country of national origin, etc. demographics. This is one reason why good polling is so expensive.
Now if you had say 100,000 respondents on an online poll, and you were able to question them carefully about their personal characteristics, you might actually be able to extract a statistically significant number which could give you a reasonably accurate result. However you would probably find that many people would give up after being asked a certain number of questions. After all, the French word for "to ask" is "demander", and peppering people with questions is ultimately making demands on their dignity and privacy. So on those Internet questionaires I tend to bail out quickly.
There is no cutting corners. Polling is expensive, and we must know who is paying the bills, or we must assume a hidden agenda.
I look at the PDFs pretty carefully just like everyone else, and I do not see any mention of the paymasters. The only one which is consistent about labelling the paymaster is CROP.
Nanos is pollster of record for CTV. Like Forum Research is for the Toronto Star.
Live telephone polling has it's own weaknesses, particularly around cell phone numbers and the difficulty in geo locating them. IVR and online polling is widely used now.
As SocialJustice has said before, most polling outside an election campaign is commissioned to make a political point, or create a news story or cycle. Many pollsters are politically aligned and work for a particular party and not others. Polls aren't reliable, but they aren't useless either, they are better than guessing what the public thinks.
If you want to see the bias or expected outcome for the poll, it's best to see the precise wording of the questions and the question order. That is where things can be tweaked to get the desired outcome.
I did say that media companies usually order polls, but that does NOT mean the pollster's result are somehow aligned with the media outlet's editorial position. Con-leaning Forum Research and the Lib-leaning Star are a good example of this. CTV usually shills for the Cons, yet they sponsors Nanos, which does not tend to over-estimate Con support.
And some pollsters, such as Nanos, are very reliable. Just compare their results to actual election results.
They are not necessarily aligned with the media's outlet's editorial position, true, but they frequently are. And polls are almost always commissioned with an interest and expected outcome in mind, that leads to a story or news article (or coverage of their issue if it is an advocacy group commissioning the poll). Poll are never impartial "fact finding" missions, even vote intent polls are strategically commissioned and released.
Different polling methodologies produce different results, but again, that's not a valid reason to lump all pollsters together as whole industry in disarray.
All pollsters like to claim they are the most accurate. The problem is that many pollsters are connected at the hip to some political party which of couse they like to keep hidden, and their polls are published in an attempt to manipulate voters.
Quote:
If you want to talk about 2015, our model predicted more individual ridings correct than every single mathematical model.
Sorry bud, but you, nor any pollster, has a handle on the truth. Everything is open to interpretation. EKOS say they were the most accurate pollster if you wish to use ridings. Every election a bunch of pollsters all say they were the most accurate. As I have said before, someone forecasting the election results the day before an election means not much, because every other poll published by that pollster could be incorrect. One better, more scientific way, rather than just relying on some pollster's marketing claim, is to go back for a few weeks prior to the election, and average out a particular pollster's results. Otherwise all the other polls, except the one on the day before the election, could just be manipulation. Imagine that! And if you want to seriously learn about polling spend some time reading Nate Silver and 538.
Pre-election poll hoarding is indeed a phenomenon. A week before an election, the polling results often start to look very similar, but not always. It's important to consider the long term track record. Some pollsters have historically managed to be accurate, while everyone else missed the mark. A good example would be Nanos (formerly SES) during the 2004 election. They identified a Liberal come-back, while all other pollsters said the election was too close to call. The Liberals won by 7 points, which is NOT too close to call.
Some pollsters are okay with being complete outliers even on the day before election. Look into Mainstreet Research during the last Calgary election, or at Forum Research during the 2012 Alberta Election. They both use IVR and they both predicted the wrong winner by a wide margin, exceeding 10 percentage points. Yet their polls are still published in the mainstream media and on this forum as if they are credible.
People post here like they have never ever heard of polling bias. All you have to do is poll in an area to support the results you want to manufacture. How do we know whether pollsters poll in Conservative, Liberal or NDP areas? We don't. Who cares whether it is phone or online.
A poll must be representative of the entire voting population in order to be scientific. This includes geographical representation. Live telephone polling has been proven to produce very accurate results.
Automated phone calls (IVR) and online polling do not tend to produce representative results. In those cases, the raw data is usually "weighted" by a pollster to supposedly represent the voting population. The "weighting" is the highly questionable part of such methodology. They've been wildely wrong before, as I mentioned above.
Here is an interesting article from the NYT in September 2016. The raw data from a professionally done poll were given to different analysts, and they all came up with different results. This is because reducing the raw data to percentages of the vote for each candidate requires assumptions during the processing that can and do vary depending on who is doing the assuming.
Nate Cohn wrote:
Polling results rely as much on the judgments of pollsters as on the science of survey methodology. Two good pollsters, both looking at the same underlying data, could come up with two very different results.
How so? Because pollsters make a series of decisions when designing their survey, from determining likely voters to adjusting their respondents to match the demographics of the electorate. These decisions are hard. They usually take place behind the scenes, and they can make a huge difference.
Online polling is the most vulnerable to pollster's personal assumptions. I would say it's a huge assumption that voter population can be represented online. Does an average voter have the time to do a voluntary online poll? I don't think so. On the other hand, when you are getting a live phone call, you are less likely to just hang up. The decline rate for both online polls and automated calls is disproportionate compared to live telephone calls. Both online polls and IVR tend to over-represent the Conservatives.
Live telephone polling has been proven to produce very accurate results.
I remember the days of land-line only polling, and as I recall, you could usually predict the results of an election based on what the polls were reporting. It was only with the advent of on-line polls and cell-phones that we started to get the succession of off-base prognostications that we see now(BC 2013, USA 2016).
Here is an analysis that I had heard before for the famous Dewey Defeats Truman upset in '68...
In 1948, the pollsters used (for the first time) the telephone in order to expand the sample size for their polls. The problem was that telephones were an expensive item back then. Republicans owning one outnumbered Democrats by a margin of three to one.
1948 was likely a transitionary period in regards to communication technology(the telephone being relatively new). We're likely living in a similar period right now.
I don't think that is necessarily accurate.
And now the real poll.
Today's Nanos Research data tracker update:
Lib 39.15%
Con 30.63%
NDP 18.19%
https://tinyurl.com/y7srwskl
http://www.tooclosetocall.ca/2018/03/are-polls-aggregators-bad-for-democracy.html?m=1
Looks like the Liberals have gained some ground over the past couple of weeks.
Perhaps they have recovered from some of the damage that was inflicted earlier in the year with the India trip, etc.
It never says on Nanos polls who is paying for them. Why would Nanos go to the expense to conduct and publish a poll for free? If a polling company just puts a poll "out there" with no explanation of its funding source, we have to assume the funding source does not want to be made public. Therefore we can only deduce there is some kind of hidden agenda. It looks like they just took the last election results, cooked the numbers a bit, and put it out.
Why should anyone believe this (or any other) poll?
I disagree. People in general do not control corporations, at least not the significant corporations. Those are controlled by the wealthy alone. Government is chosen by the people. We haven't been choosing very well but we have the power to change that. The Government belongs to the people.
What's more important, apples or oranges?
Your problem is you think people consider polls as precise measures. Most don't. They are indicators not predictors. They generally confirm our general notion of what is going on based on news and our perception of the people around us. For example, if Trudeau sent the military into BC to force the pipeline through his numbers would go down. If a poll showed his numbers increasing anywhere but Alberta I'd call it BS.
I expect Trudeau's numbers to be down a bit. If people were judging Trudeau alone his numbers are down a lot, but if he is being judged in comparison to Scheer and Singh, I expect his numbers to be quite stable, maybe down a bit in comparison to Scheer but nothing dramatic. I also judge polls against themselves. Angus Reid is obviously Conservative. I don't know if Nanos leans Liberal or not because generally they do seem to have reasonable numbers and they are accurate at election time. It's true that they could be accurate the day before the election and be fudging the numbers in between but there is no indication that is happening. That doesn't mean they are spot on. There is always a margin of error. Polls indicate trends. They aren't precise measures.
Are you arguing that polls are 100% meaningless? The equivalent of throwing dice in the air?
Because Nanos Research (formerly known as SES Research) has been the most accurate pollster since the 2004 election? They are also the only Canadian pollster to still utilize live telephone calls. All others have swithced to much cheaper online polls or automated telephone polls, some of which were discredited in past elections.
The government is indeed elected by the people, unlike corporation leadership. But if a government controls all of the media, they can literally get away with murder and still win elections.
On a day-to-day basis, both are controlled by very small groups of people, who may be out of touch and are prone to corruption. If they are not yet wealthy, they can become wealthy. It's just a bad idea to give too much power to one tiny group of people, whether they are a mega-corporation or a mega-government.
You did not answer my question about who is paying Nanos' bills. Typical Canadian response. Ignore a question by using an irrelevant non-answer.
Nanos will not disclose who their paymasters are, so I am suspicious of their motives. Unless they clearly state who is paying for each poll, I must conclude there is an ulterior motive, no matter what the results.
Of course Nanos discloses their paymasters. All you have to do is search their website and read their reports. http://www.nanos.co/our-insight/
Many of their polls are specifically ordered by CTV. Nik Nanos regularly appears on CTV, so I would assume their weekly polls are also a part of the CTV package deal.
There has to be a conspiracy, like the Russia meddling in the US election.
Russia beside pushing false info, also apparently used message boards to influence people's take on the politics and politicians.
It can be organised, as in the Russia thing, or it could also be individuals learning from the Russians tactics and freelancing.
Where is Babble's McCarthy
How quickly we forget!
https://www.theglobeandmail.com/news/politics/survey-says-the-future-of-polling-is-hard-topredict/article26042112/
Lumping all the pollsters together is the same as lumping all politicians together. It's just carelessness and lack of nuance.
If you don't trust anyone, you will never be disappointed. Now you are pushing Nanos. Why are you doing that? What is your hidden agenda? What is CTV's agenda? To get the Communist Party elected?
Trust the most reliable sources available, based on past record and methodology. Otherwise you're in your own bubble.
As for CTV, I think their Con bias is very obvious. Anyone still remembers the hatchet job they did on Stephane Dion in 2008?
The Cons also get their Senators directly from CTV.
Nevertheless, CTV ditched Ipsos and hired the most accurate pollster instead.
Maybe that's why Darrell Bricker blocked me on twitter.
Who are you? Nik Nanos?
All the pollsters claim their polls are the best A few years ago we did an analysis of the polling firms here Many had political connections At that time Angus Reid and Ipsos Reid were considered the most accurate I think based on the average of their polls in the period prior to the election But my hunch is that it changes and one pollster is close in one election but not so close in the next one
Right now for example in the Ontario election Ipsos’s most Recent poll says PCs then NDP then Libs and more recently Darrell Bricker of Ipsos says there has not been any change Ipsos also polls for the Ontario Government But who knows for sure Probably no one
Use public information to assess pollster's accuracy, not their own self-evaluations. Both Ipsos and Angus Reid used to conduct live telephone polling, but they no longer use that methodology. They both switched to online polling.
Imagine that. They are operating in the 21st century
Regular people are much more likely to use a cellphone in the 21st century, rather than doing an online survey. Online polling and IVR polling has produced wildly inaccurate results in several provincial and municipal elections.
Online polling has no "margin of error" because it is unscientific. There is no way to randomize. The polling companies get people to join their "forum" to answer survey questions of all kinds, for example, what laundry detergent you use. People join and answer the surveys because you can win prizes or air miles points. It isn't modern it's cheap.
If you want to make a statistically significant poll, you have to select your sample carefully. Thus if 55% watch TV, then 55% of your poll sample should watch TV. Your sample should match age, urban and rural, gender, identity, country of national origin, etc. demographics. This is one reason why good polling is so expensive.
Now if you had say 100,000 respondents on an online poll, and you were able to question them carefully about their personal characteristics, you might actually be able to extract a statistically significant number which could give you a reasonably accurate result. However you would probably find that many people would give up after being asked a certain number of questions. After all, the French word for "to ask" is "demander", and peppering people with questions is ultimately making demands on their dignity and privacy. So on those Internet questionaires I tend to bail out quickly.
There is no cutting corners. Polling is expensive, and we must know who is paying the bills, or we must assume a hidden agenda.
We do know who pays the bills. It's almost always some media company.
I look at the PDFs pretty carefully just like everyone else, and I do not see any mention of the paymasters. The only one which is consistent about labelling the paymaster is CROP.
There is usually a media outlet which has the right to release a poll first. They are the sponsor of that poll. It's not always in the report.
Nanos is pollster of record for CTV. Like Forum Research is for the Toronto Star.
Live telephone polling has it's own weaknesses, particularly around cell phone numbers and the difficulty in geo locating them. IVR and online polling is widely used now.
As SocialJustice has said before, most polling outside an election campaign is commissioned to make a political point, or create a news story or cycle. Many pollsters are politically aligned and work for a particular party and not others. Polls aren't reliable, but they aren't useless either, they are better than guessing what the public thinks.
If you want to see the bias or expected outcome for the poll, it's best to see the precise wording of the questions and the question order. That is where things can be tweaked to get the desired outcome.
https://www.youtube.com/watch?v=884qXhIqsKU
I did say that media companies usually order polls, but that does NOT mean the pollster's result are somehow aligned with the media outlet's editorial position. Con-leaning Forum Research and the Lib-leaning Star are a good example of this. CTV usually shills for the Cons, yet they sponsors Nanos, which does not tend to over-estimate Con support.
And some pollsters, such as Nanos, are very reliable. Just compare their results to actual election results.
http://www.cbc.ca/news/politics/grenier-new-bloc-1.4644751
They are not necessarily aligned with the media's outlet's editorial position, true, but they frequently are. And polls are almost always commissioned with an interest and expected outcome in mind, that leads to a story or news article (or coverage of their issue if it is an advocacy group commissioning the poll). Poll are never impartial "fact finding" missions, even vote intent polls are strategically commissioned and released.
Ipsos: L 36 C 35 NDP 20
https://globalnews.ca/news/4186049/liberals-gain-more-public-approval-as-memory-of-india-trip-fades-poll/
That new Ipsos poll is interesting.
It shows the Liberals up 5 points, and the Conservatives and NDP each down 3 points.
Ipsos usually leans Conservative, so the fact that the Liberals are up is probably a good sign for the LPC.
I Dunno. It sure looks like the polling industry is in complete disarray.
https://www.ipsos.com/en-ca/news-polls/copyright-ownership-Canadian-political-polls
dp
https://mobile.twitter.com/EarlWashburn/status/988255481436168192
Different polling methodologies produce different results, but again, that's not a valid reason to lump all pollsters together as whole industry in disarray.
Nanos Research poll released on 2018-05-08:
Liberals 38, Conservatives 33, NDP 18, Green 5
http://www.nanos.co/wp-content/uploads/2018/05/Political-Package-2018-05-04.pdf
All pollsters like to claim they are the most accurate. The problem is that many pollsters are connected at the hip to some political party which of couse they like to keep hidden, and their polls are published in an attempt to manipulate voters.
Haven't we been through this before? Claims are irrelevant, while track record and methodology speak for themselves.
The truth is NOT in the eye of the beholder.
Sorry bud, but you, nor any pollster, has a handle on the truth. Everything is open to interpretation. EKOS say they were the most accurate pollster if you wish to use ridings. Every election a bunch of pollsters all say they were the most accurate. As I have said before, someone forecasting the election results the day before an election means not much, because every other poll published by that pollster could be incorrect. One better, more scientific way, rather than just relying on some pollster's marketing claim, is to go back for a few weeks prior to the election, and average out a particular pollster's results. Otherwise all the other polls, except the one on the day before the election, could just be manipulation. Imagine that! And if you want to seriously learn about polling spend some time reading Nate Silver and 538.
Pre-election poll hoarding is indeed a phenomenon. A week before an election, the polling results often start to look very similar, but not always. It's important to consider the long term track record. Some pollsters have historically managed to be accurate, while everyone else missed the mark. A good example would be Nanos (formerly SES) during the 2004 election. They identified a Liberal come-back, while all other pollsters said the election was too close to call. The Liberals won by 7 points, which is NOT too close to call.
Some pollsters are okay with being complete outliers even on the day before election. Look into Mainstreet Research during the last Calgary election, or at Forum Research during the 2012 Alberta Election. They both use IVR and they both predicted the wrong winner by a wide margin, exceeding 10 percentage points. Yet their polls are still published in the mainstream media and on this forum as if they are credible.
Ipsos Moves to Protect Copyright of its Canadian Political Polls
https://www.ipsos.com/en-ca/news-polls/copyright-ownership-Canadian-poli...
Ipsos went downhill since they switched to online "polling". Even CTV won't hire them anymore.
People post here like they have never ever heard of polling bias. All you have to do is poll in an area to support the results you want to manufacture. How do we know whether pollsters poll in Conservative, Liberal or NDP areas? We don't. Who cares whether it is phone or online.
A poll must be representative of the entire voting population in order to be scientific. This includes geographical representation. Live telephone polling has been proven to produce very accurate results.
Automated phone calls (IVR) and online polling do not tend to produce representative results. In those cases, the raw data is usually "weighted" by a pollster to supposedly represent the voting population. The "weighting" is the highly questionable part of such methodology. They've been wildely wrong before, as I mentioned above.
Here is an interesting article from the NYT in September 2016. The raw data from a professionally done poll were given to different analysts, and they all came up with different results. This is because reducing the raw data to percentages of the vote for each candidate requires assumptions during the processing that can and do vary depending on who is doing the assuming.
Online polling is the most vulnerable to pollster's personal assumptions. I would say it's a huge assumption that voter population can be represented online. Does an average voter have the time to do a voluntary online poll? I don't think so. On the other hand, when you are getting a live phone call, you are less likely to just hang up. The decline rate for both online polls and automated calls is disproportionate compared to live telephone calls. Both online polls and IVR tend to over-represent the Conservatives.
I remember the days of land-line only polling, and as I recall, you could usually predict the results of an election based on what the polls were reporting. It was only with the advent of on-line polls and cell-phones that we started to get the succession of off-base prognostications that we see now(BC 2013, USA 2016).
Here is an analysis that I had heard before for the famous Dewey Defeats Truman upset in '68...
In 1948, the pollsters used (for the first time) the telephone in order to expand the sample size for their polls. The problem was that telephones were an expensive item back then. Republicans owning one outnumbered Democrats by a margin of three to one.
1948 was likely a transitionary period in regards to communication technology(the telephone being relatively new). We're likely living in a similar period right now.
https://tinyurl.com/ya2estzu
Pages