Archive for June, 2009

BREAKING NEWS: This morning, right-wing forces staged a coup d’état in Honduras

Paulitics breaking news 3

Honduras protestersReports are coming in from multiple press outlets that right-wing forces have staged a military coup d’état against the democratically-elected leftist president of Honduras, Jose Manuel Zelaya.

After Zelaya’s election in 2005, Honduras joined the Bolivarian Alternative for the Americas, a progressive international cooperation organization which is also known by its Spanish acronym “ALBA”.  ALBA (which also means “dawn” in Spanish) is an international cooperation agreement Hondurasdesigned as an alternative to the U.S.-backed neo-liberal Free Trade Area of the Americas (FTAA).  It was largely because of opposition to the FTAA from countries like Honduras, Venezuela, Bolivia, Ecuador, Dominica, Nicaragua and Cuba (all of which are now ALBA members) that the U.S.-backed FTAA failed.

The coup comes as Zelaya was expected to win a popular referrendum allowing him to stand for re-election in 2010 by removing term limits for the office of the President.  The Honduran Supreme Court and military have both declared the vote by the people of Honduras to be unconstitutional.

Currently, heads of government in most advanced democracies including Canada, France, the UK, Germany, Italy, Australia, New Zealand, Ireland, Israel, India and Japan all have no term limits.

Unlike the US-backed coup d’état attempt in Venezuela in 2002, so far, there is no direct evidence of direct US involvement in this Honduran military coup despite the long-standing support and training that the United States has provided to the Honduran military.  It will be interesting to see whether the Obama administration — long fallaciously thought to be leftist and progressive — will stand on the side of democracy and the people of Honduras or whether he will side with the military junta.

Texas soldiers train Honduran military

This is not the first time the Honduran military has been used as a right-wing force to overthrow a surging democratic leftist popular movement.  In fall 1963, the Honduaran military overthrew the leftist government of Dr. Ramón Villeda Morales and then ruled as a US-backed military junta until the early 1980s when power was finally transferred to a US-backed civilian administration.

Noam Chomsky notes that the US involvement in Honduaras has been a bloody affair both before, during and after the period of military rule in the 1960s and 70s.  Chomsky writes:

“Congress compelled the [Reagan] administration to state repeatedly that the human rights condition was improving not only in Guatemala but in El Salvador and Honduras so that the US could continue to support the regimes.”
-Noam Chomsky, Rogue States, p. 94


Bloggers everywhere: 2; Ipsos Reid: 0

John Wright has issued a semi-substantive response to the corpus of data presented here on this blog.

Unfortunately for Mr. Wright, as he continued talking he revealed far more than he intended.  Now that he has been talking, he seems to have stumbled headfirst into what I believe is a contradiction against himself and I invite my readers to think critically for themselves about whether they share in this analysis.

In one of his initial comments here at Paulitics, John Wright claimed that the source for his polling data was “Elections Canada”.  The suggestion was, of course, that his data integrity was superior to mine and that therefore his numbers should also be more valid.  At the time that he wrote this, I thought it was a bit strange since I was unaware of Elections Canada publishing or even keeping a database of polls released by private polling firms.  I even did a search of the Elections Canada website, and found that in the last 12 months, Elections Canada hasn’t published a single article containing the word “Ipsos”.

However, Mr. Wright gave me enough material to respond to that I just let it be.  While I knew that it was unlikely, I also knew that it was certainly not impossible that I could be missing some polls from 2009 and thus I asked him for the first time to provide his source.  This week Mr. Wright is now claiming his source as and that his source this whole time has been not just 2009 polls but also some 2008 polls (presumably also from

(makes Scooby-Doo “Ruh-roh” sound)

This is a big problem for Mr. Wright as contains nowhere even close to an exhaustive list of all polls.

The Paulitics Polling Resource and Wikipedia both documented a total 33 polls between the start of 2009 and when this whole episode blew up (two new polls have since been released, neither of which is by Ipsos Reid).  At only 12 polls, Nodice has less than half of all the polls conducted in 2009 and this is the source John Wright has been using (in addition to 2008 polls which I haven’t analysed once yet, but which I would be happy to take a look at.

Mr. Wright then states a complete and demonstrable falsehood.  He writes that

In the last of our 4 polls the NDP have been in the lower range (running 14-15 points) with EKOS and Harris Decima roughly the same and the others at the higher end. This is when Paulitics made the calculations. But just previous to that are 4 of our polls with the NDP running between 18-19 points. [emphasis added] (source)

Below is a list of the last four polls Ipsos Reid released in 2008.  See for yourself if the last four of Ipsos’s 2008 polls all have the NDP “running between 18-19 points”

Last four Ipsos Reid polls, 2008

If anybody is interested in seeing for themselves, the published sources for these four polls can be found here, here, here and here.  Furthermore, the October 9th, 2008 Ipsos Reid poll, even though it shows the NDP at 18 percent is actually still an outlier because it comes directly before a Nanos, Angus and Ekos trifecta of polls which put the NDP at an average of 20%; and it came on the same day as Segma poll which put the NDP at 21%, a Nanos poll which put the NDP at 22%, and an Ekos poll which put the NDP at 19% (source).

Finally, Mr. Wright again attempts to reiterate his claim that taking the unweighted averages of several unevenly-weighted averages is not problemmatic despite multiple, solid, irrefutable proofs to the contrary.  Mr. Wright also attempts to argue that his method of calculating averages is “perfectly legit and pretty much what the original author [i.e. me, Paul] did with his handful of polls in his charts.” (source)

I am truly sorry John, but this “new math” method calculating averages is neither legitimate nor has it been used by me.  Furthermore, since I’ve demonstrated why this technique is not “perfectly legit”, I think this argument is just a particularly sad attempt to attack the messenger when the data and the arguments presented here speak for themselves.

So, to reiterate:

  1. Mr. Wright seems to have forgotten his original source for 2009 polls (originally it was elections canada, now he says it’s
  2. Mr. Wright then admits that his numbers are actually based on a database ( that contains less than 50% of the pertinent data
  3. (perhaps because of this) Mr. Wright has been unable to provide a single poll which I have missed in my data.
  4. Mr. Wright then demonstrates a shocking lack of knowledge about the polls that his own firm released in late 2008.
  5. Mr. Wright then proved unable to articulate a single methodological fault in any of my analyses.
  6. Mr. Wright then admits that he has supplemented his 2009 data with 2008 polls
  7. Mr. Wright then reiterates once again that taking the unweighted averages of several unevently-weighted averages is not problemmatic despite multiple instances in which I have provided mathematical proof to the contrary

I’ll tell you, Mr. Wright is the gift that keeps on giving.  Swing and a miss, Mr. Wright, swing and a miss.

Multimedia: A new Paulitics resource

While I take a short break before posting a final reply to Ipsos Reid (hopefully tonight), do feel free to take a look at the newest Paulitics resource — Multimedia — which is now up and running.

It currently has powerful full-length documentaries, alternative progressive media sources, inspiring audio and visual speeches and programmes as well as free live TV news from the Canadian CPAC channel and Al Jazeera English.

Feel free to watch it by clicking below or clicking on the icon in the sidebar.

As always, comments are always welcome and appreciated.

Ipsos Reid: 0; bloggers everywhere: 1

A short while ago there was a post here on Paulitics which analyzed what I saw to be a discrepancy between the Ipsos Reid polling firm’s data and every other polling firms’ data. That post elicited a somewhat condescending and ad hominem response from John Wright, the senior vice president of Ipsos Reid.

Despite the profoundly unprofessional tone of this comment, this blog issued a detailed, statistical, and fully-sourced response addressing not only each one of Mr. Wright’s utterances, but also posited even further evidence to back up the initial argument that there is a statistically significant discrepancy between Ipsos Reid’s numbers for the Conservative and New Democratic parties and every other polling firm’s numbers for those parties. This response was in three parts and can be viewed at each of the urls below:

A Response to Ipsos Reid, Part III

A response to Ipsos Reid, Part II

A response to Ipsos Reid, Part I

After an initial non-response from Mr. Wright to these three statistical analyses of his firm’s data, Mr. Wright has now offered his only quasi-substantive retort to date. His response merits quotation to fully appreciate it:

John Wright 20 June, 2009 at 1:35 pm

I don’t need to clarify anything. My initial response was correct and I stand by it. In fact, the assertion of how I somehow didn’t calculate the right average is absurd. 20 years in the business and we do $200 Million a year in Canada alone…all I can say is that he seems to have a great deal of time on his hands to burn through a lot of graphs and charts and, well, whatever…and by the way, EKOS and us have pretty much the exact numbers out in the last 48 hours with the Liberals in front by a point. 20 June, 2009 at 3:31 pm

What I wrote previously still stands…despite all the charts and graphs and arrows and calculations, what I put up is entirely accurate.
You really should get your own polling company and test out your ideas.

So his response to the overwhelming mass of data presented here is three-fold:

  1. He attempts to push the well-known logical fallacy of an argument from authority
    – (“In fact, the assertion of how I somehow didn’t calculate the right average is absurd. 20 years in the business and we do $200 Million a year in Canada alone.”)
    – (“You really should get your own polling company and test out your ideas.”)
  2. He then attempts to engage in an ad hominem attack against me. The irony here is that his ad hominem attack is based on his belief that I have done too much research into this matter. Presumably, if I’d just started spouting off about something without backing it up with a detailed, statistical, and fully-sourced response, Mr. Wright could then have charged that I was naïve and relying on unsourced assertions that had no statistical validity.
    – (“all I can say is that he seems to have a great deal of time on his hands to burn through a lot of graphs and charts and, well, whatever.”)
  3. He then uses the profoundly unstatistical argument that his initial support for his firm’s data is still reasonable because one of his firm’s latest data points (for the Liberal Party, no less, which was never even a subject of any analysis I’ve ever done of his firm’s data) is very similar to one datapoint for the Ekos polling firm.
    – (“and by the way, EKOS and us have pretty much the exact numbers out in the last 48 hours with the Liberals in front by a point.”)

As always, I remain interested in getting to the bottom of the kind of descrepancies we have been seeing. So, if Mr. Wright would like to issue a substantive response to the actual data and arguments presented here, I would greatly welcome it.

Paulitics site changes

Hello all, I am in the process of completing what is predominantly a cosmetic change to the Paulitics blog (although there will be one major new resource added, a “multimedia” resource).

Unfortunately, this change has necessitated that I break many of links here on Paulitics as I restructure the site’s page hierarchy. Please note that most Paulitics resources are no longer available at their old urls so you may need to update your bookmarks.

Please remain patient while the site gets completely up to date. This will likely take several days.

A Response to Ipsos Reid, Part III

This is Part III in a three part series.
For Part I, click here.  For Part II, click here.

A Response to Ipsos Reid, Part III: Comparing Ipsos Reid to the other polling firms using 2 methods and with control tests.

In Part I of the Paulitics response to Ipsos Reid, I showed two graphs of the level of support given to the Conservative and New Democratic Parties in all polls conducted in 2009 by all polling firms.

As you can see below, the graphs tell a startling tale of just how radically out of step Ipsos Reid’s estimations of the Conservative and New Democratic Parties seem to be when compared to other polling firms.  I asked Mr. Wright to explain these data as I’m sure we would all be very interested in learning why we are seeing these trends.  So far, the response has been:  Crickets.  Nothing.  Deafening silence.

2009 Analysis 2
2009 Analysis 3

Now, a quick disclaimer:  As I’ve mentioned before, these graphs don’t necessarily mean that Ipsos Reid has bad data.  What these graphs mean is that either every single other polling firm in the country on average has bad data or Ipsos Reid has bad data.  As always, I invite my readers to critically consider for themselves what Occam’s razor can tell us about this impasse.

Now, to be fair to Ipsos Reid, I started thinking about possible explanations for these graphs since Mr. Wright didn’t seem keen to offer his own explanations.  One thing I noticed was that I was using the overall average of all polls conducted by all firms in 2009 and then calculating an overall average as if the movement up and down of the Conservative and New Democratic Parties wasn’t a factor. (Mr. Wright was doing this too, but keep in mind we’re being charitable here).

This method could possibly be a flawed method if, for instance, polling firm “x” released a whole slew of polls in January when Party “A” was enjoying high levels of support from all polling firms and then polling firm “x” didn’t release a poll in early February when Party “A” was down in support.  Taking an overall average without taking into account differences over time could thus yield a favourable or an unfavourable view of polling firm “x”.

For instance, if we used this methodology in the hypothetical example of the “Purple Party” described below, we would wrongly conclude that polling firm “x” grossly over-estimated the support of the “Purple Party” when in fact, they look to be more or less right on the money.

2009 Analysis 10
So, I decided to analyse every 2009 Ipsos Reid poll against the 3 polls conducted before and after so that we’re taking into account what’s happening in party support right then and there and not several months earlier when things might have been different.  I then calculated each 7 poll grouping’s standard deviation so that I could then calculate something called a “z-score” for each Ipsos Reid poll.  For those of you not conversant in z-scores don’t worry, I have also calculated the “z-score percentile” as an easy way to understand what’s going on in each of the polls.

The way this works is that if something is exactly average, it would be in the 50th percentile (because 50% is right in the middle).  If something is extremely unlikely at one end of the spectrum, you’ll get a very very low number (like something being in the 5th or 6th percentile) and conversely something extremely unlikely at the other end of the spectrum, you’ll get a very very high number (like something being in the 96th or 97th percentile).  So the name of the game here is that Ipsos Reid wants to see medium range percentile numbers and wants to avoid really high and really low percentiles in the following tables because if the numbers are really high or really low that would suggest that Ipsos’s data is way off the mark.

Here’s what I found:

2009 Analysis 5a2009 Analysis 5b2009 Analysis 5c2009 Analysis 5d2009 Analysis 5e


So what does this mean?  Well, the data in the tables above show 2 outliers at the 95% confidence interval, 5 outliers at the 90% confidence interval (including the 2 at 95%) and 7 total outliers at the 80% confidence interval (including the 5 at 90% and the 2 at 95%).

On the face of it, that seems like a lot of outliers especially considering I only analysed a total of 10 Ipsos Reid data points.  But, again to be fair to Ipsos Reid, maybe we would see roughly the same large number of outliers if we did this to any polling firm’s data in this way.  I mean after all, 7 polls is not a particularly large population.

So, I decided to run the exact same calculations on Nanos’s numbers, Ekos’s numers and Strategic’s numbers as a control group.  You can look through the tables of each of these calculations here, here and here.

As you can see, it’s not even close.

Nanos, which had 8 total data points, had zero (0) outliers at the 95% confidence interval, zero (0) outliers at the 90% confidence interval, and only one (1) outlier at the 80% confidence interval.

Strategic, which had 12 total data points (more than Ipsos and Nanos), still had zero (0) outliers at the 95% confidence interval, zero (0) outliers at the 90% confidence interval, and only one (1) outlier at the 80% confidence interval.

And lastly, Ekos, which had 6 total data points, had no outliers at either the 95%, 90% or 80% confidence intervals.

So, to put this another way, only 7.7% of Nanos, Ekos and Strategic’s 2009 Conservative and NDP data points are outliers at the 80% confidence interval.

Conversely, a whopping 70% of Ipsos Reid’s 2009 Conservative and NDP data points are outliers at the 80% confidence interval

Also, none (0%) of Nanos, Ekos and Strategic’s 2009 Conservative and NDP data were outliers at the 90% confidence interval whereas 50% (½) of Ipsos Reid’s data were outliers at the same interval.

If you are like me, and like to view findings like this in ‘pretty graph form’, here you go:

2009 Analysis 9a
2009 Analysis 9b
Oh, and just for fun, I also ran the same basic calculations for each Ipsos Reid poll compared to the entire set of 2009 polls and came up with almost the exact same results.  You can see the tables for these calculations by clicking here.

The evidence supporting my initial contention is mounting and the responses we’ve gotten from Mr. Wright have been a mixture of snide, sardonic comments, fallacies and silence.

I believe we all deserve better than that.

A response to Ipsos Reid, Part II

This is Part II in a three part series.
For Part I, click here.  For Part III, click here.

A Response to Ipsos Reid, Part II: How to “prove” that the Canadian House of Commons actually has 429 seats instead of the 308 seats it actually has.

Here’s how you do this: just use the mathematical practice of taking the average of several unequally-weighted averages.  Normally, this would be considered a big “no-no” in the world of statistics, but since a senior official with the prestigious Ipsos Reid polling firm recently came out and endorsed this practice, we’re good to go.

First we need some numbers.

According to Statistics Canada, the most current estimates of the Canadian population are the following:

January, 2009







Nova Scotia


New Brunswick












British Columbia




Northwest Territories




Now, each province has the following number of seats in the House of Commons:







Nova Scotia


New Brunswick












British Columbia




Northwest Territories




So, we just do a little long division using our handy computer machines and we get the following estimation of the population per seat in each province:





Nova Scotia


New Brunswick












British Columbia




Northwest Territories




Now, using this methodology, there should be nothing wrong with just taking the average of these averages to come up with the Canadian average:

Doing so, we get an average population/seat value of:



… But, we already know from Stats Canada’s data that Canada has a total population of roughly 33,504,680 people.  Therefore:

33,504,680 ÷ 78,048 = 429 seats


And presto, voila!  You’ve just proven that the Canadian House of Commons has 429 seats even though it actually contains 308.

Can you spot the problem?  The problem is that this practice — which Mr. Wright publicly came here and said was legitimate — does not take into account that taking an unweighted average of several averages means that Ontario (population 12,986,857) and Nunavut (population 31,556) are considered equally weighted.

A response to Ipsos Reid, Part I

This is Part I in a three part series.
For Part II, click here.  For Part III, click here.

Paulitics investigation

John Wright, the senior vice president of Ipsos Reid polling firm — and longtime avid Paulitics reader — recently made some comments here at Paulitics that called into question the methodology I used to examine a perceived discrepancy in Ipsos Reid’s polling data.

These comments deserve to be answered in a respectful and courteous manner irrespective of the manner in which they were delivered.

Mr. Wright explains in his comment:

First, the Conservative average vote percentage January 2008-June 2009 per polling company:

EKOS 37.5
Angus Reid 36.8
Ipsos Reid 36.2
Nanos 35.6
Strategic Counsel 35.2
Harris-Decima 34.8

The Mean: 36.02

Leaving aside for a moment the issue of where Mr. Wright got those individual averages for each polling firm (my numbers are completely different and I have meticulously documented all polls covered by the mainstream Canadian press here), Mr. Wright  just committed a very serious error in the statistical world and it is a very amateurish one at that.  Can you spot it?  It’s not his math (again, leaving aside the issue of the individual polling firm averages).  The problem is that, in his attempt to refute me, Mr. Wright took the averages of each polling firm and then took the unweighted average of the average. This, as anybody who deals at all with statistics will understand, of course, is a big “no-no”.

Consider if I did the same thing with the following data set:

Do you support the Metallic Metals Act of 1898?

Polling firm x (poll #1):   Yes, 75%
Polling firm x (poll #2):   Yes, 80%
Polling firm x (poll #3):   Yes, 70%
Polling firm x (poll #4):   Yes, 85%
Polling firm x (poll #5):   Yes, 65%
Polling firm x (poll #7):   Yes, 74%
Polling firm x (poll #8):   Yes, 76%

Polling firm y (poll #1):   Yes, 30%

Polling firm z (poll #1):   Yes, 36%

Polling firm “W” (poll #1):  Yes, 45%

Average of each polling firm:

x: 75%
y: 30%
z: 36%
“W”: 45%

Therefore, using Mr. Wright’s methodology for ‘rebutting’ my argument, the average of polling firms “x”, “y” and “z” = 46.7% which just goes to show that the result for polling firm “W” is both mainstream and accurate?

Then there’s the issue of each polling firm’s numbers.  Readers can feel free to look up all of the polls published and reported on in the press for 2009 (here or here).  Based on the data in these lists (which I believe to be exhaustive), the averages for each polling firm are as follows:

polling firm averages

Now, in the interest of full disclosure, I discovered today that I had accidentally overlooked one poll conducted by Ipsos Reid on March 5th, 2009.  I fully apologize for the error, however, as the math will demonstrate, this one omission on my part does not change matters much.  With the formerly absent March poll, the Ipsos Reid sheaf of 2009 polls (that I have found) are as follows:

Ipsos Reid 2009

I am unsure of where Mr. Wright is getting his 2009 averages for each of the aforementioned polling firms, but I would be extremely interested in hearing directly from him on this matter.  His number for Ipsos itself is close enough to mine that I would conjecture that he is probably just using the decimal values for each party obtained by his firm whereas I am using the rounded whole numbers published in the press.  Unfortunately, Mr. Wright provides neither a source nor does he show his math.  I have done both and am willing to change my position if he can show me a sheaf of polls that I’ve missed.

Even with this new missing Ipsos Reid poll, however, if we compare each 2009 Ipsos Reid poll to the three polls which directly preceeded and succeeded it, we still get these startling graphs:

2009 Analysis 3

2009 Analysis 2

Can Mr. Wright explain these graphs?

Stay tuned for Parts II and III of this response to Ipsos Reid, wherein the dénouement is reached and I will break out the z-scores, standard deviation calculations and other extreme geek methods to argue that this whole thing isn’t just a matter of attractive yet deceptive graphs.

Paulitics temporarily back to old layout

Paulitics is back to its old look temporarily until I can launch a new and improved layout that allows for both the new Paulitics word logo and red flag of socialism logo as well as providing enough width for the Paulitics polling resource page (and a sidebar that doesn’t apply only to the home page and not to individual blog posts).

The process is a complicated one given the limited availability of templates suitable to a blog with as many associated pages as this one has.

Stay tuned for more information…


home page polling resource

Click below to download the

Paulitics Blog Search

Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 2.5 Canada License.

DISCLAIMER: The views and opinions expressed in the comments section beneath each post on this blog do not necessarily reflect the views and opinions of the blog's author and creator. Individual commentators on this blog accept full responsibility for any and all utterances.


Progressive Bloggers

Blogging Canadians

Blogging Change

Paulitics Blog Stats

  • 863,959 hits since 20 November, 2006