I keep hearing voices say that Canadians are paying more and receiving lower speeds than our peers in the OECD. I keep saying that the data is biased. A couple leading academic institutions cite OECD statistics in their reports – meaning their studies are equally biased.
How many of the critics have actually looked at the data that the OECD used [
xls, 286KB]?
Indulge me, please.
Download the spreadsheet and look at the tab called “BBPricing”. Look through the worksheet and you will see the basis of the speed and pricing and price per megabit that keeps getting quoted. Canada appears on lines 117-133 of the worksheet – line 133 is the arithmetic average of the 16 lines above it. It is as simple as that.
I want you to take a little time to actually go through these numbers, because I think that anyone who actually looks at the table will see the same problems that we have with the data.
Take some time to get a really good look at how the OECD numbers were developed, because these are at the core of the Oxford and Harvard studies.
Anyone wonder why Australia has 71 samples of data – some duplicates – and Canada has just 16? Why little Dansk Bredbånd (less than 100,000 subscribers in Denmark) merits 9 samples from their offerings – more than Bell Canada’s 8 and more than double the representation from Rogers and Shaw). The OECD liked Dansk Bredbånd enough to include duplication of 3 of its offers, thereby over-weighting the highest speeds.
Videotron’s high speed services were ignored by OECD, apparently because it isn’t one of the top 3 providers in Canada. So, no coverage of TELUS, Bell Aliant, Cogeco or Novus either.
The OECD didn’t limit its review of other countries to only 3 service providers. For example, 5 providers were sampled in Ireland; 4 in Iceland; 4 in France; 4 in Japan; 4 in Netherlands; 5 in Sweden; 6 in Spain; 4 in Portugal; 4 in the Slovak Republic; 5 in the US. Why only 3 from Canada?
Governments can’t set a broadband policy based on faulty data.
Why isn’t the OECD reporting of Canada more complete? That is a question that politicians should be asking.
Since you seem to know something that we don't, why don't you use all your wonderful unbiased data and expand the spreadsheet to include more offering. Remember that not every offering today was available at the time OCDE compiled theirs.
We'll see if it changes something then. It's always easier to say something is wrong than to actually prove it.
Also should you do this, don't forget to include all the source of information.
"I keep hearing voices say that Canadians are paying more and receiving lower speeds than our peers in the OECD. I keep saying that the data is biased. A couple leading academic institutions cite OECD statistics in their reports – meaning their studies are equally biased."
Yet the Oxford report acknowledged their data's bias, something researchers are supposed to do.
You can claim it's "biased" all you want. Biased is a very open term in regards to statistics. It can mean that certain methods of gathering data create bias, or that a study funded by cigarette companies, for example, that states smoking is harmless can be biased.
There's "acceptable" bias, one researchers try to account for and acknowledge in their studies, and then there bias. Like my cigarette study example.
No offense Mark, but your post reminds me of the pot and the kettle: "This OECD data is biased! Biased, biased, biased! Oh, the major Canadian ISP's funded my paper, you say? No bias here!".
You had a perfectly good opportunity to clear the air in your report and acknowledge your bias. Yet upon reading the paper, i see no admission of bias. If i'm wrong, please link me to the page of your report where you do acknowledge your bias, and i will withdraw my comment. Any bias, no matter how small, must be identified and acknowledged in a research paper, or the integrity of said paper suffers as a result.
Radar – on page "i", the first page after the cover, we state pretty clearly "This report, commissioned by a group of Canada’s largest internet service
providers…" and footnote the 7 funders of the report.
There was nothing hidden.
I did search the paper and saw this. However, i didn't see any statements that stated "because our report is largely funded by the major telecoms, there is an element of bias that is evident in this report". It is one thing to to state that these people funded your research; it is another to admit tht this is a bias, or conflict of interest.
Though technically you did disprove my "no statement of bias" comment. But my other points still stand.