Population Numbers Quality Varies Across Canada
by William Warren Munroe, January 2012

Perhaps, Statistics Canada should be called, " Statistics most of Canada" as there are three jurisdictions that have their own Official Statistical Agencies ... and guess what, BC, Quebec, and the NWT's methods are said to be better.

According to the Ministry of Labour and Citizens' Services, BC Stats' methods are "unparalleled". The Minister's Assistant Deputy refers to a 2005 feasibility study prepared by Statistics Canada (STC) for Finance Canada, stating: " ...population estimates produced by three provincial/territorial statistical agencies (Quebec, British Columbia, Northwest Territories) were all found to be of better quality when compared with those based on any of the three STC methods. " (Endnote #1)

However, Statistics Canada's Demography Division says that they did not review BC Stats' use of the change in telephone landlines hookups to estimate population change. (Endnote #2)

Is the 2005 feasibility study reliable? The authors of the 2005 feasibility study could not check the quality of the methods of these three jurisdictions or Statistics Canada’s Demography Division against the 2006 census results.

Indeed, the post censal estimates (the years after the last census) are adjusted according to the next census results; therefore, the estimates between 2001 and 2006 were changed after the 2006 census results were in, making the 2005 study results subject to change. Can we see the testing done to support the statements made in the feasibility study, as well as the real methods used by BC Stats?

Yet, Statistics Canada does ask for peer reviews of their proposed refinements in population estimation methods and updates changes in a timely manner; unfortunately, BC Stats does not adhere to these practices.

In BC Stats' population estimation methods paper dated August 2011, revised from the last methods paper from 1998, BC Stats claims that the change in telephone landline hooksup along with electrical landline hooksup was "extensively analyzed". But if BC Stats' Population Analyst (myself) did not see the testing of telephone data (I asked for this for years and insisted that the methods be published for clients and users to see to no avail), and nor did Statistics Canada's Demography Division, then who else participated in the "extensive" analysis of the use of this variable?

I asked the Canada Population Society, out of the University of Victoria, whether they were asked by BC Stats to review the use of the change in number of telephone landline hookups to estimate population change, but alas, the Society does not review government methods and models. (Endnote #3)

Perhaps, "extensively analyzed" means that a few people, or that one person, thoroughly tested the data purchased from Telus? But why wasn't a writeup done and why wasn't it publicized? To date the testing of this variable has not been made available. What is BC Stats hiding?

BC Stats can make proposed changes to the methods available to peers and clients and BC Stats can publish changes to the methods in a timely manner (before the numbers are published).

The best thing to do is for the Ministry of Labour and Citizens' Services to make the writeup of the "extensive" analysis available online with the inputs, so that the numbers can be verified. Also, BC Stats can publish the many other changes to the methods made in the last decade ... unless of course there is an effort to "guard" the integrity of BC Stats. (Endnote #4)

BC Stats' population numbers are used to justify opening and closing of public facilities including schools of which there were close to 200 closed between 2001 and 2011. We sure do not want to be closing public schools based on unreliable numbers.

If BC Stats methods are better than Statistics Canada, what should Statistics Canada do? Should Statistics Canada adopt BC Stats methods? Perhaps the rest of Canada should adopt BC Stats' methods and practices?

Unfortunately, for the rest of Canada, this would result in the loss of the ability to review the methods used to create numbers that are in turn used to justify the opening and closing of public and private facilities, including schools.


Endnotes:

1 BC Stats' effort to "guard" the integrity of BC Stats - part1, - part2.

2 Statistics Canada's Demography Division did not review the use of the change in telephone landlines.

3 Canada Population Society does not review government methods and models.

4 BC Stats' effort to "guard" the integrity of BC Stats - part1, - part2.

Here is my response to the letter from the Ministry of Labour and Citizens' Services

Back to Top


Website content, code, and design by W.W. Munroe. Copyright 2012.