On Feb. 7, the Winnipeg Free Press published an article by the chief statistician of Manitoba entitled Statistics Canada underestimated Manitoba's population. I would like to correct this claim and reiterate that the population estimate for Manitoba is unbiased, free of error and is based on a well-established and statistically sound methodology. This approach was agreed upon by all provinces and territories and was consistently applied.
Every five years, population estimates for the provinces and territories are revised using information from two sources: the most recent census and census coverage studies. The 2011 Census of Population was very successful in Manitoba, with a response rate of 97.4 per cent -- the third-highest among the provinces and territories.
The coverage studies confirmed the high quality of the census in Manitoba. The province had the third-lowest rate of net under-coverage at 1.8 per cent, a substantial improvement in net under-coverage in Manitoba from the 2006 census.
The process to produce population estimates is based on both sound and accepted statistical methods that have been used and refined over a number of census cycles. The processes, methods and findings are carefully assessed and validated through a comprehensive review involving representatives from each province and territory.
During this process, every concern brought forward by the chief statistician of Manitoba was carefully assessed and a comprehensive review was conducted. Statistics Canada reviewed all survey steps for both the census and the census coverage studies and found no evidence of any error in its processes and findings and confirmed the new population estimate for Manitoba.
Contrary to what was stated by the chief statistician of Manitoba regarding evidence of a biased sample for the census coverage study, no evidence of error in terms of a biased sample was found. Statistics Canada thoroughly reviewed and assessed the properties of the survey frame, as well as the sampling and weighting procedures to produce the new estimates for Manitoba and the other provinces and territories. The results of this evaluation were presented to all provincial-territorial representatives prior to the publishing of the new population estimates in September 2013.
The claim that Statistics Canada deleted population from Manitoba's count is a false statement. The purpose of census coverage studies is to make appropriate adjustments to the census count for those who were missed or counted more than once in the census. Through this process, the population estimate of every province and territory is adjusted upward or downward according to the findings of the coverage studies. This process is used in every census cycle.
The chief statistician of Manitoba points to increases in the number of individuals filing income tax returns in Manitoba as an indicator of population growth. Tax files on their own cannot produce accurate measures of population growth. They represent an incomplete picture of the total population, as groups such as children, aboriginal people, the self-employed and recent immigrants are not fully represented in tax records. A growth of the tax-filing population must be cautiously interpreted as increases in tax filing can be highly sensitive to changing economic conditions and tax policies.
Since Statistics Canada has received no new evidence that would warrant any further review of the Manitoba population estimate, no adjustment can be made that is scientifically defensible.
It is important that the population-estimates program for Canada, the provinces and territories be based on a common, transparent and robust statistical methodology and be free of arbitrary adjustments.
Wayne R. Smith is the chief statistician of Canada.