Today’s Globe and Mail has a great article on the importance of data quality.
Mostly people know, or quickly learn, that lots of data is available for use by real estate professionals. It takes a little longer to realize that some of it is of high quality and some of it is not. The difference is not based on the number of numbers being reported but on how much readers can rely on those numbers.
Demographic data used to be some of the highest quality data, coming from the census conducted every 5 years by Statistics Canada. That was until the formerly-mandatory “long form census” was replaced a voluntary form “National Home Survey”. A few people objected to the long form census but “few” was numbered in the low 100s out of a total population of 35 million. As part of the the Statistics Act and other regulations, Statistics Canada is limited in what it is allowed to release (and has procedures in place to implement that limitation).
Many national organizations objected to using the Survey before and after the announcement. As expected, some individuals did not respond to the voluntary survey, and selectivity contaminates analysis. For example, when comparing the demographic characteristics of the people in different locations, the fact that certain types are less likely to respond means that the reported data must be adjusted. It is unclear how much “less” since the reason to collect this information is that the true answer is not known. Not knowing is the reason why a census is conducted every five years instead of thinking that every 25 years might be adequate. As the President of the Canadian Association of Business Economics wrote, the consequences are many, serious and hidden. (If you think about the last flaw a bit, you will note that it is even worse since it is impossible to go back in time to collect missing information.)
Today’s article makes for interesting reading based on what it says and, to a thoughtful reader or one who is still learning, based on what it implies.
(A private member’s bill has been introduced to return to the long form census.)