Skip to content

Data errors paint unreal picture of housing development in small communities

An error that would go unnoticed in Edmonton can obliterate small town statistics.
Even minor data errors can have a big impact in small towns and rural communities.

According to data from the Government of Alberta, 2023 was a devastating year for housing development in Cold Lake.

The province's regional dashboard of economic indicators shows that after consecutive years of steady construction, there wasn't a single housing start in the city in 2023. None.

Cold Lake Mayor Craig Copeland chuckles over the phone when he hears this: "That's not very accurate."

"In terms of new single-family homes, we had around 30 to 40 housing starts," Copeland said. "Which is below where we want to be, but there is definitely a subdivision getting built by the hospital."

The discrepancy between housing starts appearing in government spreadsheets and the actual concrete footings poured in Cold Lake is really quite small. But the data paints a starkly different picture of what is going on in the community.

It's an example of how, in world where people look to official data to inform every decision, even tiny errors can impact small communities, says Richard G. Shearmur, a professor at McGill University's School of Urban Planning. 

"There are data errors everywhere, whether it's in the larger communities or the smaller ones. But typically in larger communities, errors will tend to balance out,"  Shearmur said.

"Whereas if an error happens in a small community, and they only have 50 units to start with, a 50-unit error can totally obliterate or really undermine their statistics."

While statistical mistakes carry more weight in small municipalities, those same towns are also at a disadvantage when it comes to catching and correcting errors. Calgary, for example, can staff economists and data specialists to can go through and make sure all its data is gathered and verified properly.

"This requires a lot of capacity within the municipality, which can be afforded by large cities," Shearmur said. "You need the same capacities in a small city. It's not as if because you only have 2,000 people that you need less statistical knowledge or less knowledge of databases to pick up these errors. Yet, very often, in a small community you just have a few people who were doing everything. And they don't have the very specific skills and knowledge that it takes to go into the databases to make sure the data has been gathered, and even to know really who to pointed out to if they do suspect that there's an error."

In a city the size of Edmonton, a data error of 50 homes wouldn't even register, Copeland suggested. With roughly 5,000 homes in Cold Lake, that same error is equal to one per cent of the entire housing stock.

"It does give a misconception that our community is not growing. So from a development standpoint, when you're trying to showcase your community for outside developers on the retail, the commercial side, the multifamily, or just investing in a quarter section to put up single family homes, seeing an inaccurate numbers is concerning," Copeland said. 

Copeland says he's not overly worried about this error, though, because the development world is small and most builders already know what to expect in Cold Lake. Anyone who looks would be able to find listings online for new homes in the city, he said, and the the local administration has recently introduced an incentive to get developers to build multifamily apartments for the rental market.

Shearmer previously co-authored an article laying out some of the problems data errors pose to small communities, such as being excluded from policies and programs that are targeted using thresholds based on population or income.

With growing disengagement from mainstream politics in rural communities, any public-facing information that is at odds with what leaders say or what residents know to be true can also fuel distrust, Shearmer said.

"I do think this source of error, where you have official statistics, which are obviously patently wrong, yet are still presented as official statistics, really doesn't help. It feeds into the wider story of distrust between the establishment, the main sort of administrations, which are often central, and are often based in provincial capitals or in Ottawa on the one hand, and the smaller outlying communities who feel really that they're not taken that seriously, for a variety of reasons."

Shearmer, with a long career spent combing through data, said he can spot and understand the ins and outs of these data errors, "but obviously most people don't have that level of time to think about it. So to them, it can just be perceived as being ignored."

Statistics for most of the cities, towns, and counties presented in Alberta's housing starts data is consistent with reporting from the Canadian Mortgage and Housing Corporation or the municipalities own information. The province had no conclusive answer to why housing data for a handful of municipalities was askew, but Josh Aldrich, press secretary to Alberta's Minister of Jobs, Economy, and Trade, offered the following explanation.

"Jobs, Economy and Trade uses several Statistics Canada tables to track housing starts. Statistics Canada gets their data from the Canadian Mortgage and Housing Corporation (CMHC). 

"The differences are likely due to larger municipalities, such as Edmonton or Calgary, being reported as the Census Metropolitan Area or Census Agglomeration on the Government of Alberta’s Open Data Portal, as well as data revisions made by the CMHC to historical data after initial publication.

We are in the process of changing our reporting to align more directly with data published on CMHC’s website."

About the Author: Brett McKay, Local Journalism Initiative Reporter

Read more


push icon
Be the first to read breaking stories. Enable push notifications on your device. Disable anytime.
No thanks