Census highlights flaws in data gathering
Survey may not be asking the right questions for collecting race-based info
OTTAWA—When Ashley Pleasant filled out the long-form version of Canada’s census earlier this month, one question gave her pause.
It asked respondents to select the population groups to which they belong, all of which, aside from white, are designated as visible minorities. People are free to tick off multiple boxes, or input another answer of their choosing.
Pleasant’s mother is Black and her father is white. But checking off both the “Black” and “white” boxes didn’t feel quite right.
“I am not a Black person plus a white person,” the Hamilton resident told the Star. “I’m this new, emergent thing that I think should have ... language to identify it.”
Pleasant’s experience with the census comes as Canada is rethinking what types of racebased data it collects, and for what purposes.
And it highlights the fact that this year’s census — the government’s most powerful tool for collecting that data — might not be asking the questions that need to be answered as Canada’s economy emerges from its pandemic shutdown, experts say.
COVID-19 exposed the economic and health inequities facing racialized and Indigenous communities, prompting the federal government to invest in collecting better data on these issues. This year’s federal budget proposed $172 million over five years to implement a “disaggregated data action plan” to improve the country’s understanding of diverse populations. Millions more in funding is earmarked to bolster data collection efforts that would better inform Canada’s justice system and the RCMP.
But the preoccupation with diversity data in this country came as the 2021 census was already well underway.
The questions that appeared on Canadians’ computer screens were crafted through an extensive consultation and testing process that began in 2017 and ended last summer.
The short-form version of the census that 75 per cent of Canadians received excluded questions about visible minority or ethnic backgrounds. The remaining 25 per cent of respondents completed the long-form census, which posed questions about population groups, Indigenous identity and ethnic and cultural origin.
“We feel that 25 per cent should be sufficient to generalize the result to the total population,” said Tina Chiu, Statistics Canada’s director for diversity and socio-cultural statistics.
Discussions did take place, Chiu told the Star, over whether to move some questions from the long version to the short version. In the end, ensuring Canadians were not overwhelmed by the “response burden” of a lengthy questionnaire won out.
Dr. Kwame McKenzie, CEO of the Wellesley Institute and director of health equity at the Centre for Addiction and Mental Health, says including such questions in the version the majority of Canadians received would provide a clearer sense of the country’s diversity.
“If we collect that data on the short-form census, what we can also do is not only get a full and proper picture of the country, but it allows us to use that data for the whole country to try and understand other things that are going on,” McKenzie said.
The question that initially stumped Ashley Pleasant became part of the census in 1996.
The purpose of asking people to identify with one or more of the 11 listed population groups is closely tied to the Employment Equity Act, which necessitated collecting information about visible minorities.
Visible minorities are assigned to a mix of groups: some broad — South Asian and Latin American, for example — and others more precise, such as Filipino and Japanese.
Behind the scenes, Statistics Canada’s advisory committee on immigration and ethnocultural statistics discussed the merits of the groupings, but never reached consensus on the issue.
“We have committee members who think that yes, it’s not perfect, but it is the best right now to collect that information. Some would debate about the term visible minority,” Chiu said.
“But we have to respect the Employment Equity Act.”
Neither version of the census uses the word “race” because the term is complex and didn’t test well, Chiu said.
The decision to use imprecise wording is part of the problem, McKenzie said.
“The problem with the longform census is it’s not clear what it wants to know. It’s not clear that it wants to know about race, or whether it wants to know about culture or heritage. It wants to know a bit about everything and so people get confused.”
In one question, respondents are asked about the ethnic or cultural origins of their ancestors. The wording briefly puzzled Pleasant: “ancestors” was not defined, so she had to Google how many generations back she should be looking.
“Asking about people’s ancestors, that becomes fairly difficult fairly quickly,” McKenzie said. “If you go far back enough ... people say that Africa is the centre of civilization, so what do they mean?”
Chiu says the language that winds up in every census is carefully considered, adding that the ancestry question was left up to the respondent’s interpretation to give them more freedom in their answer.
But that doesn’t mean the agency isn’t open to future changes.
“Our consultation continues, and leading to the 2026 census, we’ll be testing more versions of the questions,” Chiu said.
For McKenzie, getting it right is critical.
“In the world at the moment, data is a particular currency,” he said. “If you’re not seen in the data, you are invisible.”
“We feel that 25 per cent should be sufficient to generalize the result to the total population.”
TINA CHIU
STATISTICS CANADA’S DIRECTOR FOR DIVERSITY AND SOCIO-CULTURAL STATISTICS