Painting a Portrait of Canada: The 2021 Census of Population
7. Data quality

Skip to text

Text begins

The census is designed to provide timely, high-quality information on Canadian communities from coast to coast to coast. It provides a statistical portrait of Canada that is timely and relevant, and it gives Canadians direct access to a vast repository of information they can draw upon to better understand their communities. Statistics Canada is confident that its census data are of high quality and provide an accurate portrait of the resident population in Canada.

Quality assurance

Quality assurance takes place throughout the census process. It starts before the data are collected and ends after they are released.

Statistics Canada draws widely on expertise from within the agency to ensure the data are of high quality. Various specialists provide expert advice on what content to collect, the coding of written responses, edit and imputation rules, and the certification of data outputs. They also support field operations, processing activities, dissemination, and data evaluation and analysis. They consult with stakeholders and provide professional advice and assistance to data users on how to properly use the census data.

Reliability of census data

Following the collection of information from both the online and paper questionnaires, the information provided by respondents is processed and quality assured.

Data quality assessment is an evaluation of the overall quality of census data. The results of this assessment are used to inform users of the reliability of the data, make improvements for the next census and adjust census data. Two coverage studies—the reverse record check and the Census Overcoverage Study—are used to produce official population estimates. These steps ensure that all Canadians are included in the census, and only once.

Identifying types of error

No matter how well a census is designed, the data collected will inevitably contain omissions and errors. Errors can occur at virtually any stage of the census process, from material preparation to the creation of the dwelling list, and from data collection to processing. Census data users should be aware of the types of errors that can occur so that they can assess the usefulness of the data for their own purposes.

Coverage errors

Coverage errors occur when dwellings or individuals are missed, incorrectly enumerated or counted more than once. Statistics Canada measures coverage errors by using sample surveys and carefully analyzing census records.

Statistics Canada takes several steps to improve coverage, including

All operations—particularly data collection procedures—are designed to ensure accurate coverage.

Non-response errors

Non-response errors occur when some or all information about individuals, households or dwellings is not provided. This can happen when household members are away throughout the census collection period or—in rare instances—when the head of household refuses to complete the questionnaire. However, what happens more frequently is that questionnaires are returned by mail or submitted online, but responses to certain questions are missing.

For the short-form questionnaire, an analysis is performed to detect significant cases of partial non-response, and follow-up interviews are attempted to obtain the missing information. However, despite these efforts, a small number of responses will be still missing at the end of the collection stage.

Missing responses are corrected during processing by imputation—a process by which a missing response is replaced with the corresponding response from a similar record.

For the long-form questionnaire, weighting is used for complete non-response. Weighting is when responding households are assigned a weight to represent a number of other similar households, resulting in the long-form sample representing the entire Canadian population in the final results. Missing responses are also corrected during processing by imputation to provide substitute values. 

Imputation and weighting are common and statistically sound approaches used to ensure that the final results are representative of the entire population.

Measuring data quality

Coverage studies

After data collection is completed, the agency conducts data quality studies to assess the impact of errors and understand how and where errors occur.

The census defines the population to be enumerated, as well as the rules by which the population will be counted. Coverage error occurs in the application of these definitions and rules.

The main sources of coverage errors are when a dwelling is omitted, which results in that dwelling’s residents not being counted, and when a respondent does not include all people who should be included or excludes people who should not be excluded as part of the household.

The counts usually produced from a census slightly undercount the population. This is called net undercoverage, which indicates the extent to which the number of enumerations included in the census data is lower than the actual population. Both undercoverage and overcoverage may produce a bias in official counts and estimates because the characteristics of people who are not included may differ from the characteristics of people who are included, and the characteristics of duplicates may differ from the characteristics of people who are included only once.

Statistics Canada conducts three studies to measure coverage errors:

Dwelling Classification Survey (DCS): One type of census coverage error is the misclassification of dwelling occupancy. This can occur during census operations when an occupied dwelling is incorrectly classified as unoccupied, or when an unoccupied dwelling is incorrectly classified as occupied. This misclassification can affect any dwelling for which a census questionnaire is not received. The purpose of the DCS is to study this type of classification error on a sample basis and provide estimates of occupied dwellings for which no questionnaire was received. The survey results are used to make adjustments to the information in the census database.

Reverse Record Check (RRC): Another type of coverage error is undercoverage (i.e., people not included in the census count), which occurs when the list of people in a household on the census questionnaire is incomplete or when an entire household is missing. The RRC is a sample study that identifies and provides estimates of people missed by the census and not included in the census counts. Undercoverage estimates are not used to adjust the released census counts, but are an input to the demographic population estimates. These estimates are updated and released regularly between censuses.

Census Overcoverage Study (COS): A third type of coverage error is overcoverage (i.e., people counted more than once in the census). This occurs when two questionnaires are received from the same household or when one person appears in two households and, as a result, on two questionnaires. The COS identifies pairs of people in the census database who are likely to be the same person and determines—on a sample basis—which of these pairs are duplicates. An estimate of the number of people counted more than once is then produced. Overcoverage estimates are not used to adjust the released census counts, but are an input to the demographic population estimates. When combined with the results from the RRC, the COS results provide an estimate of net coverage error in the census data.


Certification consists of several activities that rigorously assess the quality of the census data at specific levels of geography to ensure that the quality standards for public release are met. This evaluation includes the certification of population and dwelling counts and variables related to dwelling and population characteristics.

During certification a number of data quality measures are used, including response rates, edit failure rates, and a comparison of data before and after imputation. Tabulations for the census are produced and compared with corresponding data from past censuses, other surveys and administrative sources. Detailed cross-tabulations are also checked for consistency and accuracy.

External experts may also be consulted to support certification activities for specific census variables.

Depending on the certification results, census data can be

Formal reviews

When Statistics Canada releases population and dwelling counts from the census, data users sometimes question the validity of the counts for a particular geographic area, such as a municipality or submunicipal area.

For the 2021 Census of Population, as in past census cycles, Statistics Canada will conduct a formal review for each official request received.

In cases where no significant error in the population and dwelling counts is detected, Statistics Canada will confirm the published census counts with a written explanation.

In the rare cases where errors in the population and dwelling counts are confirmed, Statistics Canada shares its research results and the revised counts. It also notifies the province or territory in question—and any municipalities or other places affected—of the revised counts.

Report a problem on this page

Is something not working? Is there information outdated? Can't find what you're looking for?

Please contact us and let us know how we can help you.

Privacy notice

Date modified: