Sampling and Weighting Technical Report, Census of Population, 2021
8. Conclusion

The 2021 Census long-form questionnaire saw the introduction of considerable new content, and its collection took place during a pandemic. Both of these factors posed methodological and operational challenges. In particular, significant efforts were made to achieve response rates similar to those in 2016. Additionally, administrative data were introduced in the imputation process to try to minimize non‑response bias in the rare areas where high response rates could not be achieved. Despite those challenges, sampling and weighting methods were by and large a continuation of the 2016 Census long-form methods.

The introduction of a new dissemination system allowed Statistics Canada to revamp and improve the dissemination of data quality indicators. In an effort to empower data users to make valid statistical inference, new confidence interval methods were developed and implemented. Confidence intervals became the key variance-based data quality indicator and are disseminated with the majority of long-form data tables. Additionally, detailed non‑response and imputation rate statistics per question were made available to data users to make the data quality more transparent. For more information, see the 2021 Census Data Quality Guidelines, Statistics Canada Catalogue no. 98-26-0006.

• Previous
• Next