Do we need a census in the middle of a pandemic?

Gabriel Borges, OLAC

Censuses have been delayed, interrupted or cancelled in many countries due to the COVID-19 pandemic. Only a few countries conducted them in 2020. The censuses in the United States and in Mexico had already started when those countries were hit by the pandemic, and they made what seemed to be the most sensible decision: continue the census, with a few last-minute adjustments, mainly to protect the health and safety of the National Statistical Offices staff and the public.

Impact on field-based enumeration – Censuses scheduled in year 2020
Source: United Nations Statistical Division (UNSD)

Countries that were planning to conduct their censuses in late 2020 had to cancel or postpone them. Brazil and Ecuador, for instance, postponed their censuses and are now faced with adversities to take them in 2021, both because of the out of control situation of the pandemic, and the difficulties to get the census budged.

In addition to the difficulties to conduct this massive operation under such adverse circumstances, one technical question that remains unanswered is whether it is worth it to carry out a census that provides a snapshot of the population that might be distorted by the transformations of the pandemic. A census during or right after the pandemic, particularly in hardest hit countries, will capture the consequences of the pandemic and the arrangements society had to make to “survive” it. Besides the obvious impact on mortality, the pandemic affects the economy and the labour market, education, fertility, migration and household composition, among others. Some of these effects might be permanent and some may be just a short-term shock reflecting the pandemic situation.

These issues have been discussed in the context of planning the 2021 United Kingdom Census. Some defended to postpone the census to 2022, as future decisions risk being based on skewed figures, while others argued that in reality, the 2021 census may prove more important than ever, as it will show exactly how and why the pandemic has affected different areas and population groups of the country.

Both arguments are compelling and a reasonable alternative has been to conduct an extra census in 2026, which could assess how much the country has recovered in the five years from the previous census and could also show a picture of the country in less extraordinary times.

Going back to the original question of this post, the short answer is: yes, we need a census during or right after the pandemic, as long as it is safe to do so. Most countries have had a decade or so without census data and they need this information updated. Furthermore, it is important to assess the consequences of the pandemic for the development and implementation of post-pandemic recovery plans.

However, it is clear that we need a different census than that planned before the pandemic. Changes in data collection techniques, questionnaire design and basic concepts are probably necessary. Some countries are working on introducing new data collection modes to replace, at least partially, face-to-face interviews. Furthermore, a more extensive census questionnaire has been replaced to a short form focusing on few core topics in order to shorten duration of the interview.

Conceptual changes might also be necessary to take into account the pandemic. The 2021 UK Census, for instance, developed and tested additional guidance to support respondents in answering Census 2021 questions in this context. The updated questionnaire includes instructions such as “if the coronavirus pandemic affected your [subject of question], select the answer that best describes your current circumstances”.

This all brings a lot of concerns, as last-minute changes are always risky. Thus, the need for changes in the census planning has to be considered simultaneously with the ability to do so without compromising the quality of the census. Also, some of those changes, such as changing the data collection mode, might require a significant additional investment, which contrasts with the cut in the census budget in many countries.

Finally, a census carried out in extraordinary times may not be able to fulfil some of its most essential roles, such as being used as benchmarks for statistical compilation, sampling frame for sample surveys and research and analysis, like population projections. A snapshot of the population distorted by the pandemic do not meet the requirements to be used for many years (up to a decade) as benchmark for sample surveys and population projections, for instance.

Given these perspectives, countries should plan to conduct their censuses in the coming months, but also start considering a new census before the 2030 census round.

3 pensamientos en “Do we need a census in the middle of a pandemic?

  1. Thanks for putting this piece out, nice to see several sides, Gabriel. To be clearer in some points, here some questions. When you say: “countries should plan to conduct their censuses in the coming months”. Please, could you be more specific about the meaning of “coming months”? Also, can you be more specific about this statement: “A snapshot of the population distorted by the pandemic”? What kind of distortions? You mean distortions due to bad data collection on population by age groups and sex, or urban/rural, and other demographic/geographic attributes? If so, why would that be distorted by the pandemic and not by the procedures at place? Would that be more distorted than the 2010 data we are using, since no 2015 population count was carried out? And/or you mean distortions by income attributes for sampling frame? If so, I would say that maybe sampling frames should not be based on household income attributes in the first place. We need sampling frames that are not so dependent on an attribute that is so changeable, not only in pandemic times. Maybe the pandemic can help data producers, statisticians and data analysts to be more open and try new methodologies and make use of alternative datasets. My experience says all administrative data only improve in quality with continuous use, so, let’s use parsimoniously and efficiently so it can be of better quality.

    Me gusta

    • Hi Suzana,
      Thanks for your comments.
      By coming months I mean conducting a census in 2021 or 2022.
      By “distorted snapshot” I refer to all the conditions that are temporary due to the pandemic and the measures to restrict the circulation of people.
      I agree that this may be a good opportunity to innovate, though, as I said in the post, last-minute changes are always risky.

      Me gusta

  2. Pingback: Ao lado de Botsuana e Índia, Brasil está no grupo de países sem previsão para o Censo - Geocracia


Introduce tus datos o haz clic en un icono para iniciar sesión:

Logo de

Estás comentando usando tu cuenta de Cerrar sesión /  Cambiar )

Google photo

Estás comentando usando tu cuenta de Google. Cerrar sesión /  Cambiar )

Imagen de Twitter

Estás comentando usando tu cuenta de Twitter. Cerrar sesión /  Cambiar )

Foto de Facebook

Estás comentando usando tu cuenta de Facebook. Cerrar sesión /  Cambiar )

Conectando a %s