Bangladesh needs independent census evaluation
Population and housing census in any country is critical as it is the source of reliable, updated, and comprehensive data about a population's size, distribution, and structure. The quality of census data is crucial for many reasons, including building public trust and understanding the national statistical system. Thus, the government needs to ensure that the census provides accurate, result-giving data that is required for future plans and policies.
The main objective of the census is to count all the population and households residing within a nation's territory at a point in time, from where the socioeconomic and demographic characteristics of the people and the households are reflected. The Bangladesh Bureau of Statistics (BBS) under the planning ministry conducted the sixth decennial and first digital population and housing census of Bangladesh on June 15-21, 2022, which was extended for a week for data collection in the flood-affected areas of Sylhet, Sunamganj, Moulvibazar and Netrakona districts. The Computer-Assisted Personal Interviewing (CAPI) process was used for the digital census.
It is good to state that, for the first time, the BBS was quite fast to release the preliminary report within a month of the census operation. Still, questions have been raised from different corners in terms of the figure of the total population of the country (165,158,616). There are also a few striking features reflected in the findings, like the sex ratio (98 males per 100 females) for the first time in history, floating population (only 22,185), slum population (1,800,486), partial information of households (17,507) or partial information of population (85,957), residents of Dhaka city corporations (10,278,822), etc, which drew massive attention on social media as well as print and electronic media over the exclusion of people on counting. Currently, the preliminary report lacks information on the size of the population living abroad or remittance-receiving households, number of foreigners, etc, but it is expected to include such important information later in the final report.
It is well-recognised that a population census is not perfect. No one can claim that a census is error-free. The errors can be categorised into two types: coverage errors and content errors. Coverage errors occur due to omissions or duplication of persons or housing units in the census enumeration, whereas content errors are those that arise in the incorrect reporting or recording of the characteristics of the persons, households, and housing units enumerated in the census. In this regard, many countries, including Bangladesh, have recognised the need to evaluate the overall quality of their census results. Generally, census errors are omissions, duplication, erroneous inclusions, and gross versus net errors. It is crucial to conduct the post-enumeration check (PEC) following the operation guidelines of the United Nations Principles and Recommendations for Population and Housing Censuses (Rev 2), mainly for an independent evaluation of the quality and coverage of the primary census conducted. Independence, quality assurance, control of non-sampling errors, etc are good PEC data collection programme attributes.
As the Bangladesh Institute of Development Studies (BIDS) has been assigned to conduct the PEC, they should be meticulous and cautious in this regard. They should be given complete independence to conduct the PEC as its purpose is to give users confidence when utilising the data and to explain errors in the census result. As a result, field enumerators and supervisors in the PEC should not be BBS employees, but people from the BIDS. Considering the flood in the localities and other matters, the enumeration areas (EAs) should be determined and kept secret from all field staff. The current census is digital so that the PEC can be done within a limited time interval.
For a census to be successful, three things need to be looked at: how the data was collected, how it was compiled, and whether it was published or not. First, we need to look into how the data was collected. The BBS decision to conduct the census during the rainy season was not prudent. Floods in the northern regions meant data could not be collected on time. This was not a population but also a housing census; what happens when people's houses are washed away by flood? There was also the Cumilla city election, which coincided with the census. Given the above scenarios, were we able to count all the people and the households accurately?
The census needs to answer questions about both individuals and households, like the number of people per household, age and sex, the type of households they live in, the kind of electricity they use, education level, access to water, sources of fuel, sources of drinking water and types of sanitation, if anyone is abroad, remittance received in the last two years, the number of foreigners in the country, etc. There were 35 questions, and we need to examine if all of them were asked. If not, the government needs to know how the data collectors acquired the information from unasked questions. What is the female household response rate during the enumeration? In this regard, the PEC is important, because it will show how many people have been left behind in the primary census and how they can be considered to adjust the total size of the population. For example, in 2011, our fifth census primarily showed that we had a population of 144,043,697. After the PEC, the population increased to 149,772,352, including an additional 5,728,655. So now, we see the primary census as a first step.
The second step will be the PEC conducted by the BIDS as a completely independent agency to ensure public trust under the demanding situation. Then the BBS will complete the socioeconomic and demographic survey on a sample population. The PEC will reveal how well the census was conducted. We need it to be of quality, and it has to be correct. We cannot work with a faulty census, so the quality cannot be compromised. The PEC will adjust the errors and give a more precise picture. If we don't have the correct data, we won't know the ground reality to gauge our progress towards achieving the Sustainable Development Goals (SDGs) by 2030 and beyond, and to undertake the right policies and plans, seeing as the population is at the centre of development.
Dr Mohammad Mainul Islam is a professor and former chairman of the Department of Population Sciences at Dhaka University.