The ongoing Census 2027 was to be an apolitical statistical exercise to collect data on population and economic indices, but it has stirred up a hornet’s nest with the introduction of an open-ended question on caste and certain other sensitive queries in the questionnaire. The population count in the crucial second phase will collate the socio-economic data of individuals through a set of 40 questions notified by the Union Ministry of Home Affairs. Some of them are seen to be intrusive.
Why is the exercise drawing criticism?
The open-ended question on caste has become the most controversial in Census 2027 with the Opposition attacking the government for not introducing a structured drop-down menu. In the open-ended format, respondents state their caste and the enumerators record it verbatim, instead of selecting from a pre-loaded list of castes.
Raising concerns over the Census questionnaire, AICC general secretary (communication) Jairam Ramesh alleged the methodology for collecting caste data is “deliberately flawed”. He asked why the religion, date of birth, and the birthplace of the parents of the respondents were being sought, pointing out that such ‘micro questions’ did not feature in previous Census exercises.
What are the Census questions?
On August 14, the Registrar General and Census Commissioner of India through the MHA notified a set of 40 questions to be asked during the Population Enumeration phase, adding 11 questions or data fields that were not part of the Census 2011 questionnaire. Census 2011 had 29 questions.
The new questions seek information on the spouse’s name, nationality, father’s and mother’s particulars, digital literacy, permanent residential address, place of Covid-19 vaccination, number of bank accounts, passport number, driving licence availability and mobile, Aadhaar and voter ID numbers.
Do the Census questions mirror those of the contentious 2019 NPR?
Some analysts see a parallel between the additional questions and those proposed to be recorded while updating the National Population Register (NPR) in 2019. The NPR exercise was panned by the Opposition parties who alleged that it was the first step towards the creation of the contentious National Register of Citizens (NRC). NRC intends to record genuine Indian citizens and identify the infiltrators.
According to a former deputy Registrar General, some of the questions notified by the MHA may be intended to gather information for the NPR. He is of the view that details such as respondents’ names, parents’ names and particulars and parents’ addresses are unlikely to generate meaningful Census data because they do not appear to be designed for statistical tabulation.
What do experts have to say about the caste question?
Speaking to this newspaper, former chief statistician Pronab Sen said keeping an open-ended caste column means risking and compromising the Census data. “As a statistician I must say that for the purity of data, there should have been only one column ‘SC/ST/OBC/Others’. This way, the Census official could better tabulate and present the data for future use,” he said.
According to Sen, there are over 4 million castes and sub-castes in the country. “As the 2011 socio-economic survey was conducted without a drop-down menu, we were not able to separate them in any meaningful numbers,” he said.
“This debate took place earlier also and suggestions had emerged that attempts should be made only to collect entries only through SC/ST/OBC/Others and not through specific caste categories,” Sen added.
He pointed out that the Census has to be done in the shortest possible period of time to avoid duplication of response. “In 2011, the exercise took a long time to collect caste data,” he said.
What is the problem of caste spelling?
According to officials, although an individual’s OBC status will not be captured during population enumeration, the digital nature of the exercise, combined with the availability of advanced AI-driven tools, could help statisticians arrive at an aggregate estimate of the population belonging to specific castes.
However, discrepancies could arise if respondents provide different spellings for the same caste, or if a religious sect is entered as a caste, either because of ignorance or inadequate understanding of the caste system. Officials said the decision on which spellings should be treated as valid for a particular caste category will have to be taken by the government and Census authorities at a later stage.
The caste census will help examine the social, economic and educational indicators of individual caste groups. But officials cautioned that members of some castes could understate their income or other socio-economic indicators in a bid to portray their community as ‘backward’ and strengthen the claim for inclusion in the OBC category.
Statisticians, however, would be able to identify misleading or manipulated information through cross-checking, comparison of values and other indicators. If discrepancies are found, the information could be classified as ‘unreliable’ and unsuitable for use in policy formulation.
What is different this time around?
Explaining the key difference between the Socio-Economic and Caste Census conducted in 2011 and caste information being collected as part of Census 2027, the official said the latter provides confidentiality regarding an individual’s caste declaration.
Under the Census Act, individual-level information cannot be used as “legal evidence” for determining government benefits or for public disclosure. The second major difference is the earlier difficulty of compiling and separating caste data can now be addressed through AI-powered technological solutions.
Why was the drop-down menu proposal dropped?
During discussions on the population-enumeration questionnaire, officials considered categorising non-SC and non-ST respondents into ‘general’ and providing a drop-down menu option for OBC status. However, the proposal was not considered practical because the Central and state OBC lists do not always tally.
The Central list is prepared on the basis of recommendations and maintained by the National Commission for Backward Classes, whereas the respective state backward classes commission is responsible for its state list.
“Since the two lists do not completely tally, the question for Census authorities was which list should be followed - the Central or state list. Even within a state, the OBC lists maintained by the Centre and the state government differ,” the official noted.
According to officials, the structured drop-menu model may have been used in the state caste surveys conducted in Bihar and Telangana, but it may not work in a much larger and more complicated central exercise like Census 2027.
Why is the Centre facing political trouble?
The Congress pointed out that the government had in 2021 recommended a drop-down menu for selecting castes in its affidavit to the Supreme Court. Citing an affidavit filed by the Centre in the Supreme Court, Congress leader Jairam Ramesh said that on September 21, 2021, the government had categorically rejected the plea for a caste census. “However, para 12 (d) in the affidavit says the following: ‘There was no registry of caste prepared prior to the conduct of the 2011 Census. It would have been ideal for the Registry of Castes that there should have been a drop-down menu for selection of the castes, which could have made some consistent data available which can be relied upon,” Ramesh said.
What is the Telangana Model?
The Telangana Socio-Economic, Educational, Employment, Political and Caste (SEEEPC) Survey, 2024 is widely regarded as an important model for conducting a comprehensive caste survey, as it combined caste enumeration with detailed socio-economic information about households, enabling the government to understand social inequalities in a multidimensional manner.
A structured questionnaire was used to gather information on caste, sub-caste, religion, gender, age, education, occupation, employment, income, landholding, housing, assets, migration, access to government schemes and political representation. The questionnaire contained 56 main questions covering around 75 fields, making it a comprehensive socio-economic database.
Caste identification was carried out using standardised caste and community codes, with separate identification of Scheduled Castes (SC), Scheduled Tribes (ST), Backward Classes (BC) and Other Castes (OC). This helped in systematic classification and analysis of caste groups and their sub-groups.
Another feature was the attempt to cross-check and validate the collected information, thereby improving the reliability of the database. The survey was intended to provide empirical evidence for designing welfare programmes, assessing inequalities and informing policies relating to reservations and social justice.
What is the Bihar Model?
The Bihar Caste-Based Survey 2022-23 was undertaken to collect comprehensive information on the caste composition and socio-economic conditions of the state’s population. The methodology combined caste enumeration with socio-economic data collection and was designed as a state-wide household survey rather than a sample-based exercise.
The household survey was the basic unit of enumeration. Enumerators visited households across Bihar and collected information from residents using a structured questionnaire. Details were recorded for each family member, including name, age, gender, caste, religion, education, occupation, income and other socio-economic indicators. The survey covered all major caste categories, including SC, ST, BCs, EBC (extremely backward classes), etc.
A major feature of the Bihar model was its emphasis on digital data collection. Enumerators used mobile apps and electronic devices to enter information, which helped in standardising data collection and speeding up processing.