— ✍️ Aniruddh Singh Vidrohi
Counting caste is only the beginning. The real test is whether India can convert millions of self-declared identities into accurate, comparable and policy-useful data.
India is entering an important moment in its demographic and public-policy history. The Government of India has decided to include caste enumeration in the second phase of Census 2027, and the Office of the Registrar General & Census Commissioner, India (ORGI) notified the Population Enumeration questionnaire on 14 August 2026.
The decision to enumerate caste is significant. But an equally important question now deserves serious national attention: How will caste be counted accurately?
Counting is not merely a matter of asking a question. A modern census must transform individual responses into reliable, comparable and statistically usable information. That requires identification, coding, classification, mapping, verification, quality control and transparent statistical methodology.
The notified Population Enumeration questionnaire contains Question 10, headed “Scheduled Caste (SC)/Scheduled Tribe (ST)/Caste.” For persons who are not SC/ST in the State/UT, the caste name is entered in an open field; the schedule also provides options including “Does not want to declare Caste” and “No Caste.” SC and ST respondents are handled through the applicable State/UT list rather than the same open field.
The notified questionnaire does not contain a separately labelled OBC field or checkbox. This should not be misrepresented as proof that OBC information can never be derived from caste data. The legitimate methodological question is: what transparent, scientifically verifiable process will be used to identify and classify OBC/SEBC communities from the caste responses?
This is not an argument that every respondent should automatically be classified as OBC. It is an argument for clarity in the statistical architecture and for a publicly understandable methodology before the resulting data are used for major policy decisions.
India is one of the world’s most socially, linguistically and regionally diverse societies. A community may be reported through different regional names, synonyms, alternative spellings, transliterations, sub-group names, traditional or local names, and phonetic variations.
This does not mean self-identification is wrong. Self-identification is important. The challenge is what happens after self-identification.
If two people identify the same community using different spellings or regional names, the statistical system must recognise the relationship between those responses. At the same time, different communities must not be incorrectly merged merely because names or surnames appear similar.
In July 2015, the Government of India reported that the Socio-Economic and Caste Census (SECC) 2011 had produced 46,73,034 distinct caste names. The Government said these included caste/sub-caste names, synonyms, surnames, clan/gothra names, phonetic variations, sections and sub-groups. The Government also reported 8,19,58,314 errors in caste particulars communicated to States/UTs for rectification. Of these, 6,73,81,119 had been rectified, while 1,45,77,195 remained to be rectified at that stage.
These figures should not be interpreted as an argument against caste enumeration. They demonstrate something more important: caste enumeration requires a robust data architecture.
A searchable Caste Master List, supported by standardised codes and regional-name mapping, could improve consistency. A sound architecture could follow this chain:
Self-declared response → Standardised caste code → Synonym/regional-name mapping → Verified statistical classification
No person should be forced into an incorrect category simply because the system cannot recognise the response. If a declared caste is not immediately matched with the available Master List, the system should provide a clearly defined UNMATCHED/PROVISIONAL ENTRY mechanism, followed by appropriate verification and coding.
Scientific standardisation does not mean reducing India’s social diversity into a small number of labels. It means creating a transparent relationship between what a person declares and how that declaration is statistically coded. The original response should remain traceable within the protected data architecture, while statistical coding makes aggregation and analysis possible.
Why OBC Identification Dseserves Particular Clarity
The debate is not only about counting individual caste names. It is also about how the resulting data may eventually be used to understand social disadvantage, representation, equality of opportunity and public policy.
The absence of a separately labelled OBC field in the notified question does not by itself prove that OBC information cannot ultimately be derived or classified. But it raises a legitimate methodological question: What transparent and scientifically verifiable process will be used to identify and classify OBC/SEBC communities from the caste responses?
The methodology should be sufficiently transparent that researchers, policymakers and citizens can understand how caste responses are coded; how synonyms and regional names are mapped; how sub-groups are treated; how OBC/SEBC classification is determined; how unmatched responses are handled; how errors are corrected; and how final data quality is independently assessed.
The Government has stated that Census 2027 will be conducted digitally and that a separate Code Directory will be provided for several descriptive/non-numeric questions in the second phase, including Scheduled Caste/Scheduled Tribe and other fields. It has also described validation checks and supervisory monitoring as part of the Census architecture. That is a positive step forward.
The objective should now be to ensure that the caste component receives the same level of methodological robustness. The question is not whether India should use technology—it should. The question is how transparent, comprehensive and independently verifiable the caste coding and mapping system will be.
Wrong identification → Wrong coding → Wrong counting → Less reliable data → Greater difficulty in evidence-based policymaking.
A scientifically robust system should follow:
RIGHT IDENTIFICATION → RIGHT CODING → RIGHT COUNTING → RELIABLE DATA → BETTER ANALYSIS → EQUAL OPPORTUNITY → SOCIAL JUSTICE
Clear caste self-identification — Every person should be able to declare their caste without being forced into an inappropriate predefined category.
Standardised caste codes — Each recognised statistical category should have a consistent and traceable code.
Synonym and regional-name mapping — Different names referring to the same community should be systematically mapped where appropriate.
Unmatched/Provisional Entry — Unmatched responses should not disappear or be forced into an unrelated category.
Transparent OBC/SEBC methodology — The classification methodology should be clearly documented.
Correction and verification — There should be a defined process for correcting coding/classification errors.
Independent data-quality audit — The final statistical dataset should receive appropriate independent quality assessment.
Post-enumeration verification — A scientifically designed verification exercise should test reliability.
The demand for accurate caste data should not be viewed through the lens of one caste, political party, organisation or ideology. It concerns the entire Indian population. Every community has an interest in being correctly identified, correctly counted and accurately represented in official data.
A credible caste census should not be designed to increase or decrease the numerical strength of any community. Its purpose should be to document the facts as accurately as possible. Public policy, social justice and equal opportunity require evidence—and evidence requires reliable data.
Census 2027 is an opportunity to demonstrate that a large and complex social-data exercise can be conducted with high standards of technology, statistical discipline and transparency. The Government’s decision to include caste enumeration is historically significant. The next step should be to ensure that the methodology behind the counting is as credible as the decision to count.
India should not have to choose between self-identification and scientific standardisation. It can and should have both. The respondent should have the freedom to declare their identity. The statistical system should have the scientific capacity to code that identity accurately. The process should be sufficiently transparent to command public confidence.
Count Everyone. Identify Everyone Correctly. Code Every Response Scientifically. Verify The Data. Protect Every Ccommunity's Identity.
A successful caste census will not simply be one in which caste information is collected. It will be one in which the collected information can withstand statistical scrutiny, methodological examination and public trust.
Accurate data is the foundation of evidence-based policy.
- Aniruddh Singh Vidrohi is the National President, Mandal Army, Mainpuri, Uttar Pradesh