·The Hindu·15 marks·250–350 words

The integrity of a digital Census depends as much on its classification frameworks as on its technology. Discuss in light of recent controversies over religion and caste categories.

New · District CollectorCan you hold the district for 72 hours?Floods, a heatwave, an outbreak, a grievance queue. Take the decisions, see the consequences.Play →
In this answer
  1. Technology: necessary, not sufficient
  2. Religion: when a code list becomes a prompt
  3. Caste: no clean choice
  4. Trust: the binding constraint

Census 2027 is India's first digital Census. It offers a self-enumeration (SE) portal in 16 languages [2] and is the first since Independence to count caste [1]. Recent controversies show that software can only record what its classification allows. A flawed code list gives flawed data, however modern the platform.

Technology: necessary, not sufficient

  • SE IDs and digital forms reduce transcription errors and processing delays [2].
  • Enumerators verify self-entered data against the same back-end lists, so a design error carries into both modes.

Religion: when a code list becomes a prompt

  • The SE portal briefly showed 287 "Other" religion options, including ISKCON, Shia, Sunni and Pentecostal. These mix sects, denominations and organisations [3].
  • An internal coding dictionary is normal, but a visible drop-down suggests answers. A Sunni respondent who picks "Sunni" may drop out of the Muslim count.
  • The list was withdrawn silently, without public notice. This left two versions of the form and drew charges of "backdoor software changes" [3].
  • Category design also shapes tribal self-identification, as the long-standing Sarna code demand shows.

Caste: no clean choice

  • Caste is an open field with no drop-down. OBC groups want a drop-down of all castes in Phase 2 [3].
  • Free text produced spelling variants and sub-caste entries in SECC 2011. Its caste data was never fully released.
  • A drop-down needs one validated master list, which India lacks because Central and State OBC lists differ. A rigid list risks exclusion errors; a loose one risks inclusion errors.
  • The stakes are high. The data feeds delimitation (Art. 82), reservation (Arts. 330, 332, 243D) and the debate on the Indra Sawhney 50% ceiling.

Trust: the binding constraint

  • The Centre put caste in the main Census rather than a separate survey to protect the social fabric from political pressure [1]. That aim also needs transparent categories.

So the integrity of the Census rests on technology, classification and transparency together. The RGI should publish the religion and caste code lists and consult States and communities before Phase 2 in February 2027 [2]. It should also keep a public change log for the portal. A credible classification framework would make Census 2027 a tool of evidence-based social justice, true to the constitutional promise of equality.

Sources

  1. 1PIB: Cabinet approves Caste enumeration in the upcoming Census (30 April 2025)CCPA decision; caste counted in the main Census rather than a separate survey, to protect the social fabric
  2. 2PIB: Census 2027 – India's First Digital Enumeration Exercise (April 2026)first digital Census; SE portal, SE ID, 16 languages; Phase 2 Population Enumeration in February 2027
  3. 3The Hindu: "Opposition, OBC group voice concerns on 287 'other' religions in the Census" (17 Nov 2026)287-option religion drop-down and its silent removal; "backdoor software changes" charge; open caste field and the OBC demand for a drop-down *(the exact article page could not be checked because thehindu.com blocks automated access, so this links the domain root)*

More from this note