Free-text versus pre-coded responses: critically evaluate the methodological choices in enumerating caste and religion in India's Census.
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 →Census 2027 is India's first digital Census with self-enumeration [2][3]. It will count caste for the first time since 1931, after the CCPA's decision of 30 April 2025 [1]. Whether answers are captured as free text or chosen from a pre-coded list will decide whether this data can actually be used.
Free-text responses
- Merits: respondents describe their identity in their own words. This avoids exclusion errors, because India has no single national OBC list; the Centre and each State keep separate lists.
- Demerits: it produces spelling variants, sub-castes and clan names, all of which need heavy post-enumeration coding. SECC 2011 used open entries, and its caste data has still not been released [4].
Pre-coded drop-downs
- Merits: answers are standardised and machine-readable, so processing is faster. This suits a digital portal.
- Demerits: a drop-down is only as good as its master list. The self-enumeration portal briefly showed 287 "Other" religions. The list mixed sects (Shia, Sunni), denominations (Pentecostal) and organisations (ISKCON), and it was removed without notice after media queries [5]. A list on screen also suggests answers: a Sunni respondent who picks "Sunni" may not be counted as Muslim.
Evaluating Census 2027's choices
- Inconsistency: religion got a long drop-down, but caste, where standardisation matters most, got an open field. This led the All India OBC Students' Association to demand a caste drop-down in Phase 2 [5].
- Positive: counting caste within the main Census rather than in a separate survey adds legitimacy [1]. It does not, however, solve the coding problem SECC faced.
- Transparency deficit: the code lists are not public, and changes are made silently. Critics compared this to the SIR of electoral rolls [5]. The stakes are high because Census data drives delimitation, reservation and the debate on the Indra Sawhney 50% ceiling.
Way forward
- Hybrid design: a State-wise validated drop-down plus an "Other (specify)" free-text field.
- Publish the religion and caste code lists before Phase 2 (February 2027) [2], so communities can flag missing or wrong names.
- A public change log for the portal, and expert review of the classification.
Neither method works well on its own: free text loses comparability, and pre-coding loses inclusivity. A transparent hybrid design that is open to scrutiny can make Census 2027 a credible base for evidence-based social justice and the equality promised in Articles 14–16.
Sources
- 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
- 2PIB – Census 2027: India's First Digital Enumeration Exercise (April 2026)first digital Census; caste enumerated in Phase 2 (February 2027)
- 3PIB – Self-Enumeration for Phase II of Census 2027 Commencesself-enumeration facility for Census 2027
- 4PIB – Socio-Economic Caste Census (Lok Sabha reply, 2022)SECC 2011 caste data not released
- 5The Hindu – "Opposition, OBC group voice concerns on 287 'other' religions in the Census" (17 Nov 2026, news report)287-religion drop-down and its silent removal; OBC group's caste drop-down demand; SIR comparison