Compare the 'open column' and 'pre-defined list' approaches to caste data collection, citing examples from SECC 2011 and Bihar's caste survey.

Census 2027 will be the first Census since 1931 to enumerate the caste of all Indians, and the first fully digital one [1]. The Phase-II rehearsal (6–20 July 2026, 16 States/UTs) trialled an "open column" where respondents write in their caste [2], reviving an old methodological choice — self-declaration versus selection from a fixed list.

The 'open column' approach — SECC 2011

• Method: enumerator records caste as stated by the respondent; no fixed menu.
• Merit: captures self-identity, sub-castes and unlisted communities; avoids the State pre-judging categories.
• Demerit: SECC 2011, run by the Ministry of Rural Development outside the Census Act, 1948, generated roughly 46 lakh distinct caste entries through duplication, spelling variants and sub-caste fragments — the caste data was never officially released [3].

The 'pre-defined list' approach — Bihar 2022-23

• Method: respondents choose from a State-specific coded list of castes.
• Merit: clean, machine-readable, directly comparable data; Bihar's survey (data collection from 7 January 2023) yielded usable shares — EBCs about 36% and OBC+EBC roughly 63% of the State's population, released 2 October 2023 [4].
• Demerit: an incomplete list can exclude or misclassify communities and freezes contested identities; it is also a State survey, not a Census exercise.

Points of comparison

The trade-off is thus between fidelity to self-identification and usability for policy. A hybrid path is workable: retain SC/ST recording through existing community codes, allow an open write-in for others, and have the ORGI reconcile entries against State OBC/NCBC lists using digital, standardised classification — the safeguard SECC 2011 lacked [1]. With the final questionnaire due by September 2026 [2], sound methodology will decide whether caste data strengthens evidence-based social justice under Articles 15(4) and 16(4).