tnfacts
Menu

Methodology

Notes on how sources are chosen, how disagreements between them are handled, and what is and is not done to the published numbers.

Source authority

Sources are ranked into five tiers, based on institutional independence and statutory authority:

  1. Tier 1. Constitutional and statutory bodies: RBI, CAG, ECI, RGI (SRS, Census).
  2. Tier 2. Central government ministries: MoSPI, NCRB, MoHFW, MoE.
  3. Tier 3. Major surveys and institutional research: NFHS (IIPS), UDISE+, ASER (Pratham), AISHE.
  4. Tier 4. State government: TN DES, departmental policy notes.
  5. Tier 5. Private and academic: CMIE, Indiastat, ISB, NIPFP.

When sources disagree

More than one authoritative source often publishes different values for the same indicator. NFHS and HMIS report different IMR figures; MoSPI, RBI, and CMIE report different GSDP figures. In those cases the page shows:

  • A primary value, taken from the highest-tier source whose methodology most closely matches the indicator definition.
  • Up to three alternates behind a disclosure on the datapoint, with the percentage difference from the primary and a short note on why they differ.
  • Methodology and reference period for each, so readers can judge for themselves which figure to use.

Series breaks

When a source's methodology changes mid-stream (for example, GSDP being rebased to 2011-12 prices in 2015), the break point is annotated on the chart. The old and new series are not spliced by default; they are shown as two segments. Splicing is done only when a specific method is documented in the data point's notes.

Refresh cadence

Data is refreshed quarterly through a reviewed update process. Annual sources (NCRB, RBI Handbook, UDISE+) are picked up whenever a new release lands. Every change is attributed and timestamped; the change history for any indicator is recorded in version control.

What is not done

  • No new estimates are constructed from raw data. Every value here was already published by a primary source.
  • No editorialising appears on data pages. Any analysis lives in separate write-ups.
  • Aggregations across incompatible methodologies are flagged, not silently combined.