The India Power Atlas
Part I · Framework
Chapter 4

Data, Coding and Limitations

The Atlas rests on 16,303 research rows, each a coded observation with its source, and 10,089 money flows. Rows include, for example, a constituency result, a television channel's audience and audited content lean, a university's vice-chancellor and the circumstances of the appointment, a court ruling and its outcome, or a grant from a foundation to an Indian organisation. Every row is assigned to a side under a written classification codebook and graded for the quality of its source.

4.1Source grades

Sources are graded from A to E. Grade A denotes official, filed or audited sources (statutes, gazettes, Election Commission and registrar filings, court judgments, answers in Parliament, annual reports, FCRA records); B, peer-reviewed or systematic investigation; C, reputable reporting with documents; D, a single or partisan source, of either side; and E, unverified or anonymous material. Wikipedia is not treated as a source, only as a route to the sources it cites. Of the graded rows, 15,768 of 23,578 (67%) are grade A. Grade distributions are reported in each chapter.

4.2Coding

Alignment is coded from the documented public record: statements, affiliations, funding, editorial output and votes. It is never inferred from an individual's religion, caste or name. Content is coded on samples drawn and audited under the codebook. Where evidence is insufficient, units are recorded as unclassified and excluded from the two-sided share rather than forced to a side.

4.3Limitations