# Source Coverage and Methodology

This reference documents the scope, source provenance, variable definitions, and known limitations of the **Electoral Research Studio — Lok Sabha Statistical Archive (1951–2024)**.

## 1. Archive Scope & Election Cycles

The archive is compiled directly from **47 official Election Commission of India (ECI) statistical report volumes** and **5 official ECI reference tables** (`data/reference/`), covering **18 nationwide Lok Sabha General Elections** (1st through 18th Lok Sabha, 1951–2024) and **2 delayed-election supplementary volumes** (`1985` for Assam and Punjab following the 8th Lok Sabha election of 1984, and `1992` for Punjab following the 10th Lok Sabha election of 1991).

Across these 20 report volumes, the database preserves:
- **9,353** parliamentary constituency (`PC`) contest records (`9,313` in the 18 main general elections + `40` in the 2 delayed supplements)
- **99,593** candidate contest rows
- **8,235** parliamentary assembly-segment (`AC`) records (`1999` and `2004` Vol. III, representing assembly segments nested inside Lok Sabha constituencies rather than standalone State Legislative Assembly / Vidhan Sabha elections)

| Report Year | Lok Sabha | Cycle Type | Parliamentary Records | Candidates | Assam PCs |
| --- | --- | --- | ---: | ---: | ---: |
| 1951 | 1st Lok Sabha | General Election | 401 | 1,874 | 10 |
| 1957 | 2nd Lok Sabha | General Election | 403 | 1,594 | 10 |
| 1962 | 3rd Lok Sabha | General Election | 494 | 1,985 | 12 |
| 1967 | 4th Lok Sabha | General Election | 520 | 2,369 | 14 |
| 1971 | 5th Lok Sabha | General Election | 517 | 2,783 | 14 |
| 1977 | 6th Lok Sabha | General Election | 542 | 2,439 | 14 |
| 1980 | 7th Lok Sabha | General Election | 529 | 4,629 | 2 |
| 1984 | 8th Lok Sabha | General Election | 514 | 5,312 | 0 |
| 1985 | 8th Lok Sabha | Delayed Supplement (Assam & Punjab) | 27 | 180 | 14 |
| 1989 | 9th Lok Sabha | General Election | 529 | 6,160 | 0 |
| 1991 | 10th Lok Sabha | General Election | 521 | 8,668 | 14 |
| 1992 | 10th Lok Sabha | Delayed Supplement (Punjab) | 13 | 81 | 0 |
| 1996 | 11th Lok Sabha | General Election | 543 | 13,952 | 14 |
| 1998 | 12th Lok Sabha | General Election | 543 | 4,750 | 14 |
| 1999 | 13th Lok Sabha | General Election | 543 | 4,648 | 14 |
| 2004 | 14th Lok Sabha | General Election | 543 | 5,435 | 14 |
| 2009 | 15th Lok Sabha | General Election | 543 | 8,070 | 14 |
| 2014 | 16th Lok Sabha | General Election | 543 | 8,251 | 14 |
| 2019 | 17th Lok Sabha | General Election | 543 | 8,054 | 14 |
| 2024 | 18th Lok Sabha | General Election | 542 | 8,359 | 14 |

## 2. Numerical & Methodological Definitions

- **Valid Votes vs. Summed Candidate Votes**: `validVotes` is the PC summary figure printed in the ECI report; `candidateVotes` is the exact sum of candidate votes excluding `NOTA`. Where the two differ (notably in 2009 due to rejected postal ballots), both numbers are preserved and flagged rather than silently overwritten.
- **Candidate Vote Share (`votePct`)**: Calculated as `100 * votes / candidateVotes` (excluding `NOTA`). `reportedVotePct` separately retains the percentage printed in the original ECI table.
- **Effective Number of Parties (`enp`)**: Laakso & Taagepera's index (`1 / Σ p_i^2`), grouping candidates by party abbreviation and treating each Independent (`IND`) candidate as a distinct unit. `NOTA` is excluded.
- **Vote–Seat Disproportionality (`Gallagher`)**: `sqrt(0.5 * Σ (V_i - S_i)^2)` calculated across all contesting party and individual independent units in single-member contests before small parties are grouped into a visual "Others" bar.
- **Deposit Forfeiture (`forfeitedDeposit`)**: A derived statutory threshold flag (`votes * 6 * seats <= candidateVotes` for non-winning candidates under [Section 158(4) of the Representation of the People Act, 1951](https://www.indiacode.nic.in/bitstream/123456789/2096/9/A1951-43.pdf)), not an official ledger of deposit refunds.
- **Candidate Sex (`sex` / `gender`)**: Canonical `F` (Female), `M` (Male), `O` (Third Gender), or `U` (Unknown / unrecorded in historical tables). Women's representation metrics (`WOMEN` lens) overlap `GEN` / `SC` / `ST` seat reservations and never form a fourth mutually exclusive reservation partition.
- **Party Identity & Classification**: `partyDictionary` maps 1,966 party codes across 1951–2024 using historical ECI party lists and same-cycle ECI symbol tables (`data/reference/party-symbols-2009.pdf`, `2014.pdf`, `2019.xls`, `2024.xlsx`). Recognition status (`National Party`, `State Party`, `Registered Unrecognised Party`) is election-specific and is never carried forward or backward across cycles. Where a single ECI report lists two different party names under the same abbreviation (`ASP` in 2014; `CPIM` in 2019), both alternatives are retained and a single resolved name is withheld.

## 3. Source Exceptions & Known Archival Limits

- **2024 Uncontested Surat PC**: The supplied 2024 ECI statistical tables include 542 contested seats and omit uncontested Surat (`Highlights(Surat PC)`). No synthetic vote or electorate counts are invented for Surat.
- **1967 Damaged Source Page (New Delhi & South Delhi)**: Page 187 of the 1967 Volume I PDF is physically damaged in the original scan. Recoverable text is retained; unreadable candidate votes and party codes remain `null` / `UNKNOWN`, and constituency-level ENP and margins are withheld for those two records.
- **1951 Multi-Member Constituencies**: Unlabelled 1951/1957 constituency reservations are recorded as `Unknown`. In 1951 North Bengal (3 seats), where the summary lists two winners, the third winner is identified by vote rank and flagged for source verification.
- **2008 & 2023 Delimitation Orders**: Seat reservations for 2009 and 2014 are verified against all 543 entries of the ECI Consolidated Delimitation Order of 2008 (`data/reference/delimitation-2008.pdf`: 412 GEN, 84 SC, 47 ST). Cross-cycle lineage links (`studio/crosswalk.js`) identify statutory boundary transitions (1951, 1956/1961, 1967, 1976, 2008, and Assam 2023) and withhold automatic cross-delimitation swing claims.