Public trust standard

Editorial Standards and Corrections

These standards explain how Think Politically separates verified facts from analysis, estimates, examples, and client-confidential work. They also explain how readers can report an error and how substantive corrections are disclosed.

Effective: 29 July 2026 Owner: Think Politically editorial team Review cycle: At least annually

Scope and responsibility

These standards apply to public research, election explainers, data pages, articles, comparisons, and case studies published on ThinkPolitically.com. Commercial service copy is reviewed for the same basic requirements of accuracy, clarity, and non-deception.

The named author is responsible for the draft. The editorial reviewer is responsible for checking sourcing, labels, material claims, and corrections before publication or substantial republication. A byline does not transfer responsibility away from Think Politically as publisher.

Source and evidence hierarchy

For election results, turnout, constituencies, polling stations, schedules, and legal or administrative rules, we prefer primary public records. In India this commonly means the Election Commission of India, Chief Electoral Officer publications, legislation, court records, government statistical releases, and original documents from the relevant institution.

Reputable academic work and established news reporting may add context, but a secondary source should not replace an available primary record for a central numerical claim. Every material statistic should identify its period, geography, unit, and source. Provisional figures must be labelled provisional.

Facts, analysis, estimates, and examples

We distinguish four evidence types:

  • Verified fact: supported by a cited, identifiable source.
  • Analysis: an interpretation of disclosed facts or data; reasonable alternatives may exist.
  • Estimate or model: a calculated result that must state its assumptions, inputs, and material limitations.
  • Illustrative example: a teaching scenario, not a claim about an actual client or election outcome.

Case studies must also say whether they are permissioned client work, anonymised client work, public-result analysis, internal research, or an illustrative model. We do not present correlation as proof that a campaign activity caused an election result.

Review before publication

Before publishing trust-sensitive civic content, the reviewer checks the title and summary against the underlying evidence; verifies names, dates, totals, percentages, and denominators; opens key source links; confirms that caveats appear near the claim they qualify; and checks that structured data does not make a stronger claim than the visible page.

Automated tools may assist with transcription, formatting, translation, calculations, or drafting. They are not treated as sources. A human editor remains responsible for checking any assisted output against the cited material.

Corrections and updates

We correct substantive errors openly. A correction note should state what was wrong, what changed, when it changed, and the source used to verify the correction. We do not silently rewrite a material election result, research finding, attribution, or methodological limitation.

Minor spelling, grammar, formatting, and broken-link repairs may be made without a formal correction note when they do not alter meaning. A page may show a new “last reviewed” date only when its evidence or substance has actually been reviewed.

When reliable sources conflict or a final result is unavailable, we disclose the uncertainty rather than selecting the most favourable number.

Report an error

Readers can report a suspected error by emailing contact@thinkpolitically.com. Please include the page URL, the exact sentence or data point, the reason it may be wrong, and any supporting source.

We aim to acknowledge a clear correction request within two business days. Review time depends on the evidence required. If a material error is confirmed, the public page will be corrected and a dated note will be added. If we do not make a requested change, we may explain the evidence behind the existing wording.

Policy history

29 July 2026: First public edition adopted following a site-wide review of election claims, datasets, author entities, and structured data.

Chat on WhatsApp
Call WhatsApp