1 What counts as a trackable promise

The same test is applied to every president and governor, so the counts on one scorecard mean the same thing as the counts on another.

A promise being hard to assess is not a reason to leave it out — it's a reason to rate it In progress and keep checking.

2 Promise ratings

Every tracked promise — and every entry in Inherited Fixes — gets one of four ratings, based on the evidence linked to it.

When sources disagree, the more independent and more recent one is preferred, and the disagreement is noted in the assessment. Absence of evidence is not treated as failure: if nothing is verifiable either way and the deadline hasn't passed, the rating stays In progress; if the deadline has passed with no sign of action, that is Broken.

3 Status terms by category

"Kept vs. broken" doesn't fit a court case or an appointment, so every other category has its own status vocabulary.

Each fraud case also carries a separate government response rating — Pursuing, Stalled, Politicised, Abandoned, or No interference — describing how the government itself has handled the allegation, independent of the case's legal outcome.

4 Budget terms

Where a figure isn't publicly available, the field is left blank rather than estimated — see Sourcing & review.

5 Other terms

6 Sourcing & review

Every entry — a promise, a fraud case, a budget line, an order — is dated and linked to its original source. Nothing is rated without a linked reference, and ratings are a judgment call made against that evidence, not an automated calculation.

NGScorecard doesn't publish estimates dressed up as facts. Where an official figure hasn't been published — an unreleased state budget line, for instance — the field is left blank rather than filled with a computed guess. Blank means "not published," not "zero."

Each administration carries a reviewed date: the last time every entry under it was re-checked against current evidence. Sitting administrations are re-reviewed on a rolling basis — at least twice a year, and after major events such as a budget, an election, or a significant court ruling. Former administrations stay frozen at their final review unless new evidence emerges.

If you think an entry is wrong, outdated, or missing context, use the "Report an issue" button on that entry's card — it opens a short form (what's wrong, a better source if you have one, an optional way to reach you) that goes straight into a moderation queue. If you'd rather not use the form, you can email mubaraqsanusi908@gmail.com instead. Either way, corrections are made promptly and noted on the entry. If this methodology itself changes in a way that affects existing ratings, those ratings are re-checked against the new standard within one review cycle, and the change is recorded in the change log below.

7 Coverage & gaps

Not everything is tracked to the same depth. This section says what is covered, how far back, and what is deliberately or unavoidably thin — so a short scorecard is read as a limit of the record, not a verdict.

Depth follows the evidence. A recent administration has a contemporaneous paper trail; a 1999–2003 state government often does not, and a 1979–1983 or 1992–1993 one rarely does. Where the public record only supports a handful of checkable entries, that is all that is published — an empty section is better than one padded with unsourced claims. In particular, budget and economic-indicator series for most states before about 2010 were never published in a form that can be cited, so those sections are blank for every pre-1999 governor and many later ones. The Third Republic scorecards are uniformly short: that tier of government lasted under two years before it was dissolved.

A source review in September 2026 found and replaced a large number of dead or unverifiable links across the older records. Some were replaced with encyclopaedic or archived references while the underlying primary documents are tracked down. Entries still resting on a weaker source (see the hierarchy in section 6) are being worked back up it; a card's source is always the current best available, and "Report an issue" is the fastest way to point us at a better one.

8 Independence & openness

A tracker that grades politicians is only worth reading if the grades aren't for sale — and it's more useful to everyone if the work is free to read, check, and reuse.

9 Change log

Material changes to how promises are selected or rated are recorded here, newest first. Wording fixes and clarifications that don't move any rating are not logged.