How the assistant answers
The assistant answers from the same public record as the rest of the site — it does not browse the open web at answer time. Every figure, legal rule, contact and named allegation it states is meant to come from a tool result gathered for that question: federal Open Treasury payment reports, the accountability graph, the 1000 Reasons archive, and loaded national datasets (budgets, contracts, FAAC revenue, debt, audit findings, inflation, health and school registers, and others). When a query returns nothing, the assistant is instructed to say so rather than fill the gap from general knowledge.
- Each answer shows a Sources used line you can expand into an evidence drawer — the actual records, entities and archive items behind it, each with a label and a link to open the record.
- The data is batch-loaded, not live. Treasury figures run to the latest published daily report; several datasets have explicit coverage gaps and as-of dates, which the answer should name.
- The assistant does not adjudicate. It is civic information, not legal advice; for anything that turns on your specific facts, it points you to a lawyer, the Legal Aid Council, or the relevant body.
- Absence from these records is not proof of innocence — only that we have not surfaced a documented public record here.
What the labels mean
So a reader never mistakes “mentioned in a filing” for “found by a court”, evidence carries one consistent set of labels across answers and source cards:
- Recorded — Present in a dataset or public record.
- Alleged — Claimed in a filing, report or source — not, on its own, proven.
- Flagged — Marked by a rule, pattern or archive classification — a lead, not a verdict.
- Confirmed — Supported by an authoritative or legal outcome.
- Unverified — Mentioned, but with insufficient evidence in the records here.
- No match — No relevant internal record found — absence is not proof of innocence.
The archive’s own status taxonomy (alleged, arraigned, convicted, under investigation, and so on) maps onto these: a charge is still Alleged until there is an outcome; a conviction, acquittal or plea is Confirmed; an investigation or pattern flag is Flagged — a lead, not a verdict.
Data sources
Entries draw from public court records, agency reports (EFCC, ICPC, Office of the Auditor-General for the Federation, INEC, the National Assembly), investigative journalism by Nigerian and international outlets (Premium Times, SaharaReporters, BBC Hausa, Reuters, The Continent, and others), and contemporaneous public statements by the named individuals or institutions.
Each source is cited with its publisher, date, and URL. Where a source has since been removed from its original location, we link to the Internet Archive snapshot.
Verification process
- Each named individual is confirmed against at least two independent public sources before being added.
- Each cited amount, date, court name, or case number is verified against the original document where possible.
- Each entry is reviewed by at least one editor besides the author before publication.
- Status tags are assigned per the policy in our editorial documentation. The default tag is alleged; stronger tags require court proceedings or formal investigation.
Who reviews entries
Every entry is written by one contributor and reviewed by at least one other editor before it is published. The reviewer checks each cited amount, date, court name, and case number against the linked source, and confirms the legal-status tag matches the most recent ruling or filing on record. Reviewers are recorded against each entry, and the date an entry was last reviewed is shown on the entry itself.
Status tags are not assigned by the author alone. A stronger tag than alleged — arraigned, convicted — requires a formal court document or agency record in the citations, verified at review.
Independence & funding
1000 Reasons is not affiliated with any political party, campaign, candidate, or government body, and accepts no funding from any of them. Both ruling and opposition figures appear in the archive on the same terms and against the same evidentiary floor. The project takes no advertising and runs no paywall.
The standard is the same regardless of who an entry names: cite the public record or omit the claim. Where we get something wrong, we correct it in full and preserve the original in the audit log (see below).
Inclusion criteria
Allegations require at least two independent media sources or one formal record (court filing, audit, agency press release). Single-source social-media accusations are not enough. Family members of named individuals are not separately listed unless they have a documented role in the matter.
Limitations
- Absence from the archive does not imply absence of wrongdoing; it implies absence of documented public evidence we have surfaced.
- Court rulings can be overturned. The status tag reflects the most recent ruling we have on file as of the date listed on the entry.
- Untranslated sources may carry nuance lost in the English summary. When we summarise a non-English source, we link to the original.
- The archive does not adjudicate. We present the record and the citations; the reader judges.
Right of reply
If you are named in an entry and the record is wrong, write to corrections@1000reasons.vote or use the corrections form. We acknowledge within 48 hours. Accepted corrections are published in full; the original entry is preserved in the audit log.