Press secretaries and agency spokespeople speak for the government. Every administration, of either party, expects them to put its record in the best light, and message discipline is a legitimate part of the job. Some of Citizen Analyst’s profiles of government communications officials carry a Spokesperson Misinformation Assessment (SMA): a structured judgment about where that official’s statements crossed from advocacy into knowing misinformation. This page sets out the method in full: what is measured, how a score is reached, how a designation follows from the scores, and what a designation does and does not claim. Corrections notices on our profiles point here.

What the assessment measures

The unit of analysis is a public statement made in the official’s communications role: a briefing answer, an official social-media post, a television appearance made as spokesperson. The assessment scores statements, not actions; documented actions by officials are the subject of a separate instrument.

Three filters apply before any statement is considered.

• The tenure window. Only statements made while the person held the role, and spoke in that capacity, are eligible. Speech before the appointment and after departure is excluded entirely. Most former press secretaries go on to paid commentary; that is private opinion, not government messaging, and it is never scored. Each profile states the dates of the window it assessed.
• An institutional fact-check. Every scored statement must have been rated false or misleading by an established fact-checking organisation, such as PolitiFact, FactCheck.org, Snopes, AP Fact Check, Reuters Fact Check or The Washington Post Fact Checker. We do not substitute our own rating for theirs. A statement with no such rating is not scored.
• The profile itself. The assessment runs only on a spokesperson profile that already documents the official’s significant statements and their fact-check record, after the subject has passed our inclusion gate and the evidence has been verified.

The first question: advocacy or misinformation

Every in-window statement goes through one gate before it can be scored. The question is whether the statement could serve legitimate message discipline or favourable spin, even one-sided, or whether it required the speaker to know the claim was false or misleading.

• Advocacy is never scored. It includes favourable framing of true facts, declining to comment, predicting a policy’s benefits, disputing an opponent’s characterisation, hyperbole, a combative tone without a false factual claim, an isolated error that was corrected or not repeated, and relaying an administration position the spokesperson did not originate.
• Misinformation is the assertion of a false or materially misleading claim, in circumstances showing the speaker knew or should have known, whose function is to confuse or to erode trust.
• Borderline statements are tested against five factors: repetition after a published correction; the same false framing pushed across channels or spokespeople; contradiction of the speaker’s own record or the agency’s own data; targeting public trust in courts, elections, the press or official data; and attacks on fact-checking itself in place of rebuttal. Three of the five make it misinformation.

Each scored statement on a profile quotes the statement with its date and venue, cites the fact-check, states which factors were met, and describes the truthful version of the same message the speaker had available.

How a statement is scored

A statement classified as misinformation receives a Misinformation Severity Score (MSS) from 1 to 5. The score measures the evidence that the speaker knew, not how widely the claim spread.

A “false” rating alone never establishes knowledge above MSS 1. A spokesperson is, however, expected to know their own agency’s published data.

How a designation is reached

A profile’s overall designation is computed from the scores the profile prints, by fixed rules, and our publishing system recomputes it before anything goes live. A profile whose stated designation does not match its own scores cannot be published. The rules are checked in order, and the first one that holds decides:

The method also allows Sustained where the record documents a claim repeated again and again after correction, a pattern a list of scores cannot fully show. A Sustained designation on that basis must still rest on scores that reach at least a Pattern, and an editor confirms it by hand before publication.

A profile with no designation still carries the assessment section, stating that no designation applies and why.

The comparative baseline

Each assessment also places the spokesperson in an ordinal band (within the norm, elevated, or an outlier) against a bipartisan reference class of press secretaries from Republican and Democratic administrations. Fact-checkers examine controversial claims rather than a random sample of statements, so the band is a signal for scrutiny, not a measured rate. We cite raw counts and never publish a “lie rate.” The band gives context; it never raises or lowers a designation.

What a designation is, and what it is not

A designation is an evaluative conclusion about communicative intent, reached by applying published criteria to documented statements. The factual predicate (that on a given date an outlet rated a given claim false) is cited and true. The conclusion drawn from it is analysis, offered openly with its reasoning.

A designation is not:

• a finding that the speaker committed a crime;
• a penalty for advocacy; vigorous, one-sided messaging for an administration’s agenda is protected message discipline;
• a judgment on tone;
• a judgment on anything said outside the role;
• a combined score. The Spokesperson Misinformation Assessment is one of six instruments, each with its own unit of analysis: a spokesperson’s statements here; a person’s actions in the Democratic Malice Assessment; public speech in the Dangerous Rhetoric Assessment; a single bill or order in the Legislative Impact Assessment; a market in the Policy Winner-Loser Map; a federal institution in the Agency Damage Assessment. They are never added together. The full list is in Scoring & Inclusion Systems — Complete Inventory.

The legal disclaimer every assessment carries

Every assessment section on a profile ends with the same disclaimer, whether or not anything was scored:

The Spokesperson Misinformation Assessment is an analytical framework applying defined criteria to documented, in-role public statements by a government spokesperson. Each score rests on a cited institutional fact-check; the Misinformation Severity Score is an evaluative conclusion about communicative intent, not a statement of criminal guilt or a finding that the speaker committed a crime. The factual predicates are cited to primary sources and institutional fact-checkers; the evaluative conclusions are protected expression. New York Times Co. v. Sullivan and its progeny apply to this assessment of a public official’s public communications.

The tenure window is part of that protection: by scoring only in-role government speech, the method stays within the public-official framework the law applies to government spokespeople. Our wider legal posture is set out in Legal Framework and First Amendment Protections.

Updates and corrections

Assessments change when the evidence does, in either direction. When a published score changes enough to change a designation, we treat it as a correction: the profile carries a dated notice, and the change appears in our public corrections log. Every correction is decided by an editor.

Anyone who believes a score or designation rests on an error of fact, a statement outside the tenure window, or a misapplication of the criteria above can request a review. The process, the address and our response commitment are on the Corrections & Transparency Policy page. Requests that identify the specific scored statement and the fact-check in question are the quickest to resolve.

Further reading

• The working version of this method, as our editors apply it: the Spokesperson Misinformation Assessment skill.
• Every scoring and inclusion system Citizen Analyst uses, side by side: Scoring & Inclusion Systems — Complete Inventory.
• The companion method for documented actions by officials: The Democratic Malice Methodology.

Expanded with AI research and drafting assistance from the primary sources cited; editorial review pending.