How the record works, and its limits.

Voluntary participation

Anyone in the United States can contribute an experience. People find State of Work through news coverage, social media, community organizations and paid outreach. Because people opt in, the record will over-represent some groups — for example, people who are online, people motivated to speak, and communities where outreach happened to reach — and under-represent others. We report on who is participating so readers can account for this.

Verifying people, not claims

Contributors are asked to confirm they are a real person through a sign-in or phone verification. This helps limit duplicate and automated submissions. It establishes that a contribution likely comes from a real, unique individual.

It does not verify what the person describes. State of Work does not investigate individual accounts, contact employers, or determine whether events happened as described. We label contributions accordingly:

Verified participant. Self-reported experience.

Only an opaque identifier from the verification provider is retained, so that the same person can return to update their account. Verification details are never published.

Structured questions and narratives

Each contribution has two parts:

• Structured answers — which experiences apply (multiple choices allowed), location, industry, occupation, employment status, approximate timing, job-search duration and income change, in broad categories.
• A narrative — written or recorded in the participant’s own words.

Experiences are classified into three kinds so they can be analyzed correctly: events (being laid off, having hours cut), current status (unemployed, underemployed) and ongoing consequences (reduced income, financial strain). One person can report several of each, and may return later to add to their record as circumstances change.

How an account becomes a mark

Each account is drawn as a square in four parts. A quarter is filled when the account reports something in that group, and left empty when it doesn’t. The mark shows the shape of an experience, never its details.

Marks are generated from the structured answers only. They are published only as part of aggregate views, and sharing a personal mark is always optional.

Geographic classification

Participants provide a city and state for where they live, and separately for where the employment event happened if that is different. These participant-provided locations are the only source of geography; we do not infer location from IP addresses. Places are matched to standard Census identifiers (places, metropolitan areas, states and regions) so the record can be compared with other data. We never collect street addresses or precise coordinates.

Aggregation and privacy thresholds

Public figures are counts and shares of contributions. Any cell — a state, a metro area, an industry within a state, a category — with fewer than 10 contributions is withheld and shown as such. Filters that would narrow results to very few people are disabled. Story excerpts display only broad attributes (state, industry, season and year), never a combination precise enough to identify someone.

Maps that show totals are adjusted for population, so they don’t simply show where most people live. Even so, a darker area means more people there chose to contribute — not necessarily that conditions are worse.

Editorial review and publication

Contributions are private by default. Contributing gives permission to include information in de-identified aggregate statistics; it does not give permission to publish anything a person wrote or recorded.

Publication requires separate, explicit consent — one for written excerpts and another for audio. Editors review anything considered for publication, remove identifying details, and may decline to publish. Because a voice can identify someone even without a name, audio is never described as anonymous. Participants can withdraw publication consent at any time before publication; previously published material will be removed on request, though copies made by others may persist.

Limitations of self-reported data

• Accounts reflect individual perspectives and memory, and may be incomplete.
• Participation is uneven across places, industries and demographic groups.
• Changes in participation over time can reflect outreach, news events and attention, not just labor-market conditions.
• Categories are broad by design, which protects privacy but limits precision.

We encourage researchers and journalists to treat State of Work as qualitative and descriptive evidence that complements — and never replaces — official statistics and rigorous surveys.

How this compares with other measures

Descriptions are summaries for orientation. Refer to each source for its precise definitions. Before launch, this table should be reviewed against current published methodology.

Questions about method? See Contact.