Editorial methodology

Documentation is only as useful as its provenance. This page describes how every piece of content reaches the platform and what we promise readers about its accuracy.

Source tiers

The platform recognizes four tiers of sources, in descending order of authority:

1. Primary — government and court. Federal and state statutes, Supreme Court and circuit-court rulings, federal commission reports, agency underwriting manuals, federal statistical series (Census, BLS, BJS, CDC, OCR), state attorney-general filings.
2. Primary — archival and contemporaneous. The Library of Congress Manuscript Division (including the Federal Writers' Project Slave Narratives), the National Archives Record Groups, university digital-scholarship lab repositories (the Mapping Inequality project at the University of Richmond, the Stanford Open Policing Project), contemporaneous newspaper coverage from rights-cleared archives.
3. Peer-reviewed. Journal articles in social-science and public-health venues with documented replication or established methodology. Working papers from the National Bureau of Economic Research where they are the originating publication for a finding that has since been peer-reviewed.
4. Institutional reporting. Material from civil-rights organizations (the NAACP Legal Defense Fund, the Equal Justice Initiative, the Brennan Center for Justice, the ACLU, the Sentencing Project) whose institutional review processes we consider sufficient for citation. We tend to cite the underlying primary sources these organizations themselves cite when those are accessible.

Material from social-media posts, advocacy blogs, anonymous opinion pieces, or wiki sources is not cited. We use Wikipedia as a finding aid but not as a citation.

Testimonies

Two kinds:

• Public-domain archival — from the Federal Writers' Project (1936–1938) and similar sources whose copyright has lapsed. Re-published verbatim where possible; light edits noted in the source block.
• Contributor-submitted — require informed consent and an editorial review pass that confirms the account is non-libelous, properly anonymized at the contributor's request, and consistent with documented patterns. We never publish a name without explicit written permission.
• Composite scenarios — clearly flagged with an editorial note at the top of the entry. These reconstruct documented patterns of discrimination drawn from paired-tester studies, federal civil-rights enforcement data, and peer-reviewed empirical research. Composite scenarios are never representations of specific real cases. Each one names the empirical source the underlying pattern is drawn from.

Why composite scenarios? Because the documentary record of contemporary discrimination is dominated by enforcement statistics, paired-tester studies, and complaint counts — not by named individual cases. The lived experience that those statistics describe is rarely accessible as primary-source narrative because contemporary contributors face retaliation risk. Composite scenarios let the platform render those patterns as readable prose while keeping a clear editorial firewall between empirical evidence and fictional reconstruction.

Pathways

Each pathway is verified against the organization's own published guidance. The retrieval date in the source block tells you when we last checked — if it's old, click through to the org page and confirm before acting.

Phone numbers and intake emails are copied from the organization's public-facing site. We do not maintain back-channel intake details. Filing deadlines (the 180-day window for OCR complaints, the 300-day window for EEOC charges in deferral-state jurisdictions, the 365-day window for HUD housing-discrimination complaints) are quoted from the enforcing agency's own published guidance and double-checked against the underlying statute or regulation.

We do not publish pathways for jurisdictions where the enforcement mechanism is so distorted in practice that the listing would mislead readers about what filing actually accomplishes. Where an enforcement agency exists but has documented backlog or substantive-policy issues that affect typical outcomes, we note the relevant published reporting alongside the contact information.

History entries

Every entry cites a primary or near-primary source: federal commission report, court ruling, federal statistical series, or contemporaneous news coverage from a reputable outlet. Where the source is publicly archived (e.g., the Oklahoma Commission Report, Library of Congress records), a link is provided.

We organize entries by year and tag each one with a coarse region. Entries that span multiple decades use the year the most consequential decision or statute was issued and note the broader range in the title (e.g., "1890–1968: sundown towns"). The dating convention prioritizes readability of the timeline over taxonomic precision.

Where two well-credentialed sources disagree on a fact — a death toll, a date, an attribution — we cite the range the sources span rather than picking the most dramatic number. The Tulsa Race Massacre death toll is the canonical example: the 2001 Oklahoma Commission report described a range, and we quote that range rather than collapsing it.

Imagery

Testimony pages on racist.systems carry no decorative imagery. First-person accounts of harm should not share visual real estate with stock photographs of unrelated people. This is the editorial posture taken by the Equal Justice Initiative, the Mapping Inequality project, and the Library of Congress slave-narratives collection.

History-entry pages carry public-domain photographs of the civic buildings, historical landmarks, signing ceremonies, and monuments referenced in each case — sourced from Wikimedia Commons under public-domain, CC-BY, or CC-BY-SA licenses with full author and license attribution preserved in the source.

We do not modify the rights metadata on any image we use. We do not crop images in ways that alter their evidentiary value. Where a CC-BY or CC-BY-SA license requires attribution and a license URL, both are preserved in the seed comment immediately above each entry.

Legal aid directory

Indexed by ZIP-3 prefix (the first three digits of US ZIP codes), which is coarse enough that small location errors still surface useful providers. Each entry's source block names where we found it — typically the provider's own site or a national legal-aid directory.

Providers are drawn from Legal Services Corporation grantee directories and the state-bar pro-bono referral pages. We include both LSC-funded providers (which have a federal eligibility cap of 125 percent of the federal poverty line) and non-LSC providers (which apply their own eligibility rules). The coverage note on each provider summarizes the area served and the services offered so readers can self-screen before calling.

Corrections workflow

Found something wrong? Email reach-us@racist.systems with the URL and the correction. We treat correction requests as bugs: logged, traced, fixed, with the change history kept.

Registered readers can also file a correction directly against any published entry via the in-platform reporting flow. Filed corrections go to a moderation queue, run through an automated abuse-screen pre-pass, and are then resolved by the editorial team. Substantive revisions are noted on the entry itself; minor typo fixes are not.

What we don't do

• Provide legal advice. We're a directory.
• Publish unverified accusations against named individuals.
• Originate political action. Campaigns and surveys link to the coalition or lab that owns them.
• Use stock photography. Testimony pages have no decorative imagery; history pages use Wikimedia public-domain or CC-licensed civic photographs only (see Imagery above).
• Republish copyrighted news content. Where contemporaneous news coverage is the best available source, we summarize and cite, never quote verbatim beyond the boundaries of fair use.
• Run user-generated comments. Reader feedback comes in through the corrections workflow above, which routes the message to a person rather than to a public-facing thread.

What we revise, and why

The platform's editorial position is that documentary archives improve with revision. Entries are updated when better primary sources surface, when a cited statute is amended, when a federal commission updates a report we relied on, or when a reader correction identifies an error. We do not retroactively edit primary-source quotations; we do retroactively correct paraphrases, dates, name spellings, and citation strings. Where the revision is substantive enough to change a reader's understanding of the entry, a brief revision note is added at the bottom of the entry.