About this edition · historical cutoff August 31, 2026
About and research
The Oregon Governance Atlas helps readers understand how Oregon’s legislature works, how it developed, and what its decisions changed. It is an academic and informational project, with historical accounts, institutional studies, and sources readers can inspect.
Editorial independence
The Atlas has no legislative agenda and does not endorse candidates, parties, legislation, ballot measures, or constitutional changes. Its studies investigate how institutions operate, including documented successes, harms, exclusions, and unresolved problems.
Comparisons describe specific rules, costs, consequences, and uncertainties. Hypothetical scenarios state their assumptions. We do not rank politicians, parties, policies, or alternatives, or give government a composite score.
A documented problem can have several explanations: the governing rule, implementation, resources, outside events, or deliberate choices. A study examines these explanations and relevant counterexamples before drawing a conclusion. Different views receive attention according to their evidence.
Explore Institutional Studies for the questions, existing evidence, and investigations connecting decisions to their consequences.
Method
A decision history begins with a question and a defined period. The research follows the people, competing texts, procedural choices, funding, and implementation, then examines later revision or litigation. Case selection should include relevant counterexamples and explain why each episode was chosen.
A quotation needs a source, date, and setting. Evidence of agreement is needed to describe a coalition. A regional benefit needs a defined geography, population or other denominator, and period. When the reason a proposal stopped is unknown, the account should leave it unknown.
The research protocol separates what a source reports, what has been checked, and what the analysis infers. Estimates, assumptions, and simulated results have their own labels. It requires reading logs with the versions and pages, tables, or timestamps examined; a download or text search does not count as reading a whole document.
Authorship and review
The research and website have been developed with AI assistance. The current workflow uses the conversational model for interpretation and local tools for retrieval, extraction, search, calculations, and mapping. There is no separate batch-LLM or hosted-embedding pipeline in this workflow.
A model-assisted check can identify errors or contrary evidence. It does not constitute independent human review or academic peer review. A study may claim human review only when it records the reviewer, date, scope, and outcome; an unspecified review status should not be read as approval.
A study’s disclosures should identify its contributors, assistance, review coverage, funding, and relevant conflicts. Where a declaration is absent, it should be treated as undisclosed. Academic affiliation or institutional endorsement requires an explicit, documented statement.
Publication status
An available reading, sufficient source coverage, and a recorded review are separate milestones. The session counts below come from release atlas-2026-09-04, generated September 4, 2026. They do not establish that a new institutional investigation is complete.
20
Account available; evidentiary review not complete
25
Reviewed account
115
Sources being assembled
4
Account available; publication status under reconciliation
Session-by-session status and checks
What is missing, and where
The research queue holds 2,466 distinct evidence needs: 280 named by session accounts and cases, 2099 record classes not yet acquired for particular sessions, 84 tasks that need a person to write to an archive or visit one, and 3 datasets still to be built. A reader-facing page shows its linked needs once; the detail lives here.
The full retrieval queue · 468 archive and access tasks drafted, none sent without authorization.
How sources are gathered
The local collector retrieves approved public documents, checks for existing copies, and preserves originals with checksums. It observes access rules, download limits, and a disk-space reserve. The power and regions collection records 122 preserved items in this edition. Acquisition, reading, verification, and publication are tracked separately.
Repositories in the research registry
Top 25 of 356 repositories by record count. Records describe custody, not reading.
Changes
- September 4, 2026: Introduced one publication-status definition; separated draft, identity, and source checks; flagged status conflicts instead of resolving them by label.
Corrections and disagreement
A correction should identify the page, the claim, and the evidence that changes it. Substantive revisions should preserve the previous edition and explain what changed. New source material triggers review of the findings that depend on it; it does not automatically replace a published account. A dedicated public corrections channel has not yet been published.
Editions and dating
Statements about officeholders and operative rules describe the edition dated August 31, 2026. The date a document was retrieved is not the date a rule took effect; each preserved source records both where the server supplied them. A later live edition will carry its own cutoff and effective-date checks.