Inspect and review a candidate

Choose reviewers and eligible adjudicators under Review team on the study page. Reviewers sign in before they can be selected. Each reviewer inspects sources and records an independent assessment on the same page. The owner controls execution and allocation. Review and annotation writes are disabled during active processing.

Open a run, select Distribute reviews, and set Maximum pending reviews per person to enable automatic distribution for that run. The portal assigns independent pairs when the team changes, someone submits a review or a run makes new candidates available. It preserves existing assignments and skips runs while they are processing. Candidates wait until a pair has capacity and an eligible adjudicator is available. Leave the limit blank to use manual preview and confirmation.

Find your tasks

After signing in, the home page shows pending review and adjudication counts. Open My reviews to see Pending reviews, Adjudications, Submitted or All assignments. Each task identifies its study, run, candidate and category. Select Inspect and review to open the evidence and decision form. Submitted tasks stay available for corrections and indicate when the other review or adjudication is pending.

The queue uses candidate-level assignments. Older runs preserve the reviewers already assigned through the original study plan. Newly distributed tasks appear on the next page load. With SMTP configured in Django, automatic distribution sends a grouped notice when a reviewer receives work for an empty queue in a run. Other new assignments for that run are grouped into a later email. An empty queue can mean that the run is waiting for capacity or that the owner has yet to distribute candidates.

Inspection checklist

  1. Select a category in Results and open Inspect. Read the hypothesis, discovery reason and missing evidence.
  2. Verify endpoint identity, IRTF/IETF affiliation, chronology and the actual object/agenda. A name, reference or shared author is only a clue.
  3. Locate the passage in a preserved source. Inspect the relevant revision, context, provenance and contradictions.

Inspect and record your assessment

Read the located passages and context, link additional preserved evidence where needed, and record gaps. The independent decision form is on the same candidate page. Opening it does not change candidate data or mark inspection complete. Scientific decisions and labels belong to each reviewer's assessment; shared preparation does not replace that judgment.

Review checklist

  1. Apply the T1–T6 and object criteria. Select confirmed, rejected or uncertain, and justify the decision. Pending means no consolidated decision yet.
  2. Record evidence with code, source url, stored artifact, matching sha256, excerpt, location, event_date and date precision. Confirmed findings require the applicable mechanism and object conditions; O16 requires a description.

Evidence and participant annotations reference this run's sources; they do not fetch the linked URL or verify identity automatically. Manual hypotheses can be registered in all three categories without fabricating automatic signals. Record gaps when the necessary sources are missing. The interface does not provide a general upload workflow for arbitrary complementary sources.

Independent decisions and corrections

First and second reviewers see only their own decisions until both have submitted. Consolidation uses matching state, mechanism and object labels, or adjudication after both reviews. Disagreement includes differing mechanism or object label sets even when both decisions match. A third person outside the pair is assigned automatically for distributed cases. One positive review does not establish independent confirmation. Unresolved disagreement remains distinct from confirmed findings.

Submitting a correction appends history rather than erasing the prior review. Reports use active decisions. An annotation does not silently rewrite a submitted scientific judgment. Use an explicit correction when new evidence changes the decision.

Human mediation and T6

Human discovery has two routes: HUM_SHARED_AUTHOR starts from a documentary candidate; HUM_AUTHOR_INTERSECTION can start without one. For an independent lead, first determine whether there is an identifiable contribution connecting its documents. Matching addresses, similar topics, editing or a presentation alone do not establish that contribution or a mediation role.

For either route, check the actual person, participation on both sides and mediation role. A confirmed mediation must have:

  • T6 and a linked non-mediation base in the same run;
  • independent consolidated T1–T5 confirmation of that base;
  • located evidence of the person's mediation role, not authorship, chair status, citation or lineage alone.

These gates enforce review dependencies, not automatic truth of the passage. A correction that removes the base confirmation makes dependent mediation pending again without deleting its reviews. Filter Human mediation by base status to prioritize eligible cases; a confirmed base is not confirmed T6.

Associate an event with an exploratory lead

  1. Locate or register a documentary or organizational hypothesis in the same run that describes the contribution being investigated.
  2. On the human candidate, expand Link or change the related event, enter that candidate's ID and explain the contribution connecting the lead to the event. Leave the ID blank with a justification to unlink it. An event from another run or a human-mediation candidate is not an eligible base.
  3. The application records the previous and current link and your justification. IDs, evidence and submitted reviews are preserved. Changing the base does not carry confirmations about a different base into the new association; reviewers must evaluate the new link.
  4. Confirm the base through independent agreement or adjudication, then submit the role-bearing evidence needed for T6. Uncertainty or rejection can be recorded before a base is established.

Manual mediation registration still requests a base and a proposed mediator role. Base-free automatic leads are generated by the independent discovery action; their existence does not change the scientific conditions for T6.

Review possible agenda adoption

For ORG_AGENDA_ADOPTION, verify that the attribution and mandate concern the same research result and that the originating work was available before the mandate. Distinguish adoption of work from replacement or succession of organizations. Inspect the event source for each charter date; an administrative observation is not a publication or approval event. The signal supplies two passages and possible charter chronology, not a T5 decision.

The Results guide explains filters and exports. Full review history remains in the run export; respect the study's blinding procedure when distributing it.