Candidate Disposition Log¶
Traceability record — instantiates Progressive Narrowing
Records the fate of every candidate at every stage — advanced, held, merged, eliminated, or reopened, with the reason and the evidence — so a narrowing set never shrinks silently.
A Candidate Disposition Log is the running ledger of what happened to each item in the live set — not a screen, a score, or a judgment, but the durable record of every judgment already made. Its single job is traceability: for each candidate it captures the disposition (advanced, held, merged, eliminated, deferred, reopened), the stage and gate that produced it, the rationale, and the evidence cited, so the history of the set is reconstructable long after the meeting ends. It owns no criteria and casts no vote. What it does is make hidden narrowing impossible, because a candidate that leaves the live set without a logged reason is a gap the log exposes.
Example¶
A film festival receives about 4,200 feature submissions for roughly 30 program slots. Each film is routed to first-round screeners; some advance to a programming committee; a handful become finalists. The disposition log is the spine of the whole operation. Every film has a row recording who screened it, the round-one verdict, the reason ("out of scope — TV pilot, not a feature"; "held — strong, but slot conflict with a similar title"; "advanced to committee"), and a pointer to the screener's notes.
Three weeks before the festival, a board member asks why a well-known director's film isn't in the lineup. The log answers in seconds — eliminated in round two; three of four committee members rated it below threshold; notes attached — instead of triggering a frantic reconstruction from memory. When a confirmed finalist withdraws, the log's "held" rows are precisely the reserve pool to promote from. The log decided none of this. But without it, the festival could neither defend its choices to a skeptical board nor recover its near-misses when a slot reopened.
How it works¶
- One row per candidate, append-only. Each disposition is a new event, not an overwrite, so the full trajectory (entered → held → reopened → eliminated) stays legible.
- Every event carries stage, gate, rationale, evidence pointer, and owner. A disposition recorded without a reason is flagged incomplete rather than accepted.
- Reopening is a first-class event. When new evidence revives a candidate, the log links the reopening to the original elimination so the two can be read side by side.
- Reconciliation. The count of open rows must equal the option set minus everything eliminated or deferred; a mismatch means a candidate left the set unrecorded.
Tuning parameters¶
- Disposition-code granularity — a coarse advanced/eliminated split is cheap but blurs held, merged, and deferred; a rich code set captures nuance but burdens screeners and invites inconsistent coding.
- Rationale depth — a checkbox reason versus a written justification. Deeper rationale makes later review and reopening trustworthy but slows the pipeline.
- Access and visibility — who can read the log (auditors only, all reviewers, candidates themselves). Wider visibility deters arbitrary cuts but can chill candid notes.
- Retention window — how long dispositions are kept. Longer retention supports appeals and learning; shorter retention limits liability and stale-data noise.
When it helps, and when it misleads¶
Its strength is that it converts an ephemeral sequence of judgments into an auditable, recoverable history: silent narrowing becomes detectable, and the reserve and reopening machinery finally has something concrete to act on. It is the component most fairness reviews and appeals actually start from.
Its failure mode follows from what it is: a log records but does not judge, so a beautifully complete log can document a biased process exactly as faithfully as a fair one. Completeness is not correctness. The classic misuse is treating the existence of the log as evidence of a good process — "we documented everything" — when the rationales themselves are thin or written after the fact to survive an audit, a bureaucratic cousin of security theater.[n1] The guarding discipline is to audit a sample of rationales for substance rather than mere presence, and to check that reopening events actually occur: a log in which nothing is ever reopened may mean the reopening rule is dead, not that every cut was right.
How it implements the components¶
option_set— the log is the authoritative roster of the live pool; its open rows are the candidates still in play.decision_gate— each gate outcome (advance, hold, eliminate, merge, defer, reopen) is written as a disposition event, giving every gate decision an owner and a timestamp.reopening_rule— reopening is logged as an event linked back to the original elimination, so the record shows when and why a candidate returned.elimination_rationale_record— its core: every removal carries the reason and the evidence cited, which is exactly this component.
It stores the history but sets no survivor_criteria and computes no comparison_frame of its own — those are weighted_scoring_matrix — and applies no filter_sequence — that is successive_screening; the log is the ledger those screens write to, not a screen itself.
Related¶
- Instantiates: Progressive Narrowing — supplies the traceability the archetype's "eliminations must be traceable" invariant demands.
- Consumes: the gate decisions produced by Successive Screening, Hiring Shortlist Process, and the other screening siblings — the events it records.
- Sibling mechanisms: Successive Screening · Diagnostic Narrowing Protocol · Design Downselection Review · Hiring Shortlist Process · Procurement Shortlisting · Research Hypothesis Elimination · Legal Issue Narrowing · Weighted Scoring Matrix · Funnel Process
Editorial Notes¶
Form Classification¶
Form family: Record, Log & Register
Rationale: The mechanism appends every candidate's advance, hold, merge, elimination, and reopening event with stage, reason, evidence, and owner, so its operative form is a disposition history.
Nearest alternative: Decision, Gate & Allocation — It records decisions made elsewhere rather than selecting candidate outcomes itself.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Organizational & Management Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Stage-gate and portfolio governance supply the repeated dispositions—advance, hold, merge, eliminate, reopen—and the responsibility to record their reasons.
Related originating lineages:
- Accounting & Auditing — Ledger discipline contributes completeness, dated entries, evidence references, and reconstructability of every narrowing step.
- Law & Governance — Administrative reason-giving supplies the expectation that adverse dispositions remain reviewable.
Review resolution: Organizational process management is primary because the log records each candidate's review state and disposition. Accounting supplies auditable state records and law supplies retention, fairness, and defensibility requirements; their combined artifact is a multi-domain Encyclopedia synthesis.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
The log is usually the only component that outlives the people who made the decisions. When a fairness review, an appeal, or a "why did we drop this?" question lands a year later, it almost always starts here — not with the criteria, which were in someone's head, but with the record of what was actually done.
[n1] Security theater — Bruce Schneier's term for measures that deliver the feeling of a safeguard without its substance. A disposition log kept for appearance's sake, its rationales written to look thorough rather than to be true, is the same failure in bureaucratic form: the record exists, but it certifies nothing. ↩