Structured Review Form¶
Documentation template — instantiates Bias-Specific Decision Audit
Turns a bias review into a filled record — the decision's context, the vulnerability map, the checks run and what they found, what changed, and the residual risk accepted — so a later reader can see what actually happened.
A Structured Review Form is the template that turns a bias review into a durable record. It has a field for each thing the review is supposed to produce — the decision's context and stakes, the vulnerability map, the specific checks run, what each check found, what changed as a result, and the residual risk knowingly accepted — and the review is not "done" until the fields are filled. Its defining idea is capture, not judgment: the form neither decides which biases are plausible nor performs a single check; it makes the review legible after the fact, so a later reader — an auditor, a successor, an inquiry — can see what was examined and what was skipped. Where a checklist prompts a reviewer forward into action, the form records backward what the review actually did.
Example¶
A national benefits agency is about to roll out an automated eligibility-scoring tool, and policy requires a bias review before launch. The review itself happens in a workshop; the Structured Review Form is what survives it. Its context block records the stakes (tens of thousands of claimants), reversibility (hard — appeals are slow), evidence channels (historical claims data), and affected parties (applicants who never sit in the room). The map block lists the three pathways the team judged live: selection bias in the training data, framing effects in how caseworkers read the score, and an affected-party blind spot.
Under each pathway, the check run and finding fields are completed — a data-representativeness review flagged two under-sampled regions; the affected-party review found no appeal explanation had been drafted. The change field records the two fixes made; the residual risk field records the one accepted: the tool ships without a fairness guarantee for a rare claimant category, flagged for the first post-launch review. Months later, when a journalist asks how the tool was vetted, the completed form answers in one page instead of a scramble of recollections.
How it works¶
The form is a fixed set of fields whose structure enforces completeness. Three properties make it a trace rather than a memo:
- It is field-complete — an empty map field or an empty change field is a visible gap, not a silent omission.
- It is paired — every listed check has a finding, and every finding has either a change or an accepted-risk note, so the form cannot record scrutiny that led to nothing.
- It is archived against the decision — the record outlives the meeting and the people in it.
What the form deliberately does not do is generate content: it inherits the map and the findings from the review and merely gives them a shape a later reader can walk.
Tuning parameters¶
- Field granularity — a few free-text boxes versus a fine-grained schema. Finer fields make records queryable and comparable across decisions but push toward form-filling for its own sake.
- Mandatory vs. optional fields — which blanks block sign-off. Hard-required fields guarantee the map and residual risk are never skipped; too many turn the form into an obstacle people route around.
- Retention and access — how long records are kept and who can read them. Durable, discoverable records enable learning and accountability but create sensitive documents that can be subpoenaed or weaponized.
- Linkage — whether fields point at the actual artifacts (the data review, the revised criteria) or merely describe them. Linked records are auditable; described ones drift into unverifiable summary.
When it helps, and when it misleads¶
Its strength is traceability: it converts a conversation that would otherwise evaporate into a record that shows which vulnerabilities were checked, what changed, and what risk was consciously carried — the raw material for both accountability and cross-decision learning.
Its failure mode is that a record of a review is easily mistaken for a good review. A fully completed form can document a shallow, generic audit and lend it false confidence — the box is ticked, so the decision feels vetted. Worse is goal displacement[n1], where filling the form becomes the objective and the thinking it was meant to capture atrophies into boilerplate. And because the record is durable, it can be written defensively — to read well to a future auditor rather than truthfully. The discipline that guards against this is to treat the form as evidence of a review, never a substitute for one: require the map and residual-risk fields to be specific to this decision, and read a form that could have been copied from any other decision as a red flag.
How it implements the components¶
evidence_and_process_trace— its core: the completed fields are the record of what was checked, what each check found, what changed, and what risk remains.decision_context_boundary— the context block fixes the stakes, reversibility, evidence channels, participants, and affected parties the rest of the record is read against.
It does not produce the vulnerability analysis or the fix it records (bias_vulnerability_map, process_revision — those are Bias Audit), and it does not itself run any check (targeted_bias_check — the check siblings such as Blind or Masked Review).
Related¶
- Instantiates: Bias-Specific Decision Audit — the Structured Review Form supplies the traceability strand, making the audit legible after the fact.
- Consumes: Bias Audit — it records that audit's map, findings, and revisions; with no audit to capture, the form is empty scaffolding.
- Sibling mechanisms: Bias Audit · Decision Checklist · Hiring Review Rubric · Blind or Masked Review · Independent Estimation · Diagnostic Debiasing Check · Decision Log · Reference-Class Forecasting
Editorial Notes¶
Form Classification¶
Form family: Record, Log & Register
Rationale: Structured Review Form operates as a persistent ledger, log, register, or case record that preserves history and traceability because it turns a bias review into a filled record — the decision's context, the vulnerability map, the checks run and what they found, what changed, and the residual risk accepted — so a later reader can see what actually happened.
Independent corroboration: The frozen evidence defines Structured Review Form as 'Turns a bias review into a filled record — the decision's context, the vulnerability map, the checks run and what they found, what changed, and the residual risk accepted — so a later reader can see what actually happened', so its operative form is Record, Log & Register.
Nearest alternative: Assessment, Review & Assurance — Structured Review Form includes features of a bounded evaluation of existing evidence or work that produces a finding or disposition, but its defining operation is a persistent ledger, log, register, or case record that preserves history and traceability.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Accounting & Auditing
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: A controlled form that requires the same evidence, findings, disposition, reviewer, and date at every review is audit documentation made repeatable. NARA metadata policy grounds required, consistently defined record elements and provenance; auditing supplies review assertions and sign-off.
Related originating lineages:
- Human-Computer Interaction — human_computer_interaction contributes human-computer interaction and interface design to this mechanism's defining operation—Turns a bias review into a filled record — the decision's context, the vulnerability map, the checks run and what they found, what changed, and the residual risk accepted — so a later reader can see what actually happened—without displacing the selected primary historical lineage.
- Law & Governance — Reasons support accountability.
- Library & Information Science — library_information_science contributes library and information-science stewardship to this mechanism's defining operation—Turns a bias review into a filled record — the decision's context, the vulnerability map, the checks run and what they found, what changed, and the residual risk accepted — so a later reader can see what actually happened—without displacing the selected primary historical lineage.
- Organizational & Management Science — Forms standardize review.
- Systems Thinking & Cybernetics — Systems thinking, feedback control, and cybernetics supplies a parallel or contributing lineage for the mechanism's defining operation: turns a bias review into a filled record — the decision's context, the vulnerability map, the checks run and what they found, what changed, and the residual risk accepted — so a….
Review resolution: The blind reviewers disagree on primary lineage (accounting_auditing versus organizational_management). Authoritative or primary research supports accounting_auditing as the best historical origin: A controlled form that requires the same evidence, findings, disposition, reviewer, and date at every review is audit documentation made repeatable. NARA metadata policy grounds required, consistently defined record elements and provenance; auditing supplies review assertions and sign-off. The cited U.S. National Archives, Metadata Requirements for Permanent Electronic Records; U.S. National Archives, Records Management Profile directly supports the mechanism's defining operation. All independently supported contributing domains are retained without an arbitrary cap. origin_mode=cross_disciplinary_synthesis records lineage, while domain_reach=universal records later applicability separately from provenance.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- U.S. National Archives, Metadata Requirements for Permanent Electronic Records
- U.S. National Archives, Records Management Profile
Notes¶
A Structured Review Form and a Decision Log both leave a durable record, but of different things: the log captures why a choice was made and what it governs across many decisions; the form captures how one decision's bias review was conducted. A log entry can cite a completed form; they are not substitutes.
[n1] Goal displacement — Robert K. Merton's term for the bureaucratic inversion in which following the procedure becomes the goal and the procedure's original purpose is lost. A review form is especially prone to it, because a filled form is so easy to mistake for a completed review. ↩