Question–Evidence Matrix¶
Template — instantiates Relevance-Substitution Detection and Correction
Maps each claim or decision to the evidence types allowed to bear on it, and flags supplied cues that fall outside the grid.
Question–Evidence Matrix is a grid built before the evidence arrives: down one axis it lists the questions to be decided, and across the other the kinds of information that would legitimately bear on each, with the cells marking which evidence type is admissible for which question. It is a pre-registration of relevance. When a cue then shows up in deliberation, the matrix answers a single question — is this the kind of evidence we already said counts for this claim? — and anything that lands outside the grid is flagged as off-issue by construction. The idea that makes this mechanism distinct is that it works across a whole decision's worth of claims and fixes the standard of admissibility per question in advance, rather than dissecting a single comparison or discounting a single feeling in the moment. It is a map of what should count, drawn before the persuasion starts.
Example¶
An internal-medicine team builds a Question–Evidence Matrix for a diagnostic workup on a patient with fatigue and weight loss. The rows are the live differential hypotheses — hyperthyroidism, occult malignancy, depression, anemia. The columns are the evidence types that can legitimately move each — for hyperthyroidism: TSH and free-T4 levels, specific physical signs, symptom pattern; explicitly not the patient's articulate anxiety or the fact that a charismatic consultant "has a feeling." Each cell is filled in advance: labs and exam findings are admissible for the endocrine row; the consultant's confidence and the patient's demeanor are not. When the case is discussed and someone leans hard on how convinced the senior consultant sounds, the matrix flags it — persuasive, but it sits in no admissible cell for any row. The workup returns to ordering the tests the grid says would actually discriminate the hypotheses, and the "confidence" cue is set aside as not the kind of thing that decides a diagnosis.
How it works¶
- List the questions. Every claim, hypothesis, or decision in play becomes a row, so relevance is defined per question rather than for the case as a whole.
- Declare admissible evidence. For each row, name the evidence types that would genuinely bear on it and why; this is the standard, written before exposure to the actual cues.
- Fill the grid. Mark which evidence type is admissible, inadmissible, or conditional for which question, producing an explicit map of what should count where.
- Screen incoming cues against cells. As information arrives, place it in the grid; anything that fits no admissible cell for the claim it is aimed at is flagged as off-issue and kept out of the basis.
Tuning parameters¶
- Row granularity — how finely the questions are split. Fine rows give precise per-claim standards but a large grid; coarse rows are manageable but let a cue relevant to one sub-claim justify itself for a whole cluster.
- Admissibility strictness — how demanding the "counts as evidence" bar is per cell. Strict cells catch more substitution but risk excluding legitimate weak evidence; lax cells admit more and screen less.
- Conditional cells — whether cells are binary (in/out) or allow "relevant only if." Conditionals capture real epistemics but complicate the screen.
- Lock timing — how firmly the grid is fixed before evidence is seen. Locking early prevents motivated redefinition of relevance; allowing revision handles genuine surprises but reopens the door to gerrymandering.
When it helps, and when it misleads¶
Its strength is that it defines relevance before anyone is looking at a persuasive cue, which is the only time relevance can be defined disinterestedly. It is the working equivalent of a legal relevance rule: a standard that admits information only if it has a tendency to make a fact of consequence more or less probable, decided independently of how compelling any particular item happens to feel.[n1] By committing the standard in advance, the matrix denies a seductive cue the chance to argue its own way in.
Its failure mode is rigidity against genuine novelty: a pre-built grid can exclude a legitimately new kind of evidence simply because no one anticipated its cell, and a matrix that is never allowed to update becomes a bureaucratic cage. The classic misuse is gerrymandering the grid — quietly adding the admissible cell that lets a favored cue in, or defining rows so that only the desired evidence qualifies. The guarding discipline is to lock the standard before exposure, log any later change to the grid with its reason, and treat a cue that seems relevant but fits no cell as a prompt to reexamine the standard openly rather than to bend it silently.
How it implements the components¶
question_at_issue_record— each row is an explicit record of a claim or decision being evaluated.relevance_standard— the admissible-evidence declaration per row is the standard for what counts, fixed before exposure.issue_relevance_map— the filled grid is a map of which evidence types bear on which questions, against which incoming cues are screened.
It does not test whether a persuasive similarity structurally transfers (relevance_test) or register the source-domain features of a comparison (substitute_signal_register) — that analogy-internal dissection belongs to Analogy Mapping Table, its template twin.
Related¶
- Instantiates: Relevance-Substitution Detection and Correction — provides the pre-registered, per-question admissibility grid that screens cues by construction.
- Sibling mechanisms: Affect–Evidence Split Prompt · Red Herring Filter Checklist · Analogy Mapping Table · Decision-Basis Disclosure · Reweighted Update Log · Cue-Validity Audit · Proxy-Relevance Audit · Blinded or Masked Review
Editorial Notes¶
Form Classification¶
Form family: Representation, Specification & Plan
Rationale: Question–Evidence Matrix operates as a static representation, map, specification, schema, or prospective plan that externalizes information because it maps each claim or decision to the evidence types allowed to bear on it, and flags supplied cues that fall outside the grid.
Independent corroboration: The frozen evidence defines Question–Evidence Matrix as 'Maps each claim or decision to the evidence types allowed to bear on it, and flags supplied cues that fall outside the grid', so its operative form is Representation, Specification & Plan.
Nearest alternative: Decision, Gate & Allocation — Question–Evidence Matrix includes features of a case-specific gate, selection, routing, prioritization, or resource disposition, but its defining operation is a static representation, map, specification, schema, or prospective plan that externalizes information.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Law & Governance
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Predeclaring which evidence may bear on which question most directly operationalizes legal relevance and admissibility: Federal Rule of Evidence 401 defines relevance relationally between evidence and a consequential fact.
Related originating lineages:
- Medicine & Healthcare — Differential diagnosis and evidence-based medicine materially shape question-specific evidence grids.
- Organizational & Management Science — Decision matrices supplied the cross-tabulated governance artifact.
- Philosophy — Epistemology supplied the distinction between reasons that bear on a claim and merely salient cues.
- Statistics & Experimental Design — The statistics_experimental_design tradition materially shaped Question Evidence Matrix through randomization, prespecification, validity, and controlled inference.
Review resolution: The blind reviewers disagreed on primary lineage. Light authoritative research resolves the defining form in favor of law_governance: Predeclaring which evidence may bear on which question most directly operationalizes legal relevance and admissibility: Federal Rule of Evidence 401 defines relevance relationally between evidence and a consequential fact. The other materially formative traditions are retained as alternates; current breadth of use remains separate as domain_reach=universal.
Attribution caveat: The matrix form is an encyclopedia synthesis, though the page explicitly invokes the legal relevance rule. The named matrix is an encyclopedia synthesis rather than a standard method.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; medium confidence.
Sources consulted:
Notes¶
[n1] The evidentiary notion of relevance — information counts only if it makes a fact of consequence more or less probable (as in Federal Rules of Evidence 401) — is exactly what each admissible cell encodes, decided before the persuasive item is in view. ↩