Domain-Specific Trust Evidence and Exposure Matrix¶
Test or assessment — instantiates Evidence-Bounded Trust Governance
Replaces one global trust label with a grid of specific behaviors crossed against separated trust dimensions, each cell carrying an evidence grade and its contrary evidence.
The Domain-Specific Trust Evidence and Exposure Matrix is a static cross-tabulation: down one axis, the specific behaviors you are actually relying on someone for, in a named domain and at a stated stakes level; across the other, the trust dimensions those behaviors depend on — ability, reliability, honesty, care, and external constraint. Its one defining move is decomposition: it refuses the sentence "I trust them" and forces "I have strong evidence they are accurate, weak evidence they are honest about sources, and no evidence at all they are reliable under deadline." Each cell holds the grade of the best evidence for that behavior-by-dimension pair and, crucially, any evidence pointing the other way. The matrix does not gather new evidence, run a trial, or watch outcomes over time — it arranges what is already known so that a weakness in one dimension can no longer hide behind a strength in another, and so the cells that are both high-stakes and thinly evidenced light up as the ones to worry about.
Example¶
A regional newsroom is deciding how far to rely on a new stringer who has filed three strong pieces. "She's great — use her" is the global label the editor wants to avoid. Instead the news desk builds the matrix. The rows are the behaviors they might rely on her for: filing routine council coverage, breaking a fast-moving story, and handling a sensitive source in a domestic-abuse investigation. The columns are the dimensions: factual accuracy, honesty about sourcing, deadline reliability, and care with vulnerable subjects.
Filling the grid is sobering. Accuracy on routine coverage is well-evidenced (three clean pieces, no corrections). But honesty-about-sourcing on the sensitive story is an empty cell — nobody has ever watched how she protects an anonymous source under pressure — and that cell sits at the highest stakes on the page. Care-with-vulnerable-subjects has one faint positive signal and one contrary one (a subject once complained she felt rushed). The output is not a score. It is a picture: reliance on routine filing is warranted now; reliance on the sensitive investigation is not, because the two cells that matter most are blank or contested. That single view tells the desk exactly where to withhold reliance and exactly what evidence would change the answer.
How it works¶
- Name the behaviors, not the person. Each row is a concrete relied-on action in one domain at a stated stakes level, drawn from the purpose-and-stakes frame — never "trustworthiness" in the abstract.
- Separate the dimensions into columns. Ability, reliability, honesty, care, and constraint each get their own column so a single observation cannot be smeared across all of them.
- Grade each cell, and record the contrary. Every cell gets the grade of its best evidence (direct representative performance beats corroborated reference beats self-report) and a note of any contrary or absent evidence. A blank cell is a finding, not a gap to paper over.
- Read the hotspots. The cells that are simultaneously high-stakes and low-grade are the reliance you cannot yet justify. The matrix hands those to other mechanisms to resolve.
Tuning parameters¶
- Row granularity — how finely behaviors are split. Splitting "handles sources" into "routine" and "vulnerable" surfaces the dangerous cell; over-splitting produces a grid too sparse to read.
- Dimension set — which columns you carry. Adding authority and incentive alignment catches conflicts a five-column grid misses, at the cost of more blank cells to fill.
- Grade scale resolution — a coarse strong/weak/absent scale is fast and legible; a finer provenance-tagged scale captures recency and independence but slows the read.
- Contrary-evidence burden — whether a cell may be marked strong on one confirming instance, or must survive a search for disconfirming cases first. Raising it fights confirmation but costs effort.
- Stakes shading — whether cells are weighted by the frame's stakes so the eye goes to high-stakes/thin-evidence cells first.
When it helps, and when it misleads¶
Its strength is that it is the antidote to the halo effect — the tendency for one salient positive impression to color unrelated judgments — because a grid physically prevents a strong cell from filling in a blank one.[n1] It makes the shape of your evidence legible in one view, and it turns "do I trust them?" into the far more useful "which specific cells are blank, and do they matter?"
Its failure mode is that a matrix is only a snapshot of evidence you already hold, so it inherits every blind spot in that evidence and can lend a false air of completeness — a fully-inked grid can still rest entirely on the trustee's own self-reports. The classic misuse is treating a filled cell as a verified one when the "evidence" was never independently checked, or quietly summing the grid back into a single average and re-creating the global label the matrix existed to destroy. The guarding discipline is to keep the grades honest about provenance (an unchecked claim is a weak cell, not a strong one), to leave blanks visible rather than averaging them away, and to hand thin high-stakes cells to a verification or trial mechanism rather than filling them by inference.
How it implements the components¶
trust_purpose_relationship_domain_and_stakes_frame— the rows are the frame made concrete: each is a specific relied-on behavior in a named domain at a stated stakes level, which is what stops a global label from forming.trustee_ability_reliability_honesty_care_and_constraint_profile— the columns separate the trust dimensions and the cells preserve each one's evidence grade and contrary evidence, keeping the dimensions distinct as the invariant requires.
It does not independently produce or authenticate the evidence in its cells — that rigorous source-triangulation is the Independent Reference, Claim, and Conflict Verification sibling's proportional_evidence_signal_verification_and_privacy_plan; the matrix only arranges evidence already in hand. It also does not profile recoverability and alternatives or stage a live tier (trustor_exposure_recoverability_power_and_alternative_profile, domain_specific_trust_tier_and_reliance_envelope — the Graduated Reliance trial), nor track drift over time (trust_calibration_outcome_dependency_and_drift_review — the audit).
Related¶
- Instantiates: Evidence-Bounded Trust Governance — the matrix supplies the decomposed evidence picture the rest of the lifecycle reasons over.
- Sibling mechanisms: Graduated Reliance and Bounded-Exposure Trial · Independent Reference, Claim, and Conflict Verification · Commitment, Disclosure, Exception, and Outcome Log · Trust-Breach Triage, Containment, and Repair Protocol · Trust-Calibration and Dependency Audit
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Domain-Specific Trust Evidence and Exposure Matrix operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it replaces one global trust label with a grid of specific behaviors crossed against separated trust dimensions, each cell carrying an evidence grade and its contrary evidence.
Independent corroboration: The frozen evidence defines Domain-Specific Trust Evidence and Exposure Matrix as 'Replaces one global trust label with a grid of specific behaviors crossed against separated trust dimensions, each cell carrying an evidence grade and its contrary evidence', so its operative form is Assessment, Review & Assurance.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Security Studies & Intelligence Analysis
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Security and intelligence practice cohered behavior-specific trust judgments tied to evidence quality, contrary evidence, and bounded exposure rather than a global reputation label.
Related originating lineages:
- Law & Governance — Risk-based authorization and due-process traditions supplied separated permissions and reviewable evidence.
- Psychology — Halo-effect research explains why one overall impression must not contaminate unrelated trust dimensions.
Review resolution: Security and intelligence source evaluation is primary, while authorization law and psychological debiasing genuinely shape the evidence-and-exposure matrix.
Attribution caveat: The matrix combines intelligence source evaluation, access governance, and a psychological debiasing structure.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
The nearest twin is verification, and the difference is directional: the matrix takes evidence you already have and lays it out to expose gaps; verification goes and gets or authenticates the evidence a gap demands. Run the matrix first — its blank high-stakes cells are the work order that tells verification and the trial exactly where to point.
[n1] The halo effect is the well-documented cognitive bias in which an overall impression of a person (often driven by one trait like competence or likeability) contaminates judgments of their specific, unrelated attributes. A behavior-by-dimension grid is a structural countermeasure: it makes each judgment occupy its own cell so the halo cannot spread. ↩