Convergence Confidence Card¶
Scorecard — instantiates Independent Convergence Recognition and Transfer Design
A standardized one-page record that fixes the convergence claim, its supporting and disconfirming evidence, caveats, and a graded transfer confidence in a form others can audit and reuse.
Analysis that never gets written down decays into folklore — "everyone knows those cases converged" — and folklore cannot be checked or revised. The Convergence Confidence Card is the artifact that prevents that decay. It is a fixed-format, one-page record that states the convergence claim in one sentence, lists the independence evidence and the pressure match, records the negative cases and open caveats, and stamps the whole thing with a single graded confidence level rather than a yes/no verdict. Its defining property is that it is a record, not a computation: it does not weigh the evidence or derive the grade itself — it captures, in a stable and comparable structure, what other mechanisms concluded, so the claim can be audited later, compared across cases, and refined when the world talks back.
Example¶
A nutrition research group is deciding whether an effect seen in three separate randomized trials — that a particular time-restricted eating window improves a metabolic marker — is a genuine convergent finding worth folding into guidance. Rather than let the conclusion live in a slide deck, they fill a card. The claim line: "Independent trials in three unrelated populations converge on the same directional effect." Independence evidence: different funders, different sites, no shared protocol authors. Pressure match: comparable participant profiles and adherence pressures. Negative cases: one trial in a fourth population showed no effect (recorded, not buried). Caveats: short follow-up; surrogate marker, not outcomes. Grade: moderate — enough to justify a larger confirmatory study, illustratively, but not to change clinical guidance. Six months later a new null result arrives; because the card exists, the grade is downgraded to low in one edit, with the reason attached, instead of the original enthusiasm quietly persisting.
How it works¶
- One-sentence claim. The recurrence is stated as a single falsifiable proposition, so the card is about this claim and not a vague sense of similarity.
- Structured evidence slots. Fixed fields for independence, pressure similarity, sample diversity, negative cases, and performance — the same slots every time, so cards are comparable across studies.
- Grade, don't grade-then-hide. A short ordinal confidence (e.g. low / moderate / high) with the reasoning inline, borrowing the logic of graded evidence appraisal used in medicine.[n1]
- Versioned, not frozen. Each card carries a date and a revision line so a later outcome updates the grade in place rather than spawning a competing story.
Tuning parameters¶
- Grade resolution — a three-band scale versus a finer one. Coarse bands are honest about how little precision the evidence supports; fine bands invite false confidence but discriminate close cases.
- Caveat strictness — how aggressively unresolved doubts must be logged before a grade may rise. Strict caveating slows promotion but catches over-claims.
- Evidence-slot rigidity — a locked template versus free-form notes. Rigidity buys cross-card comparability; flexibility captures the odd case the template didn't anticipate.
- Refresh trigger — whether a card is revisited on a schedule or only when new evidence lands. Event-triggered refresh is cheaper; scheduled refresh catches silent staleness.
When it helps, and when it misleads¶
Its strength is institutional memory with an audit trail: a card makes a convergence claim comparable to its neighbors, downgradable when evidence turns, and legible to someone who wasn't in the room. It is the step that keeps a moderate finding from hardening into settled fact simply because it was asserted confidently.
Its failure mode is the grade eating the evidence — readers skip to the confidence band and treat it as an oracle, so a high stamped on thin reasoning travels with more authority than it earned. A classic misuse is grade inflation to justify a decision already made. The guarding discipline is to require that the grade never be quoted without its caveat line, and to treat the card as the terminus of appraisal, never a substitute for it: if the evidence slots are empty, the grade is unearned regardless of what it says.
How it implements the components¶
convergence_claim_record— the card is this record: the one-sentence claim plus its structured, versioned evidence and caveats, held in an auditable and comparable form.recurrence_confidence_grade— the card carries the graded confidence as its headline stamp, always paired with the reasoning that produced it.
This card does not compute the grade or assemble the raw convergence_evidence_set — that weighting is done by Multiple-Origin Evidence Weighting, its nearest twin, which produces the grade this card merely records and communicates. It also does not extract the lesson (constraint_extraction_model) or fix where it travels (transferability_boundary); those are other siblings' work.
Related¶
- Instantiates: Independent Convergence Recognition and Transfer Design — the archetype's target output is an auditable convergence claim, which is exactly this card.
- Consumes: Multiple-Origin Evidence Weighting supplies the graded confidence the card stamps and records.
- Sibling mechanisms: Multiple-Origin Evidence Weighting · Negative Convergence Case Search · Lineage Independence Audit · Hidden Diffusion Checklist · Homoplasy vs. Inheritance Review · Pressure Similarity Matrix · Form–Function Decomposition · Analogy-to-Constraint Extraction Workshop · Cross-Domain Transfer Trial
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: The card appraises a falsifiable convergence claim against standardized supporting, disconfirming, independence, diversity, and performance evidence and issues a reasoned graded confidence, so its operative form is assessment.
Nearest alternative: Record, Log & Register — The dated, versioned card preserves the evolving evidence and grade, but its defining contribution is the evidence-based confidence finding rather than historical accumulation alone.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Statistics & Experimental Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Evidence-appraisal practice cohered standardized graded-confidence summaries that preserve supporting, disconfirming, independence, and limitation evidence for later revision.
Related originating lineages:
- Ethnography & Qualitative Methods — Qualitative evidence appraisal contributes concise auditable summaries that preserve supporting, disconfirming, and caveat evidence.
- History & Historiography — Source criticism contributes documented lineage independence and durable provenance of the claim.
- Philosophy — Philosophy of science contributes the distinction between convergent support and proof, plus explicit defeaters.
Review resolution: Graded evidence summaries are anchored in evidence appraisal but synthesize qualitative triangulation, source criticism, and defeasible epistemic confidence into a compact card.
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] GRADE (Grading of Recommendations Assessment, Development and Evaluation) is a widely used framework in evidence-based medicine for rating certainty of evidence on an ordinal scale with explicit reasons for up- or down-grading. The card borrows its posture — a stated grade is only as good as the recorded reasons behind it. ↩