Evidence Gaps & Diagnostic Bias

← Back to Domain-Specific Families

Abstractions about base rates, denominators, underascertainment, changing case definitions, survival estimates, reporting provenance, and biases in interpreting health evidence.

10 abstractions in this family — domain-specific abstractions that sit near one another in structural-signature space (k-means over structural-signature embeddings). Each is shown with its short description.

• Base Rate Fallacy — The systematic human tendency to underweight or ignore the prior probability of a hypothesis when vivid, specific evidence is available, so a posterior estimate collapses toward the likelihood instead of tracking Bayes' rule.
• Case-Definition Drift — The surveillance failure in which the operational definition of 'a case' silently changes across time, place, or data source while counts are still reported and compared under a stable label — conflating real epidemiological change with an artefact of what the instrument now counts.
• Kaplan–Meier estimator — A nonparametric product-limit estimator of a survival function from observed event times in the presence of right-censoring.
• Martha Mitchell effect — Diagnose the trap in which an evaluator discounts a true but improbable report as delusional purely on its low prior probability, never testing whether its external referent is real.
• Missing Denominator — Expose a count that only looks like information by firing one substrate-independent question — out of how many? — since a numerator reported without its reference population wears the shape of a rate while blocking every quantitative inference the reader needs.
• Munchausen Syndrome — The clinical pattern (DSM-5 factitious disorder imposed on self) in which a person deliberately feigns, fabricates, or self-induces medical symptoms with no external material reward, in order to occupy the valued role of the sick person — with a by-proxy variant enacted on a dependent.
• Original Reporting — The journalistic practice of generating new primary evidence — by interview, on-site observation, document acquisition, or original analysis — rather than re-processing the public record, forming the irreducible base layer that caps the reliability of all derivative work above it.
• Outbreak Underascertainment — The surveillance failure in which recorded case counts fall systematically below the true count because a multi-stage detection pipeline — symptom expression, care-seeking, testing, confirmation, reporting — filters cases with biased attenuation at each layer.
• Pratfall Effect — A small, recoverable blunder raises observers' liking for an agent already judged highly competent — humanizing them by narrowing a status gap — while the same blunder lowers liking for a mediocre agent, so the impression update reverses sign depending on the prior competence estimate.
• Sourcing Ladder — Rank the sources backing a claim by directness, independence, expertise, and accountability, then resolve disagreements by letting the higher rung win and cap a chain's confidence at its weakest load-bearing link.