AutoDefect Research & Data Center: Methodology, Thresholds & Briefs
AUTODEFECT.ORG BMW HVAC EVAPORATOR REGISTRY · INDEPENDENT
MY RECORD REPORT A FAILURE
Evidence before accusation
Source classes and case evidence levels are tracked separately. Every material claim states what it does — and does not — establish.
OPEN THE EVIDENCE CENTER →
Symptom ≠ diagnosis
Uneven vents, hissing or temporary post-recharge improvement are diagnostic clues — not proof the evaporator is the source.
READ THE DIAGNOSIS GUIDE →
Three cost layers, never blended
External benchmarks, documented estimates and completed invoices answer different questions — the registry keeps them separate.
SEE COST INTELLIGENCE →
Public record, plainly labeled
Recalls, settlements and complaints are different kinds of records. Each entry links its primary source and states its scope.
OPEN THE PUBLIC RECORD →
Statistics describe submitted records
They do not establish manufacturer-wide incidence or failure rates. Read how records are reviewed before publication.
VISIT THE DATA CENTER →
SEARCH AUTODEFECT.ORG
Prototype — search binds to Supabase full-text in the production build (see handoff-v4).
POPULAR SEARCHES
X5 G05 Evaporator replacement cost Passenger side warm 64119496922 BMW recalls Class action settlements R-1234yf leak
Analyst wall of charts and registry data visualizations
RESEARCH & DATA CENTER

AutoDefect Research

Evidence briefs, methodology, registry data summaries and dated revisions — published only when the underlying records support them, at thresholds designed to keep a small registry honest.

METHODOLOGYNEWSROOM

LIMITATION — No public dataset download exists yet. Dataset structured data will be deployed only when a real, reviewed, anonymized collection is published.

Published: August 8, 2026Last reviewed: August 8, 2026Compiled by: AutoDefect.org Registry EditorSources last checked: August 8, 2026CorrectionsManufacturer right of reply
WHAT THIS CENTER PUBLISHES

Five output types

OUTPUTSTATUS
Evidence briefsFirst brief in preparation: documenting the founder case end-to-end (symptom → diagnosis → estimate → goodwill review)
Methodology & data dictionaryPublished — variables collected, inclusion criteria, evidence levels, privacy and redaction, review steps, limitations
Registry data summariesPublished per the thresholds below as approved records accumulate
Dated revisions & correctionsEvery substantive change is dated; corrections remain visible
Downloadable figuresCharts with source notes, released alongside briefs
DATA & METHODOLOGY

What the registry collects

Variables include model year, vehicle model, chassis, mileage at diagnosis, diagnostic spend, repair estimate, completed repair cost, warranty status, goodwill request and contribution, owner-paid total, evidence classification and document types on file. Diagnostic trouble codes are recorded only when they appear in submitted documents.

Case inclusion requires document review; published records are sanitized derivatives. Contradictory and alternative-diagnosis records are retained under Level E.

FULL METHODOLOGY & DATA DICTIONARY →
PUBLICATION THRESHOLDS

Editorial safeguards, not statistical guarantees

RECORDSWHAT MAY BE PUBLISHED
n < 3Individual approved records only — no percentages, no granular cross-tabs
n = 3–9Counts and ranges; no implied stable distributions
n ≥ 10Medians, with explicit n and the self-selected-registry limitation
n ≥ 20Additional descriptive statistics may be considered — still without claiming representativeness

Current status: n=1 (founder case). The registry publishes the record, not statistics derived from it.

PUBLIC DATASET CRITERIA

When a downloadable dataset will exist

A public, anonymized dataset will be released only when it can be real: an accessible table or downloadable file derived from reviewed records, with a published data-use policy, privacy review, and version history. Until then this page carries no Dataset structured data, no invented license, and no “open data” placeholder.

Researchers and journalists welcome.

Methodology questions, data requests and verification inquiries are answered through the newsroom.

NEWSROOM & MEDIACONTACT
CONTINUE IN THE EVIDENCE ARCHITECTURE
MethodologySource libraryNewsroomRegistry recordsEvidence center