CASE 006 · WATER GOVERNANCE · VALIDATED

Flint Water Crisis: The Revised Report

How did a routine water-quality report change after a conference call with Michigan environmental officials?

RESEARCH STATUSSupported documentary sequenceHuman-reviewed · cross-verified primary-source chain

01 · THE QUESTION

What was supposed to happen?

Under the federal Lead and Copper Rule (LCR), water systems are required to collect a specified number of water samples from high-risk homes. The results determine whether the system must implement corrosion-control treatment. The process is designed around defined sampling criteria and regulatory review.

02 · WHAT HAPPENED IN THE RECORD

The report changed.

In 2015, the City of Flint submitted an initial Lead and Copper Report to the Michigan Department of Environmental Quality (MDEQ). The original report stated that 175 sample bottles and instructions had been distributed to city residents and that the city was still soliciting samples because it had not reached the minimum number for the monitoring period.

A later version of the report states that it was revised after a conference call with DEQ staff. The revised report removed two samples for not meeting sample criteria, reduced the required sample count to 60, and reported 69 samples as taken.

The documentary record therefore contains a traceable sequence: Original LCR → DEQ conference call → Revised LCR → FOIA record documenting the underlying record.

03 · WHERE DID THE RECORD CHANGE?

The divergence is visible in the procedural record.

The material finding is not that the revision was necessarily improper. It is that the official record changed in identifiable ways after a documented conference call: two samples were removed, the required count changed, and the reported sample total changed. The later FOIA record independently preserves the existence and timing of the records surrounding that sequence.

Trace the Evidence

Each source below is represented in Loom's machine-readable Flint artifact record. The hashes identify the exact repository-captured source objects; they do not, by themselves, establish the truth of an underlying claim.

Source 1 · Original Flint Lead & Copper Report · August 20, 2015

Evidence: The original report states that 175 sample bottles and instructions were distributed and that the city continued soliciting samples because it had not reached the minimum number for the monitoring period.

SHA-256: ee7a3384d57bf1fb4a0638c2fa4f4ffe4d994fe73f2c5a47edf3c003dae95576

Source 2 · Revised Flint Lead & Copper Report · September 30, 2015

Evidence: The revised report explicitly states that it was revised after a conference call with DEQ staff. Two samples were removed for not meeting sample criteria; the required count was reduced to 60; and 69 samples were reported as taken.

SHA-256: 8eefeacf0eccd826a6150bfe7083504461a5e74fd73db6b0d7d0f63165f4bbbf

Source 3 · MDEQ FOIA Appeal Response · September 28, 2015

Evidence: MDEQ's appeal response states that its September 14 production included a copy of the Flint LCR and emails discussing the Flint LCR. The FOIA record requested the original report, the conference call that caused the revision, and the rationale for invalidating two samples.

SHA-256: 7b9bade6bb783742d4bf7576ff9212cf6b3dd680be96f4d51d6aa0e98110feac

Epistemic status: Supported · Verification: Cross-verified · Review: Human reviewed

→ Inspect the raw YAML artifacts and provenance records in GitHub

04 · WHAT LOOM CAN AND CANNOT CONCLUDE

Separate the documentary finding from the unanswered question.

Supported: The documentary record shows that the original LCR report was followed by a revised report that identified two removed samples, changed the required sample count, and reported a different sample total. The revised report itself attributes the revision to a conference call with DEQ staff, and the FOIA record independently documents the related LCR records and request.

Not claimed: Loom does not conclude from this sequence alone that the revision was fraudulent, improper, or substantively unjustified. Determining whether the two samples were correctly excluded requires analysis of the specific sample criteria and the underlying sample records.

05

CRYPTOGRAPHIC TRUST SIGNALS

Inspect the artifact record.

The table below is generated from the checked-out repository during deployment. Each SHA-256 value identifies the exact bytes of a repository artifact at that commit. These hashes let a reader check whether a file has changed; they do not establish that a source document is authentic or that a research claim is true.

VALIDATION STATUS: LOADING

Loading the latest machine-readable artifact manifest…

All Flint artifacts validated — GitHub Actions workflow status

The badge reflects the workflow result. A green badge means the configured checks passed—not that every historical claim has been independently proven.

Per-object SHA-256 hashes

Manifest has not loaded.

ObjectType / recordSchemaRepository SHA-256
Waiting for the manifest…

Research artifact files

FileSHA-256
Waiting for the manifest…
06

SHOW THE WORK

Inspect the underlying research record.

Loom treats provenance as part of the finding, not a footnote. The public case page exposes the documentary chain while the repository retains the structured claims, evidence records, source metadata, validation rules, and cryptographic manifest used to audit the case.