Practical Tips · The Demer Factors · Factor Three · July 2026
Flawed methods and procedural irregularities — the third Demer factor, and the only bias evidence available on day one, before a single discovery request is served.
The first factor of the Demer framework measures the expert’s financial dependence; the second measures the expert’s pattern across other insureds’ claims. Both require discovery. The third factor is different in kind: it asks whether the expert’s opinion was formed through flawed principles, theories, and methodologies, or through procedural irregularities in how the evaluation was run — and much of that evidence is sitting in the denial file the day it arrives. Factor three comes third in the doctrine and first in the workflow.
Two layers. The first is the face of the report: does it state a methodology, cite the standards of its field, consider competing explanations, document what was examined, address the evidence that cuts against its conclusion, and qualify what it could not determine? A sound report does these things because the expert’s own field demands them. A flawed report is recognizable by their absence — boilerplate text, categorical certainty, selective photographs, contrary evidence that simply never appears.
The second layer is the process around the report: the reviewer who rejected a claimant’s credibility on paper without ever conducting an examination; the report sent back with narrower and narrower questions until it produced the answer the insurer wanted; the assessment that flipped between drafts with no explanation; the engagement letter that announced the expected conclusion before the work began.
One facial defect deserves special attention because it is nearly universal and nearly always missed: the flawed report cites no sources for the principles it applies. Every expert opinion rests on some general principle — damage of this kind indicates that cause; findings of this kind are inconsistent with disability. Read the report and ask of each such principle: says who? A treatise? A peer-reviewed study? A field standard? In flawed reports the answer, over and over, is no one. An unsourced principle cannot be tested — and an opinion that cannot be tested cannot be effectively contested at the claim stage, which is precisely the point. The omission also departs from the expert’s own professional field, whose standards require investigators to identify the basis for their determinations. Treat every unsourced premise as presumptively invented until the expert identifies its authority.
The Demer illustration. In Demer v. IBM Corp. LTD Plan (9th Cir. 2016) 835 F.3d 893, the compensation metrics raised the inference of a financial conflict — but the methodology completed the proof: reviewing physicians who discounted the claimant’s credibility without examining him and without explanation, and a physical-capacity assessment that shifted from restrictive to permissive with no stated reason. Metrics plus flawed method is the pairing that carried the decision.
In ordinary litigation, a shoddy expert opinion is impeachment material. In first-party insurance it is something more, because the insurer owes its policyholder a duty to conduct a full and fair investigation before denying — the duty California recognized in Egan v. Mutual of Omaha and codified in its claims regulations. California’s gatekeeping law (Sargon Enterprises v. University of Southern California) requires an expert opinion to rest on reliable material and supported reasoning, with the same intellectual rigor the expert’s field practices. It is this project’s position — a reasoned derivation, not yet any court’s holding — that an investigation built on an expert opinion that could not survive that scrutiny is not a full and fair investigation.
Nor can the insurer convert the report into a shield. The genuine-dispute doctrine protects an insurer only when the dispute is genuine: the court that shaped the doctrine (Chateau Chamberay) warned that an insurer may not insulate itself from bad-faith liability “by the simple expedient of hiring an expert” to manufacture one, and listed unreasonable experts, dishonest expert selection, and failure to investigate thoroughly among the circumstances that send bad faith to a jury. In Hangarter v. Provident Life (9th Cir. 2004), a $7.67 million bad-faith verdict stood on exactly that footing — an examiner with a perfect insurer-favorable record and a retention letter, from an in-house consultant who had never examined the claimant, that had already asserted there were no objective findings of disability before the exam took place.
A discipline runs through the case law in both directions. Courts routinely hold that relational metrics alone — volume, prior work, professional acquaintance — do not prove bias; flawed methods and irregularities are the “more” those decisions demand. But a single claim’s irregularities, standing alone, may not suffice either. The reliable architecture is multiplicative: metrics × irregularities × pattern (or the insurer’s failure to take any reasonable measures to secure a neutral expert — the fourth factor). One flawed report is a quality dispute; the same flawed method across many claims is bias.
No treatise yet organizes this factor into a unified doctrine — which is why its best demonstration is historical. The most fully documented biased-expert operation in the industry’s history ran for two decades on a single unsourced principle: a demand for “objective evidence” of conditions that are diagnosed by self-report, imposed by no policy term and grounded in no medical authority. Regulators found it, settled around it, and declared it fixed — and the federal courts spent the next twenty years documenting its continued operation. That story is this factor’s case study: The Reform That Kept the Machine.
Read the deep-dive on Substack → See the bias-evaluation service →
The screen distilled to a field card: Before You Believe the Denial Report: The Twenty-Check Screen. The factor’s case study: The Reform That Kept the Machine. The framework this factor belongs to: the inference-of-bias standard, Factor One (financial dependence), and Factor Two (patterns across other insureds’ files) — all set out in Demer’s Paradigm for Assessing Biased Insurance Experts (Advocate Magazine, 2024).
Distilled from the project’s own analysis: the Factor 3 doctrinal synthesis (2026), treatise ch. 7 (Assessing Experts’ Neutrality and Reliability), and Chris Dion, “Demer’s Paradigm for Assessing Biased Insurance Experts,” Advocate Magazine (July 2024). Educational and informational only; not legal advice.