How to evaluate a technical claim

Start with the actual claim

A good technical statement identifies the subject, the conditions, the quantity being measured, and the limits of the measurement. “This part is strong” is not testable; a statement about a specific material condition, load case, and observed failure threshold is.

Do not collapse a datasheet property, a simulation prediction, and a tested assembly into the same category of evidence. They answer different questions.

An evidence ladder

• Source quotation: a traceable statement with document title, revision, date, and page or section.
• Derived calculation: the equation, units, assumptions, and inputs can be independently checked.
• Simulation: solver, boundary conditions, mesh sensitivity, and uncertainties are documented.
• Bench observation: apparatus, calibration, specimen count, and raw results are retained.
• Independent replication: a second reviewer reproduces the result without relying on the author’s undocumented assumptions.

Keep conclusions proportional

“Tested once” does not mean “production qualified.” A dimensional inspection does not establish fatigue life. A successful prototype does not prove the production process is consistent.

This handbook distinguishes sourced explanation from verified experimental evidence. Draft chapters are labeled until their technical claims receive separate review.