E01¶


Research	R0050 — Journalism Disciplines
Run	2026-03-31-02
Query	Q002
Source	SRC04
Evidence	SRC04-E01
Type	Factual

FMEA's Risk Priority Number contributes a novel three-factor composite risk scoring methodology

URL: https://en.wikipedia.org/wiki/Failure_mode_and_effects_analysis

Extract¶

FMEA calculates a Risk Priority Number (RPN) using three independently scored factors:

RPN = Probability (P) x Severity (S) x Detection (D)

Each factor uses a defined scale (typically 1-5 or 1-10): - Probability: How likely is the failure? (1 = extremely unlikely to 5 = frequent) - Severity: How bad is the consequence? (1 = no effect to 5 = catastrophic) - Detection: How likely is it to be caught before causing harm? (1 = certain detection to 5/6 = undetected)

Novel concepts not in reference frameworks: 1. Detection dimension: None of the nine reference frameworks include a formal assessment of how detectable an error/failure is. The detection axis is unique to engineering safety analysis. 2. Composite risk scoring via multiplication: The RPN approach of multiplying three independent assessments to produce a single priority score is structurally distinct from GRADE's domain-based quality assessment or ICD 203's single probability scale. 3. Proactive failure anticipation: FMEA analyzes what could fail before it does, whereas most reference frameworks evaluate evidence about what has happened.

Known limitation: The RPN multiplication approach has been criticized because ordinal scales do not validly support multiplication, leading to "rank reversals" where different factor combinations produce the same RPN despite different risk profiles.

Relevance to Hypotheses¶

Hypothesis	Relationship	Strength
H1	Strongly supports	Multiple novel concepts
H2	Contradicts	Clearly novel, not redundant
H3	Strongly contradicts	Unambiguously formal

Context¶

FMEA was developed by the US military in the 1940s and has been adopted across aerospace, automotive, healthcare, and other high-reliability industries. The detection dimension is particularly interesting because it addresses a meta-level question about evidence evaluation: not just "is this true?" but "would we know if it were false?"