Decision-Making Failures in the Boeing 737 MAX Crisis: A Cognitive and Bayesian Network Analysis

Abstract

This report analyses the Boeing 737 MAX crisis as a systemic decision-making failure within a safety-critical engineering environment rather than as the consequence of a single technical defect. It examines two interconnected decision situations: Boeing’s attempt to preserve design and training continuity with previous 737 aircraft, and the subsequent reassessment of risks associated with the Manoeuvring Characteristics Augmentation System (MCAS). A cognition-driven model is first applied to evaluate how anchoring, overconfidence, bounded rationality and confirmation bias influenced the continuity decision. The findings indicate that commercial pressure to maintain the 737 platform’s operational familiarity contributed to limited pilot training, insufficient disclosure of MCAS and the underestimation of automation-related risks. A Bayesian network model is then used to illustrate how emerging warning evidence should have altered the assessment of systemic MCAS risk. The estimated probability of high systemic risk increases from 30.0% to 66.7% following severe accident evidence and to 84.2% when low pilot awareness and weak regulatory scrutiny are also considered. The analysis demonstrates that the crisis resulted from the normalisation of unsafe assumptions and inadequate updating of risk assessments. It concludes that safety-critical engineering decisions require transparent assumptions, independent scrutiny and systematic reassessment when new evidence emerges.