Repeated MCAS nose-down trim followed erroneous angle-of-attack data. The final report and participating authorities disagree about sensor failure and crew performance.
Display copies resized and converted to WebP without cropping; social copy resized to JPEG. Derivatives licensed CC BY-SA 2.0.
Geographic context
Addis Ababa departure
Nairobi intended destination
Approximate airport markers for the intended route, not an accident flight track or navigation chart. Basemap: Natural Earth, public domain.Read supporting evidence
ET302 developed false angle-of-attack indications immediately after takeoff. The crew faced warnings, repeated MCAS trim and increasing control forces. It cut out electric trim, later restored it and lost control after another nose-down command. All 157 occupants died.
Read the full incident Follow what happened, from the flight to the investigation findings.
What investigators concluded
EAIB identified repetitive uncommanded MCAS inputs as the probable cause. NTSB and BEA agreed MCAS contributed but challenged the treatment of crew actions; NTSB also rejected EAIB's electrical-failure explanation for the sensor, favouring foreign-object impact.
The flight sequence and the authorities' disagreement
1,532 words / about 7 minutes to read. Based on the investigation records cited below.
A normal takeoff followed by false warnings
Ethiopian Airlines flight 302 departed Addis Ababa Bole International Airport on 10 March 2019 for Nairobi Jomo Kenyatta. The Boeing 737-8 MAX, registered ET-AVJ, carried 157 people: 149 regular passengers, two pilots, five cabin crew and an in-flight security officer. The departure airport lay at an elevation of 7,656 feet. The aircraft lined up on runway 07R with flaps set to five degrees and stabilizer trim at 5.6 units. The captain was pilot flying. The recorded takeoff roll and initial liftoff were normal.
Shortly after liftoff, the left angle-of-attack signal changed to an erroneous value while the right signal remained consistent with the flight. The left stick shaker activated and remained active almost to the end of the recording. Airspeed and altitude values derived on the left side diverged from the right-side values. These combined effects created a demanding cockpit environment immediately after departure. The warnings and disagreements were consequences of corrupted information; they did not establish that the aircraft was actually stalled at the time.
The captain attempted to engage the autopilot. Earlier attempts failed, but it subsequently engaged briefly before disconnecting. The flight was already being affected by erroneous indications and automatic guidance before the first MCAS stabilizer command. This distinction is central to the published BEA comments. BEA argued that the initial flight phase deserved a fuller operational analysis and that concentrating on MCAS alone obscured lessons about response to stick shaker, unreliable airspeed, automatic guidance and thrust. The mechanical and operational sequence cannot be reduced to its final automatic trim command.
With the autopilot disconnected, the flaps reached the up position. At about 05:40:00 UTC, the first MCAS nose-down trim command operated for nine seconds in response to the erroneous angle-of-attack data. The stabilizer moved towards a position requiring increased nose-up control force. The captain applied nose-up electric trim, but the initial input did not restore the earlier trim condition. A second automatic nose-down command followed. Electric trim could interrupt MCAS, yet a continuing erroneous input allowed the automatic sequence to recur.
The crew discussed stabilizer trim cutout and the switches were most likely selected to cutout at about 05:40:38. A further recorded automatic command produced no corresponding stabilizer movement, supporting that switch-position interpretation. This is an important factual distinction from the claim that the crew never used the cutout procedure. However, the aircraft remained out of trim, with substantial column forces and increasing airspeed. Removing electrical stabilizer movement did not automatically bring the stabilizer back to a position appropriate for the flight condition.
The captain requested a return and the crew attempted to manage the aircraft with large control forces. Both pilots pulled on the column during portions of the flight. BEA and NTSB regarded thrust management and the amount of nose-up electric trim used before cutout as important parts of the explanation. Their comments argued that excessive speed and insufficient trimming increased the forces on the control column and manual trim wheel. This interpretation is attributed to those authorities rather than represented as an uncontested conclusion of the Ethiopian report.
Restored electric trim and the final loss of control
The aircraft continued with its trim cut out while the crew faced an abnormal flight path, warnings and high control loads. The record later showed short nose-up electric trim inputs, establishing that electric operation had been restored. At about 05:43:11, an autopilot engagement attempt was also recorded. The flight-data parameter could record those electric trim commands only with the cutout switches in the normal position. Restoration allowed the crew to use electric trimming, but also made the automatic MCAS function available again under the continuing erroneous sensor condition.
At 05:43:21, another automatic nose-down trim command operated for about five seconds. The stabilizer moved from approximately 2.3 units towards one unit. The flight path turned into a rapidly increasing descent. The report records a steepening nose-down attitude, increasing airspeed and very high descent rate before both recorders stopped at about 05:43:44. The aircraft impacted a farm field near Ejere southeast of Addis Ababa. The high-energy impact destroyed it and killed all 157 people aboard. No cabin dialogue or imagined final experiences are necessary to describe that established outcome.
The Ethiopian report identified repetitive, uncommanded MCAS nose-down inputs from erroneous angle-of-attack data as the most probable cause. Its contributing factors concentrated on design reliance on one sensor, assumptions about pilot recognition, combined cockpit effects, the missing disagreement alert and information or training deficiencies. Those findings identify substantial aircraft-system and certification concerns. They are presented here as EAIB findings, alongside the recorded sequence, because the final report was followed by formal disagreements that affect how its explanatory conclusions should be read.
The NTSB participated as the accredited representative of the state of aircraft design and manufacture. It reviewed successive draft reports and provided comments. When the Ethiopian final report was issued in December 2022, the NTSB said its most recent comments had not been appended as requested. The report instead linked an earlier version. The NTSB published its May 2022 comments separately and later issued a January 2023 response after reviewing changes in the final document. Both comment sets belong in a complete account of this final report.
In the later response, the NTSB agreed with the analysis of MCAS and related systems' roles in the accident. However, it challenged the final report's explanation that production-related electrical problems and failure of the sensor heater generated the erroneous values. It said the report did not provide supporting details for that electrical explanation. The US team instead concluded that the angle-of-attack vane separated after impact by a foreign object, most likely a bird. That competing explanation was supported by tests and analysis supplied during the investigation.
These are different causal descriptions of the event that corrupted the sensor output. Agreement that the output was wrong and that MCAS responded does not settle why the sensor became wrong. The draft therefore does not merge an electrical defect and a probable bird impact into one established initiating event. It describes the recorded erroneous values as fact, EAIB's electrical explanation as the report's finding, and NTSB's foreign-object explanation as its formal response. The distinction prevents a reader from mistaking disagreement for unanimous confirmation.
The NTSB's earlier comments proposed that the probable cause include the crew's inadequate use of electric trim and management of thrust alongside uncommanded MCAS inputs. It also proposed factors concerning preparation for stabilizer malfunctions and the foreign-object impact. The NTSB criticised statements that aircraft control was impossible, arguing that appropriate electric trim and manual thrust reduction could have maintained control. These are conclusions in its published comments, not an independent reconstruction by this website and not an allegation extending beyond the official investigation record.
BEA shared the Ethiopian report's conclusions about the contribution of MCAS but considered its operational and crew-performance analysis insufficient. Its comments focused particularly on the period before the first MCAS activation. BEA described incomplete execution of relevant abnormal procedures, insistence on engaging the autopilot, continued high thrust and limited crew coordination. It also discussed the first officer's limited experience and the airline's method of disseminating MCAS information. Some interpretations were expressly qualified, including possible stress and likely failure to identify particular indications.
BEA argued that inadequate crew actions and insufficient cockpit resource management belonged in the chain of events. It said the failure to reduce thrust, combined with insufficient trimming, made control-column and manual-trim forces increasingly difficult. The existence of these comments does not negate the technical design findings or turn them into a pilot-only explanation. It shows that participating authorities considered those findings incomplete without an operational account. A Level 3 final-report page must show that disagreement openly, rather than treating final publication as evidence that every participating authority accepted the whole explanation.
Recommendations, changes and the retained uncertainty
EAIB recommended that the manufacturer assess the combined effects of cockpit alerts on recognition and response, provide adequate MCAS training, and reconsider system design so both angle-of-attack sensors were analysed before commands were issued. Its report also documented later safety actions and responses. The NTSB comments described software changes, revised sensor monitoring and training requirements, including simulator exercises for erroneous angle-of-attack conditions. These measures relate to the documented vulnerabilities, but their report-era description is not a fresh assessment of current fleet compliance.
The operational disagreement also informs the research questions left for the reader: what cues reached the pilots, which procedures were applied, how speed and trim affected control effort, and how the airline verified understanding of new information. These questions are grounded in the official comments and should remain connected to their authors. They do not require deciding that one authority's entire report is worthless or that another authority resolves every issue. The safety account benefits from preserving the areas of agreement as well as the specific points of dispute.
The common factual core is a normal takeoff, corrupted left sensor information, persistent warnings, repeated MCAS trim, use and later reversal of trim cutout, and a final rapid descent. EAIB's probable-cause statement and the NTSB and BEA responses offer different emphases and conflicting explanations in parts of that chain. This draft retains all three institutional positions. It states what the record supports, identifies whose conclusion supplies each disputed explanation, and stops short of a new adjudication of responsibility beyond the official material.
Matching dossiers for Formal investigation disagreement: No other dossier in this collection.
Read alongside
These comparisons connect specific research topics. Shared circumstances do not establish identical causes. Evidence from both investigations is provided below.
ET302 and Air France 447 involved difficult combinations of aircraft indications and flight-deck responses. ET302 involved erroneous angle-of-attack data and repeated MCAS trim; AF447 involved airspeed inconsistency and a sustained stall. Comparing recognition and training issues does not equate these mechanisms or resolve the ET302 authorities' disagreement.
Explore the records below for additional context. They open on the publisher's site and are not all included in our report search. Source reuse terms apply.
Official English comments on BEA HTML page accompanying its 3 January 2023 statement; numbered HTML extraction, not PDF
Sources and transparency
Checked against cached official report pages and specified official comments. No independent expert review.
Source limitations
Final report findings are attributed to EAIB. NTSB and BEA formally disputed aspects of its sensor and operational analysis. Their comments are part of this Level 3 account.
Aircraft Accident Investigation Report B737 MAX 8, ET-AVJ
Ethiopian Aircraft Accident Investigation Bureau December 2022 final report, official BEA-hosted copy
Private evidence cache. Only short attributed quotations in original commentary; full report and third-party illustrations are not redistributed. Source terms / provenance.
US comments on the ET302 draft report
NTSB 12 May 2022 comments, published 27 December 2022
Private evidence cache. Only short attributed quotations in original commentary; full report and third-party illustrations are not redistributed. Source terms / provenance.
NTSB response to the ET302 final report
NTSB 13 January 2023 response, published 24 January 2023
Private evidence cache. Only short attributed quotations in original commentary; full report and third-party illustrations are not redistributed. Source terms / provenance.
BEA published observations on the Ethiopian report
BEA Official English comments on BEA HTML page accompanying its 3 January 2023 statement; numbered HTML extraction, not PDF
Private evidence cache. Only short attributed quotations in original commentary; full report and third-party illustrations are not redistributed. Source terms / provenance.
Final Report: Accident on 1st June 2009 to the Airbus A330-203 registered F-GZCP operated by Air France flight AF 447
Bureau d'Enquetes et d'Analyses (BEA) English final report, 5 July 2012
Limited quotations for source-based commentary. Full report and illustrations are not republished. Wider reuse requires checking BEA terms. Source terms / provenance.
Kevin Murphy writes the incident accounts on AirlineIncidents.com, out of a long interest in aviation safety and the lessons each investigation leaves behind.
Prepared by
AirlineIncidents editorial desk
Report citation check
Independent expert review
Not completed
Citation checks compare supporting text with the indexed report pages. They do not constitute independent expert review. How these pages are prepared.
Revision history
: Private landmark Wave 1 original account; final report and specified comments cited.
.jpg){kind=link}