What does the Air Canada Flight 759 near miss tell us about expectation bias? The 2017 incident shows how people can continue interpreting conflicting evidence through a familiar picture of the world. The U.S. National Transportation Safety Board (NTSB) found that the crew misidentified a taxiway as a landing runway, with ineffective review of closure information, expectation bias, fatigue, unused backup guidance and communication-related factors contributing. The lesson is not that experienced people should simply concentrate harder. It is that accurate perception depends partly on how information and independent checks are designed.
What happened at San Francisco in July 2017?
On July 7, 2017, Air Canada Flight 759, an Airbus A320 arriving from Toronto, was cleared to land on runway 28R at San Francisco International Airport. The parallel runway, 28L, was closed. The flight crew instead aligned with taxiway C, where other aircraft were waiting. The crew initiated a go-around at low altitude, avoiding a collision. The NTSB documented the incident in its 2018 report, AIR-18/01.
The close approach makes the event memorable. But the more instructive part is what happened before the aircraft reached the taxiway. The runway closure was documented in information available to the crew. The pilots did not retain effective awareness of it. They approached with a mental picture that did not match the airport's actual configuration, and some of the visual information they encountered appeared to fit that picture.
The NTSB determined that the probable cause was the crew's misidentification of taxiway C as the intended runway, resulting from ineffective review of notice-to-airmen information about the closure. It also identified expectation bias, fatigue, failure to tune backup instrument guidance, breakdowns in crew resource management and ineffective presentation of approach and notice information as contributing factors. Those are the Board's findings, not an AeroLeadership reconstruction of private cockpit thoughts.
Why did conflicting visual cues not correct the mistake?
Expectation bias is the tendency for an anticipated situation to shape how incoming information is interpreted. It is not necessarily a conscious refusal to accept facts. When a familiar pattern seems to be present, the mind can assign meaning to ambiguous cues before those cues are carefully tested against alternatives.
At San Francisco, the crew's familiar expectation of two parallel illuminated runways was no longer appropriate because one runway was closed. According to the NTSB, available cues did not sufficiently overcome the crew's belief that the taxiway was the intended landing surface. The investigation also found that salient cues near the airport seawall likely helped prompt the eventual go-around. In other words, some evidence failed to change the interpretation until the aircraft was much closer to the hazard.
This is a useful distinction. A signal can exist without being recognized as contradictory. More information is not necessarily better if the information arrives in a form that reinforces the existing interpretation, or if the crucial exception is not noticed at the point where it matters.
How did fatigue and information presentation matter?
The NTSB concluded that both pilots were fatigued because of circadian disruption and extended wakefulness. Fatigue likely contributed to runway misidentification, continuing expectation bias and the delayed go-around. That does not mean fatigue alone explains the event. The Board's analysis treated several conditions as interacting contributors.
The report also criticized how approach procedure and closure information were presented. An important detail can be technically available yet operationally ineffective if it is difficult to distinguish from surrounding material or is not retained during the decision. This does not excuse an ineffective review. It widens the question from whether information was supplied to whether the information system helped the user notice the exception.
The crew also did not tune the instrument landing system frequency for backup lateral guidance. An independent source of position information could have challenged the visual interpretation. The relevant principle is not that one particular instrument should be copied into business. It is that checks with genuinely different failure modes can reveal mistakes that repeated use of the same assumptions will miss.
What can organizations learn without pretending a boardroom is a cockpit?
Businesses are not airports, and the time scales, regulatory controls and consequences differ. Still, the information problem has a recognizable structure. A project team expects a supplier to deliver, so an ambiguous status update is interpreted as routine. A senior manager expects a product launch to be on schedule, so warning signs are explained as temporary noise. A financial team expects a familiar reconciliation to work, so a discrepancy receives a familiar explanation before alternatives are explored.
These are illustrative scenarios, not findings from the Air Canada investigation. The transferable question is narrower: when people have a strong expectation, what information is likely to challenge it, and can that information reach them in a usable form? A second opinion that shares all the same assumptions may not be independent. A dashboard that simply repeats the same data may add confidence without adding a meaningful check.
Organizations also have to think about the cost of challenging an apparently settled picture. If conflicting evidence is seen as obstruction, people may hesitate to raise it. If information is buried in large reports, important exceptions can be technically disclosed but practically invisible. These are organizational design considerations, not proof that every aviation safeguard will improve business performance.
Why this case is more than a story about individual error
It would be easy to end with the instruction to pay closer attention. But the NTSB's findings go beyond attention: information presentation, backup guidance, fatigue and crew coordination all mattered. Each influenced the conditions in which the pilots had to perceive and decide. A person can remain responsible for a decision while the system around that decision is also examined.
This is closely related to the broader problem of whether inconvenient information reaches decision-makers. The Air Canada case is not an employee-silence case, but both topics raise a common question: does the information needed to challenge a current understanding actually influence action?
The important insight is not that expectation bias can be eliminated. It is that familiarity can make a mistaken picture feel increasingly plausible, especially when independent evidence is missing or poorly presented. The challenge for reliable organizations is to preserve the possibility of correction before the cost of being wrong becomes irreversible.
Primary source and scope
The factual account and investigation findings in this article come from the NTSB's 2018 report, AIR-18/01. The Board's findings are also summarized in subsequent congressional testimony. The business examples and discussion of organizational transfer are AeroLeadership analysis, not NTSB conclusions. The case is used here to examine one bounded human-factors issue rather than reproduce a full accident investigation or a course case study.