Every pilot knows the scene. The passenger is already aboard, the meeting time is tight, the weather at the destination "should improve," and every new piece of information that arrives is read as confirmation that it is possible to go. When the information contradicts the plan, it is discounted: the wet runway "is not that wet," the tailwind "is only a few knots," the approach outside parameters "can still be corrected."
Aviation has had a name for this for decades: plan continuation bias, or, in cockpit jargon, get-there-itis, press-on-itis. The point that research has made clear, and that the culture of many operations has yet to absorb, is that this is not a defect of a few pilots. It is a mechanism of human cognition, one that shows up in experienced pilots and that commercial pressure turns into operational risk.
What NASA found in 19 accidents
In 2006, Benjamin Berman, former chief of investigations at the NTSB, and Key Dismukes, chief scientist for human factors at NASA's Ames center, published in Aviation Safety World the summary of a study of the 19 airline accidents in the United States, between 1991 and 2000, in which the NTSB cited crew error as a factor. The central conclusion: most experienced pilots, placed in the same situation and knowing only what those crews knew, would be vulnerable to the same errors.
Two themes run through the cases. The first is plan continuation bias, evident in at least nine of the 19 accidents. The authors recall that a Flight Safety Foundation task force had already found "press-on-itis" in 42% of the accidents and incidents it reviewed. The second is the snowballing workload: the sequence of small demands that pile up on the approach, reduce the time available, and leave the crew behind the airplane. The two feed each other: the further behind the pilot gets, the less capacity is left to reassess the plan.
The research by Judith Orasanu and colleagues at the same NASA describes the mechanism. The pilot's decision has two parts: assessing the situation and choosing the action. The bias acts on the first. The cues that the situation has changed are ambiguous, arrive gradually, and are interpreted in light of the plan that already exists. None of them, alone, seems reason enough to abandon the plan. And when the whole picture finally becomes clear, the options have already narrowed. Continuing an unstabilized approach is the typical example.
The number that shows the size of the problem
The final report of the Flight Safety Foundation's go-around project, discussed in the article on stabilized approach and go-around, measures the bias at scale: unstabilized approaches occur in 3.5% to 4% of landings, and 95% to 97% of the crews in that situation continue, even though company policy calls for a go-around. Continuation bias is not a rare case of poor judgment. It is the default behavior.
In Brazil, ANAC's Flight Alert AV 01/2016, on the Cessna 560 that overran the runway at Jacarepaguá, ends with the same recommendation to air taxi companies: improve crew judgment regarding destabilized approaches, "always considering the possibility of a go-around."
The five hazardous attitudes, and why the antidote is not enough
The FAA's Risk Management Handbook catalogs five hazardous attitudes that degrade pilot decision making, each with its antidote: anti-authority ("don't tell me what to do"), impulsivity ("do it quickly"), invulnerability ("it won't happen to me"), macho ("I can do it"), and resignation ("what's the use"). The same handbook recalls the NTSB statistic that, over the last two decades, about 85% of accidents had pilot error as a cause.
The hazardous attitudes explain part of continuation bias, but not all of it. The study by Berman and Dismukes indicates that the bias arises from the interaction of social and organizational influences, the limits of human cognition, and the situation itself. The pilot who continues is not, most of the time, the handbook's macho. It is the competent professional who has recalibrated, flight after flight, what counts as acceptable risk, because they continued many times and it worked out. With each successful landing in a marginal condition, the marginal condition becomes normal.
The pressure the company creates without noticing
In air taxi and business aviation, the passenger is the customer, sits two meters from the cockpit, and often pays for the flight just requested. The pressure to continue does not need to be spoken. It is present:
- In the schedule that makes landing at the destination the only acceptable option, with no alternate agreed before takeoff.
- In the sales side that sold the flight with a fixed time to a runway that cannot take the airplane at the planned weight.
- In the absence of a cost to continue and the visible cost of giving up: diverting means explaining to the customer, paying for a hotel, rebooking.
- In the culture where the pilot who went around or diverted must justify it, and the one who continued and got away with it is praised for efficiency.
- In experience on the type. ANAC's alert records that the two pilots of the Cessna 560 had 3,000 and 4,700 total hours, but only 100 and 30 hours on type. General experience gives confidence; experience on the model is what gives margin.
None of these items appears in a report as "passenger pressure." They appear as "crew error in judgment." The judgment failed in the cockpit, but the conditions for the failure were set up in the office.
Taking the decision off the pilot's shoulders at the wrong moment
The defense against continuation bias is not to ask the pilot to resist pressure better at the moment it is greatest. It is to move the decisions earlier, to when the bias is not yet acting, and leave to the cockpit only the execution of what has already been decided:
- Numerical criteria defined on the ground. Minimum runway, maximum tailwind, minimum ceiling and visibility, stabilization window. A number decided calmly is not negotiated on final.
- An alternate agreed before the flight, with the passenger informed that it exists. The alternate the customer already knows costs less to use.
- Tiered decision authority. A flight with high risk in the preflight assessment requires approval from the operations manager. The pilot stops being the only one who says no.
- Go-arounds and diversions without consequence, treated as normal procedure and recorded as a prevention event, not a failure.
- Measure and show. How many flights were diverted or went around this month? If the number is zero, the bias is operating, and the company does not see it.
Where Alis comes in
The Alis FRAT is the decision made beforehand. The preflight risk assessment cross-references aircraft, route, weather, nature of the flight, night flight, over-water segments, mountainous terrain, and the threats reported by other operators at the same aerodrome, and returns a score with the corresponding decision authority: low risk follows the plan, medium risk calls for a review of the factors with the pilot in command, high risk requires approval from the operations manager and mitigation of the factors. The pilot arrives at the airplane with the hard decision already made by someone who had the time and distance to make it.
The Flight Kit places beside that decision the runway dimensions, the aerodrome elevations, the weather, the NOTAMs, and the estimated weight, so that the short runway or the weight above plan shows up in the planning. And the flight record, in the logbook, leaves the trace that the diversion or go-around was made, so that the number "zero go-arounds this month" is not an illusion.
For the risk assessment method, the article on FRAT and PAVE describes how the factors are structured; and the one on TEM and CRM in practice covers how the crew manages the threat that arrived anyway.
Want to see the FRAT with decision authority by risk level? Book a demo.
This article is informational. The conclusions on continuation bias are those of Berman and Dismukes (NASA Ames, "Pressing the Approach," Aviation Safety World, December 2006) and of Orasanu and colleagues (NASA, 2001); the hazardous attitudes and the NTSB statistic are those of the FAA's Risk Management Handbook (FAA-H-8083-2A); the go-around data are from the final report of the Go-Around Decision-Making and Execution Project (Flight Safety Foundation, 2017); the Brazilian case is that of ANAC's Flight Alert AV No. 01/2016. The text does not replace the guidance of the person responsible for operations or the company's general operations manual.