The same person can use mental simulation in different instances to achieve success one time yet meet failure the next. Will Rogers III is the former commanding officer of an AEGIS cruiser, the USS Vincennes. As part of a project we did for the U.S. Office of Naval Research, I interviewed him about several incidents involving time-pressured decision making.
Example 6.1 The Circling F-4s
The Vincennes has had to cut short its training exercise in April 1988 to go to the Persian Gulf during the Iran-Iraq War. The war has escalated, and each side has been attacking merchant shipping. The U.S. Navy has intervened to protect the merchant ships, the oil tankers from Kuwait, and all the other maritime traffic. The Iranians have purchased Silkworm missiles from China, which are capable of attacking the U.S. Navy forces. At that time, the navy has only one type of vessel capable of neutralizing the Silkworms: the AEGIS cruiser, designed to handle numerous attacking airplanes in a blue ocean setting. That means that they were designed to operate in the open seas, not for narrow bodies of water such as the Persian Gulf. Nevertheless, because no other ship can do the job, the Vincennes is in the thick of the Iran-Iraq War.
The Vincennes is given a mission to escort the flagship Coronado through the Strait of Hormuz during daylight. During this mission, Captain Rogers notes that two Iranian F-4s are taking off from Bandar Abbas airport, to the east. He expects that they are on a routine patrol and waits for them to head north or south. They violate his expectancies by circling the end of the runway. He notices that the circles are getting wider and wider, and with each widening, the F-4s are getting closer to the Vincennes and the Coronado. Suddenly the lead F-4 turns and shifts from its search to its target acquisition mode on its radar (used to lock on to a target when preparing to fire a missile at it). These are not friendly actions. Under the then-existing rules of engagement for the Persian Gulf, aircraft using radars to lock onto ships were considered to have committed hostile acts.
Rogers has the prerogative to respond in kind, to protect his ship; however, he does not believe they will attack. They are so visible, so vulnerable, that he believes they are merely harassing him. Besides, the U.S. Navy is in the Persian Gulf to reduce hostilities, not to increase them. He puts himself in the boots of the pilots. Would he come out, with no maneuvering and no distractions and no covering force, in daylight, and attack a ship with better missiles and better radar? Or would he play some simple games with a U.S. Navy ship? That is something Rogers could imagine. So instead of shooting the two airplanes down, the Vincennes uses its advanced electronics to jam their radar.
Captain Rogers is careful to monitor the range of the F-4s in case they do get too close. He will not put his ship or the Coronado in danger. Soon the Iranian airplanes have had enough, and they fly off.
Captain Rogers achieved what he wanted: to protect his flagship and defend his own ship, without shooting at the F-4s. He was able to draw on his experience to imagine what was going on inside the minds of the pilots. Because he could not imagine how the Iranian pilots could realistically be preparing to attack him, he believed he had the situation well in hand and did not allow himself to become provoked. Sometimes we diagram incidents like this. Figure 6.1 gives an example. The incident starts at the top left and moves to the right. New information is shown along the top. The columns represent how Captain Rogers was thinking about the situation.
Figure 6.1 The harassing F-4s: How the situation was perceived as it evolved
Captain Rogers was involved in another incident; the shooting down of an Iranian airliner. He used the same strategy, trying to imagine what the pilot was thinking, what the pilot’s intent was. He was again able to rule out one explanation as implausible. Based on his mental simulation, he made a diagnosis that the track on his screen was not a commercial airliner. This time he was mistaken.
Chronology of the Vincennes Shootdown
On the morning of July 3, 1988, the USS Vincennes fired two missiles at an Airbus 300, Iran Air flight 655, which had taken off from the Bandar Abbas airport. (The map detail is shown in figure 6.2.) The departure time was 10: 17 A.M. local time. The missiles were fired at 10:24. The entire flight lasted seven minutes, eight seconds. Approximately three minutes, nine seconds elapsed from the time the air track was judged to be tactically significant until the missiles were launched. Shortly after, the Vincennes and the rest of the world learned that the track had been a commercial airliner on a regularly scheduled flight from Bandar Abbas to Dubai, in the United Arab Emirates.
This incident is widely used as a demonstration of faulty decision making under stress. There was time pressure, fear, and uncertainty. Many have claimed that there were dear flaws or biases in the decision making of the crew of the Vincennes.
Prior Events
Hostilities between the Iranian military and the U.S. Navy had been steadily worsening. Besides the case of the circling FAs, on April 18, 1988, there had been a battle in the same area in the Persian Gulf, involving another Navy ship, the USS Wainwright. At that time missiles had been fired at an Iranian F-4, damaging it. In mid-June 1988, the Iranians had transferred F-14s to the Bandar Abbas airport, which was used for both military and commercial activities. In one case an Iranian F-14 flying toward a U.S. cruiser had been warned and had broken off after the cruiser locked onto it with fire control radar. In another incident, a commercial airliner had taken off from Bandar Abbas with an Iranian F-4 flying just below, tucked underneath it to avoid radar detection. The U.S. Navy had modified its rules of engagement after the USS Stark had been hit by an Iraqi fighter. The new rules gave commanders more freedom to take steps to defend themselves, even if the attacker did not fire first.
Another issue was the use of an Identification Friend or Foe (IFF) system to distinguish aircraft. Commercial and military airplanes are outfitted with electronic devices to signal their characteristics. If another piece of electronic gear sends a signal, the electronic device will respond. In the United States, if the response is a Mode IV type, that shows that the aircraft is military and friendly, having been configured with the code word of the day. Commercial airplanes are rigged to send Mode III signals back when queried. Military aircraft from other countries will send back a Mode II signal. However, in the past, Iranian military aircraft had been observed to send Mode III signals, pretending that they were commercial.
Also, there had been recent U.S. operations against the Iranians, the Iraqis had just scored some military successes, and intelligence information suggested that the Iranians intended to take some provocative action during the July 4 weekend, perhaps kamikaze-type attacks on shipping. The Iranian Revolutionary Guard Corps began a series of attacks on commercial shipping the evening of July 2, 1988, prestaging some larger actions. The stage was set.
On July 3, 1988, the USS Elmer Montgomery was surrounded and attacked by thirteen gunboats belonging to the Iranian Navy. At 10:00 A.M. the Vincennes was thirty-five to forty miles away from the Montgomery. It headed over to help, and sent a helicopter ahead to inspect the situation. The helicopter was fired on, and as the Vincennes reached the area, it too was attacked by the gunboats. The gunboats split into two groups, and a surface battle commenced.
It may seem strange that the Vincennes, a large U.S. Navy cruiser, would feel threatened by small gunboats, but its job was anti-air warfare. It was not built to deal with small and fast surface targets; fire from the gunboats was capable of damaging the Vincennes’s superstructure and injuring sailors. The Vincennes had only two guns that were useful in firing at the gunboats. One of these fouled, forcing the ship to maneuver sharply to bring the other gun into play. This was the surface situation into which the Airbus flew.
Example 6.2 The Flight of the Airbus
The Iranian Airbus took off at 10:17 A.M. local time, twenty-seven minutes after its scheduled departure time. Its flight path carried it directly over the Vincennes. Flight 655 climbed throughout its flight and had reached 12,000 feet, on its way to 14,000 feet, when it was hit. Here is the chronology of how it all happened that morning.
10:17 The computer system onboard the Vincennes classified the track as “unknown, presumed enemy,” because the plane was taking off from an Iranian mixed-use airport. The Vincennes assigned it track number 4474. (Each object in the computer—whether an airplane, surface ship, or submarine—is given a track number, which is its name. This track number enables different crew members to know they are talking about the same object when they discuss it.)
10:18 A second ship in the area, the USS Sides, picked up the track of Flight 655 and gave it a different track number, 4131. The computer system network judged that tracks 4474 and 4131 were the same aircraft. The computer assigned it track 4131, the one selected by the Sides, rather than the track number the Vincennes had been using. It is unclear whether this shift in track number was announced over the Vincennes’s internal communications net. At the same time, one of the Vincennes crew members checked the commercial air schedules and noted that Flight 655 was supposed to depart at 9:50 A.M. No airliners were scheduled to depart Bandar Abbas at 10:17. Also during this minute, the Vincennes noted that an Iranian P-3 airplane was in the area. P-3s serve surveillance functions and can be used to direct other aircraft during missions. The Airbus was forty-four nautical miles away from the Vincennes at this point.
10:19 The Vincennes issued its first challenge to the unknown aircraft, warning it to avoid the area. During the following minutes, it issued several more challenges using both military and commercial radio frequencies. The track never responded to any of these challenges. Captain Rogers found this unusual. In the past he had twice issued warnings to commercial airplanes, and each time he had received an answer. He assumed that commercial pilots routinely monitored the international distress frequency as they were supposed to. The Vincennes might have tried calling air traffic control, but it did not have an extra radio on board to let it contact both air traffic control and the track of interest.
10:20 At this point several crew members in the Vincennes Combat Information Center reported they saw a Mode II return from the IFF, suggesting that they were dealing with an Iranian military aircraft, probably an F-14. We now know that this identification was mistaken. Throughout the incident, the crew in the Combat Information Center of the Vincennes believed that it was dealing with an Iranian F-14. The AEGIS computer system does not incorporate data about IFF returns, so it did not have this erroneous piece of information. The designation was entered into the large screen displays. The identification was also communicated over the internal voice net. The air engagement decision properly begins at this point, with the idea that the Vincennes was facing an F-14 flying directly at it and ignoring radio warnings. Not everyone accepted this judgment. One crew member told Captain Rogers that he thought the track was a commercial airliner.
The tactical action officer in charge of managing the air war requested that the Vincennes defend itself by firing missiles at the track when it reached twenty nautical miles. Captain Rogers denied the request. The aircraft was not using any systems that would normally be associated with aggressive actions, such as search radar or fire control radar. Rogers’s mission was to reduce conflicts, not increase them. He wanted more information before firing, even if it meant adding to the risks. He had confidence in the AEGIS system. Besides, he did not believe a fighter pilot would attack a ship as sophisticated as the Vincennes in such a reckless way. The Vincennes turned on its fire control radar, to prepare to fire at the aircraft if necessary, but also to send a message to the pilot that he was in the Vincennes’s sights. The aircraft still was not responding to radio warnings. Moreover, it was departing from the centerline of the commercial air corridor. The corridor was 10 miles wide and the track was within this space, but as a rule commercial aircraft flew down the center “like a bug on a wire,” as Captain Rogers described it. By the time it was hit by the missiles, the aircraft was 3.35 nautical miles off of center, in the direction of the Vincennes.
10:22 Captain Rogers asked, “What is 4474 doing?” Several crew members onboard the Vincennes reported that the aircraft was now descending, an unhappy sign, since a commercial airliner would have still been ascending. According to these crew members, the track was assuming a classical attack profile of increasing speed, decreasing altitude, and closing range. However, some crew members saw the track as still increasing in altitude. At any rate, the Vincennes was now within range of air-to-surface missiles that the aircraft might be carrying.
10:24 Captain Rogers could wait no longer. His own weapons systems would become ineffective if he let the aircraft get any closer. He had waited until the last possible moment. Now he ordered that the missiles be fired. The Airbus was hit at 13,500 feet, eight nautical miles away.
As far as I can see, the decision to fire was fairly straightforward, based on mental simulation. The cues were all consistent with a hostile aircraft:
The IFF report showing that it was an F-14.
The failure to respond to radio warnings.
The timing of the takeoff to correspond to the attack by the gunboats.
The final descent toward the Vincennes.
This story all fits. There were some discrepancies, as shown in table 6.1, but these were small and easily explained away. The aircraft was not using radar or other electronic means of targeting, but this could be explained because the Iranian P-3 in the area could be providing targeting data. Or the pilot might have been trying for visual identification, intending to use fire control radar when he got closer. A second discrepancy was that the aircraft showed Mode III IFF, as if it was a commercial airliner, but Iranian fighters had played that trick before. Third, the aircraft ignored the radio warnings, as expected if it was up to no good.
Table 6.1 The Vincennes Shootdown: Story Discrepancies
Attacking F-14 Story Discrepancies
Commencial Airline Story Discrepancies
Not generating any radar emissions
Flying into the middle of a surface battle
Generating a Mode III IFF signal
Not on the centerline of the air corridor
Ignoring the Vincennes’s military air distress radio calls
Takeoff was not on the schedule
Ignoring the Vincennes’s international air distress radio calls
Generating a Mode II IFF signala
Descending toward the Vincennesa
a. Later determined to be incorrect.
In contrast, the story that the plane was a commercial airliner had too many holes in it. Why would air traffic control vector an airplane directly into the center of a surface battle? Why didn’t the pilot fly straight down the center of the air corridor? Why didn’t the takeoff correspond to the commercial schedule? Why did the cockpit ignore the radio warnings issued over the international distress frequency? More troubling, why was it sending Mode II IFF signals? Most troubling of all, why was the airplane descending (as was reported by the majority of crew members)? It is hard to build a story that can fit all of these discrepancies. The diagnosis that the track was a commercial airliner was rejected without difficulty.
It is easy to blame Captain Rogers for making the wrong decision, since hindsight shows that he was mistaken. But we do not need hindsight. The commander of the USS Sides, viewing the same picture, concluded that the aircraft was a commercial airliner. Certainly the Sides did not have a potential F-14 bearing down on it and was not being attacked by a fleet of gunboats. Still, the Sides read the situation correctly, even under time pressure. Why?
The Vincennes was working with two pieces of inaccurate data: Flight 655 never squawked Mode II IFF, and it never descended. The Sides had accurate readings on both issues. Once the Vincennes’s crew became convinced that the track belonged to an F-14, that assumption colored the way they treated it and thought about it. Once they believed it was descending toward them, there was nothing more to wonder about. Of these two mistakes, the first was the more important. If the Vincennes had never received any Mode II IFF signals, the crew would have been more likely to believe that it might be a commercial airliner.
The first mistake is easy to explain. One of the crew members in the Vincennes’s Combat Information Center was in charge of identifying tracks. He used his remote control indicator to query the aircraft’s transponder, working with the IFF system. He successfully hooked Flight 655 as it departed from the airport, but left it hooked for almost ninety seconds. The way the system worked, the range gate stayed positioned at the end of the runway of Bandar Abbas airport, even as Flight 655 moved toward the Vincennes. During that time, airplanes on the ground would be interrogated by the IFF system. If a military aircraft happened to be positioned within the area covered by the range gate, it would reply with a Mode II response. By coincidence, an Iranian military aircraft did take off from Bandar Abbas just when the Vincennes received the Mode II response. The military airplane was squawking Mode II. So there was no decision error here, just a human error that let the Airbus become correlated with a Mode II signal reserved for military airplanes.
The second mistake is the more controversial one: judging that the Airbus was descending. In fact it ascended throughout its entire flight until hit by the missiles. The computer system reported continual ascent, and the USS Sides saw continual ascent. The Fogarty report, the official U.S. Navy analysis of the incident, concluded that “stress, task fixation, an unconscious distortion of data may have played a major role in this incident. [Crew members] became convinced that track 4131 was an Iranian F-14 after receiving the ... report of a momentary Mode II. After this report of the Mode II, [a crew member] appear[ed] to have distorted data flow in an unconscious attempt to make available evidence fit a preconceived scenario (‘Scenario fulfillment’).”
This explanation seems to fit in with the idea that mental simulation can lead you down a garden path to where you try to explain away inconvenient data. Nevertheless, trained crew members are not supposed to distort unambiguous data. According to the Fogarty report, the crew members were not trying to explain away the data, as in a de minimus explanation. They were flat out distorting the numbers. This conclusion does not feel right.
The conclusion of the Fogarty report was echoed by some members of a five-person panel of leading decision researchers, who were invited to review the evidence and report to a congressional subcommittee. Two members of the panel specifically attributed the mistake to faulty decision making. One described how the mistake seemed to be a clear case of expectancy bias, in which a person sees what he is expecting to see, even when it departs from the actual stimulus. He cited a study by Bruner and Postman (1949) in which subjects were shown brief flashes of playing cards and asked to identify each. When cards such as the Jack of Diamonds were printed in black, subjects would still identify it as the Jack of Diamonds without noticing the distortion. The researcher concluded that the mistake about altitude seemed to match these data; subjects cannot be trusted to make accurate identifications because their expectancies get in the way.
I have talked with this decision researcher, who explained how the whole Vincennes incident showed a Combat Information Center riddled with decision biases. That is not how I understand the incident. My reading of the Fogarty report shows a team of men struggling with an unexpected battle, trying to guess whether an F-14 is coming over to blow them out of the water, waiting until the very last moment for fear of making a mistake, hoping the pilot will heed the radio warnings, accepting the risk to their lives in order to buy some more time.
To consider this alleged expectancy bias more carefully, imagine what would have happened if the Vincennes had not fired and in fact had been attacked by an F-14. The Fogarty report stated that in the Persian Gulf, from June 2, 1988, to July 2, 1988, the U.S. Middle East Forces had issued 150 challenges to aircraft. Of these, it was determined that 83 percent were issued to Iranian military aircraft and only 1.3 percent to aircraft that turned out to be commercial. So we can infer that if a challenge is issued in the gulf, the odds are that the airplane is Iranian military. If we continue with our scenario, that the Vincennes had not fired and had been attacked by an F-14, the decision researchers would have still claimed that it was a dear case of bias, except this time the bias would have been to ignore the base rates, to ignore the expectancies. No one can win. If you act on expectancies and you are wrong, you are guilty of expectancy bias. If you ignore expectancies and are wrong, you are guilty of ignoring base rates and expectancies. This means that the decision bias approach explains too much (Klein, 1989). If an appeal to decision bias can explain everything after the fact, no matter what has happened, then there is no credible explanation.
In order to reject this scenario fulfillment or expectancy bias explanation, we need to come up with a better one. The Fogarty report pointed out an obvious culprit: the computer screens that the crew members worked with were not easy to read. A very large screen display showed the big picture, but this display did not show altitude. Instead, altitude was listed on a small alphanumeric display off to the side of the primary display. The altitude of a track was given as a four-digit number, so 13,000 feet would appear as 1300. This number was embedded in a list of other numbers showing range, speed, bearing, and so on. This small screen was hard to read, especially if a crew member had to abandon what he was watching on the large screen to search through the small one. The more critical problem noted by the Fogarty report was that the trend was not given. Crew members could not easily check whether an airplane was ascending or descending. To figure that out, the crew members had to study the changes in the four-digit altitude readout. I asked Captain Rogers how long it might take to infer altitude. He said perhaps five to ten seconds. That may not sound like a long time, but consider these sailors in a noisy room, wearing headphones with one message coming in over one ear and another over the second ear, and a third set of messages coming in over the loudspeakers in the room, and perhaps a fourth from someone else; it may be hard to remember what the original numbers were in trying to keep track of the trend. The key part of the decision here took only 189 seconds, so 10 of those seconds can seem long. The Fogarty report recommended that trend data for altitude be shown on the large screen displays, a sensible idea. The Vincennes’s crew may just have gotten confused by the welter of numbers. Yet this still leaves the problem of why so many crew members insisted that they saw the aircraft descending.
The answer may be that there was a second airplane that no one knew about, causing the confusion. I heard about the second airplane from Captain Rogers, who told me that something odd had occurred at the beginning of the incident. Recall that the Vincennes had given Flight 655 the track number 4474, while the Sides had identified the same airplane at the same time and used track number 4131. The computerized system figured out that 4474 and 4131 were the same airplane and needed to have a single track number. The computer selected the Sides’s number for Flight 655, track number 4131. Rather than risk running out of track numbers, the computer system put track number 4474 in storage, to be recycled. A few minutes later it recycled the number, assigning it to a U.S. Navy A-6 several hundred miles away. As luck would have it, during the time period 10:20 to 10:24, that A-6 was flying a course that was descending and increasing in speed. At least that was what Rogers thought, but he did not want to attach too much credibility to it.
In the Fogarty report, the account of the Vincennes during the incident sounds like bedlam, with everyone having a different idea of what the track was doing.
Roberts and Dotterway (1995) reanalyzed the evidence and concluded that there were just two sets of impressions in the Vincennes. One set of crew members reported an aircraft steadily increasing in altitude. Their reports matched well with the Airbus’s pattern. The other group reported an aircraft steadily descending, which matched well with the A-6 pattern. In other words, due to a system weakness, the Vincennes’s crew may have been seeing two different airplanes. The fact that the track number had been changed had not been clearly announced at the time it occurred. To find out the altitude of a track, you can use a trackball to hook it on the screen. If the ship is pitching violently, as the Vincennes was, it might be safer to punch in the track number into a key pad. Apparently many of the crew chose to do that, and they punched in the wrong number. When Captain Rogers asked, “What is 4474 doing?” they proceeded to find out. It was descending. So we do not need to resort to scenario fulfillment explanations or expectancy biases. The crew members used a four-digit track number that turned out to be wrong.
Why did the Airbus fail to respond to the radio warnings? One guess is that on such a short flight, the pilots kept one ear of the headphone on the air traffic controller of the Bandar Abbas airport just departed and the other on the destination airport, Dubai. This strategy would leave the pilots without any ears free to be tuned to the international air distress frequency.
The Vincennes shootdown has some similarity to example 5.2, the Libyan airliner, another commercial airplane shot down because the pilot behaved in ways that were inexplicable to the people monitoring its flight. Both instances seem to follow the same pattern: the use of mental simulation to evaluate and rule out possible explanations.
