Human Movement Science journal homepage: www.elsevier.com/locate/humov
Performing under pressure: Gaze control, decision making and shooting performance of elite and rookie police officers
Joan N. Vickers, University of Calgary, Canada
William Lewinski, Minnesota State University (Mankato), MN, USA
Gaze of elite (E) and rookie (R) officers were analyzed as they faced a potentially lethal encounter that required use of a handgun, or inhi- bition of the shot when a cell phone was drawn. The E shot more accurately than the R (E 74.60%; R 53.80%) and made fewer decisions errors in the cell condition when 18.50% of E and 61.50% of R fired at the assailant. E and R did not differ in duration of the draw/aim/fire phases, but the R’s motor onsets were later, during the final second compared to the E’s final 2.5 s. Across the final six fixations the E increased the percent of fixations on the assailant’s weapon/cell to 71% and to 86% on hits, compared to a high of 34% for the R. Before firing, the R made a rapid saccade to their own weapon on 84% of tri- als leading to a failure to fixate the assailant on 50% of trials as they fired. Compared to the R, the E had a longer quiet eye duration on the assailant’s weapon/cell prior to firing. The results provide new insights into officer weapon focus, firearms training and the role of optimal gaze control when under extreme pressure.
The ability to perform under extreme pressure is a quality sought by all, whether in the military, law enforcement, emergency medicine or sport. Why are some able to maintain their focus and make the right decision under the most trying of circumstances, while others fail in a similar situation? The human’s ability to deal with life and death situations is not well understood in science, due to the inherent difficulties associated with simulating extreme events, of testing individuals at the poles of expertise, and of harnessing technology that can provide relevant information in compressed and physically challenging time frames. Yet the prevalence of force on force encounters is growing on our streets as more and more officers are faced with serious crimes involving firearms (Barclay, Tavares, Kenny, Siddique, & Wilby, 2003). Officer involved shootings lead to a costly process where lives are lost, careers ruined and billions of taxpayer dollars spent on investigation and litigation costs (Dumke, 2009; Klinger, 2006). Highly trained special weapons attack officers (SWAT) handle a greater number of serious incidents than non-SWAT officers yet they make fewer errors (Williams & Westhall, 2003). We therefore determined the gaze, decision making and shooting performance of highly trained elite officers who had extensive experience with violent encounters and rookies who were nearing the completion of their training program in an in situ encounter where an ‘‘assailant’’ suddenly either drew a weapon from under his coat, spun rapidly and fired, or pulled out a cell phone instead.
Control of the gaze plays a major role in models of visuo-motor control (Corbetta, Patel, & Shulman, 2008; Land, 2009). Corbetta et al. (2008) explain that the gaze is controlled by a neural network that includes a ‘‘dominant ventral frontoparietal network that interrupts and resets ongoing activity and a dorsal frontoparietal network specialized for selecting and linking stimuli and responses’’ (p. 306). Feeding spatial information into both networks is the gaze system which directs attention to important objects or events within a scene in real time and in the service of ongoing perceptual, cognitive, and behavioral activity (Henderson, 2003). Currently most firearms training programs teach officers to focus their gaze on two locations, first on the sights of their gun, and secondly on the target before pulling the trigger (Hendrick, Paradis, & Hornick, 2008; Morrison & Vila, 1998). This gaze strategy works very well in training with rookies achieving high accuracy scores before graduation but once on the street and faced with a violent firearms encounter they shoot poorly, averaging between 10 and 60% accuracy (Morrison & Vila, 1998; Oudejans, 2008).
Eye movements studies show that the type of gaze control rookies are taught in firearms training differs from that used by elite athletes who perform in the pistol, rifle and shotgun sporting events. Ripoll, Papin, Guezennec, Verdy, and Philip (1985) recorded the gaze of elite and near-elite Olympic pistol shooters as they fired at a fixed target. The near-elite shooters first fixated the sights on the pistol and then aligned the sights of their gun to the target resulting in a final fixation that was shorter in duration than that of the elite shooters who fixated the target first and never let their gaze deviate from the target as they raised their pistol and aligned the sights relative to one stable line of gaze. This resulted in a longer final fixation duration and higher accuracy.
Similar results were found by Vickers and Williams (2007) who tested Olympic biathlon rifle shooters under low and high levels of pressure and physiological arousal. The athletes took standing shots at a target after exercising on a bike ergometer at individually prescribed power output (PO) from 55% to 100% of their maximum power output. Performance pressure was manipulated by testing them in a low pressure condition where they were told that the purpose of the testing was to give them information about their fixation on the target, while in the high-pressure condition they were told their shooting percentages would be used in Olympic team selections. Although the athletes did not differ in levels of anxiety, heart rate and rate of perceived exertion when under high pressure, those who choked at the highest workload (accuracy <30%) reduced the duration of their final fixation, or quiet eye, on the target while those who did not choke (accuracy >80%) increased their quiet eye duration above any level found during the low or high pressure. In effect, the use of a long duration quiet eye seemed to reduce the normally debilitating effects of high anxiety, pressure and physiological stress. Similar quiet eye findings have been found for elite athletes performing in long distance rifle shooting (Janelle et al., 2000) and the three shotgun events of skeet, trap, and double trap (Causer, Bennett, Holmes, Janelle, & Williams, 2010).
The QE was defined by Vickers (1996, 2007) as the final fixation or tracking gaze that is located on a specific location or object in the performance space within three degrees of visual angle for a minimum of 100 ms prior to the onset of a critical movement. The quiet eye has been shown to underlie higher levels of skill and/or performance in a wide range of skills, including golf (Vickers, 1992, 2004, 2007; Vine, Moore, & Wilson, 2011); basketball (de Oliveira, Oudejans, & Beek, 2008; Harle & Vickers, 2001; Vickers, 1996; Vine & Wilson, 2011); rifle and shot gun shooting (Causer et al., 2010; Janelle et al., 2000; Vickers & Williams, 2007); billiards (Williams, Singer, & Frehlich, 2002) and ice hockey goaltending (Panchuk & Vickers, 2006, 2009). Participants who have been tested in high pressure situations have a higher frequency of gaze, more fixations of shorter duration (Behan & Wilson, 2008; Janelle, 2002; Williams, Vickers, & Rodrigues, 2002; Wilson, Vine, & Wood, 2009), as well as a reduced ability to detect information in the periphery (Janelle, Singer, & Williams, 1999). Janelle et al. (2000) also found progressive quieting of the left hemisphere of elite marksmen before the trigger pull and quiet eye durations that were significantly longer for the elite shooters than for non-elite. In a recent meta-analysis Mann, Williams, and Ward (2007) found that a longer quiet eye duration was one of three predictors of perceptual-motor expertise, along with specific fixation locations and a low frequency of fixations. In every motor task there appears to be a crucial moment when the individual must fixate or track specific task information and this must occur before a critical final movement is made.
Most QE studies have been carried out in sport tasks where the participants were familiar with the conditions of performance. In the current study the officers were not aware of the scenario, but instead were faced with an unknown ‘‘assailant’’ in a strange room and events they could not predict in advance. As such, the situation was similar to that found during the commission of crime, where the intentions of the perpetrator are most often unknown until the last few seconds. A number of studies have shown that the gaze of eyewitnesses to a crime is drawn to only one location – the weapon being used by the perpetrator (Hendrick et al., 2008; Hope & Wright, 2007; Hulse & Memon, 2006; Kassin, Tubb, Hosch, & Memon, 2001; Pickel, 1999; Stanny & Johnson, 2000; Steblay, 1992). Weapon focus refers to ‘‘the concentration of a crime witness’s attention on a weapon, and the resultant reduction in ability to remember other details of the crime’’ (Loftus, Loftus, & Messo, 1987, p. 55). While weapon focus is considered a robust finding, applicable to both lay-persons and police officers, some reservations have been expressed due to there being no empirical support during an actual live encounter (for example, Turtle, Read, Lindsay, & Brimacomb, 2008). The gaze of officers has also not been determined during a situation where they must fulfill the dual role of a witness to a crime and that of an officer who must draw, aim and fire accurately under extreme pressure, or alternatively inhibit its use when a benign object such as a cell phone is pulled from a pocket in a manner similar to when a gun is drawn.
Weapon focus studies have largely been carried out using video simulations, yet there is growing evidence that the gaze differs in simulated environments from that found in real world situations where performers are able to physically perform the motor skills being investigated. Müller and Abernethy (2006) and Mann, Abernethy, and Farrow (2010) have shown that cricket batsmen step earlier and track the ball longer against a live bowler than when addressing video simulations of pitches. Dicks, Button, and Davids (2010) found similar results in a soccer goaltending penalty kick study which used five experimental conditions – a video simulation of kicks with verbal and joystick responses, and on the field responses against a kicker using verbal, step and normal interceptive responses. Not only did the goalkeepers make more saves in the in situ condition, but they focused on fewer locations and had earlier reaction times and faster movement times. Van der Kamp, Rivas, van Doorn, and Savelsbergh (2008), Mann et al. (2010) and Dicks et al. (2010) have commented that when simulators are used, critical elements are often removed leading to activation of only the ventral system, while in the in situ setting both ventral and dorsal processing occurs. The ventral system is the slower of the two systems and facilitates the reorienting of attention and cognitive processing, whereas the dorsal system is designed to control fast actions that are controlled automatically (Corbetta et al., 2008; Milner & Goodale, 1995, 2008). Elite performers use both systems, switching back and forth as needed, whereas novices may rely too much on one or the other depending on how they are assessed. When simulators are used the dorsal and ventral systems may become de-coupled as there is no real consequences or immediacy built into the task, while during in situ situations the specificity of action is maintained and perception–action coupling occurs which more accurately reflects the true nature of the visuo-motor system as it has been trained to function. Since video simulators are used extensively in police training, there is an added importance of a study where the gaze and motor responses of officers are assessed under conditions that are very similar to those encountered in the field.
In summary, it is not known if officers involved in a lethal firearms encounter control their gaze as taught in training, or as found for elite athletes in the shooting sports, or as described by the weapon focus literature. We therefore determined the gaze and shooting performance of elite and rookie officers during a gun condition where shots were always fired, and a cell condition where all aspects of the scenario were the same but a cell phone was drawn instead of a gun. Shooting performance was assessed using shot accuracy (%), location of shots (%) and shot speed (ms). Decision making was determined during the cell condition using the percentage of officers who inhibited both shots. Overall performance (low, high) took into account combined measures of shot accuracy, shot speed and shots inhibited. The officers motor phase durations (ms) were determined for the first seven seconds (prepare, unholster) and last seven seconds (assess, draw, hold, aim/fire), while the assailant’s phase durations (ms) were determined during the final seven seconds: (confront, pivot, aim/fire). Fixation variables were fixation location (%), fixation duration (ms), QE duration (ms) and final saccade location (%).
During the initial part of the encounter, we expected both the elite and rookie officers to em- ploy a weapon focus fixating locations on the assailant where a weapon could be hidden. If the weapon’s focus literature holds true the officers should fixate the weapon being used, or the cell phone, with other areas being secondary. As the assailant pivoted, aimed and fired (or appeared to) we therefore expected the elite officers to maintain a longer duration QE on the assailant’s weapon or cell prior to firing or inhibiting the shot, while the rookies QE period would be briefer and to locations that were more varied. Consequently, we expected the elite officers to shoot with greater accuracy during the gun condition and make fewer decision errors during the cell condition, while the rookies would exhibit less control over their gaze leading to lower shooting accuracy and more decision errors.
24 officers volunteered for the study, 11 elite (E) male members of an Emergency Response Team and 13 rookies (R) from the same department, 6 males and 7 females. The E officers had extensive field experience dealing with firearms incidents through-out their career, while the R were at the end of their training program. The E were significantly older, F(1, 22) = 7.11, p < .01, g2p = .24 (M = 38.82 yr- s ± 4.60 yrs) than the R (M = 30.54 yrs ± 6.55 yrs). The shooting eye of the officers was predominately right (E 9/11; R 9/13), as was their shooting hand (E 10/11; R 13/13). Shooting performance results were available for all 24 officers, but coupled gaze and motor data were available for 18 officers (elite, male n = 8; rookie, n = 10, 6 males, 4 females). Gaze data were not available for six officers due to squinting as they fired, or tilting the head laterally causing loss of scene by the eye tracker. Testing occurred within a police training academy over a period of two weeks under the supervision of trained safety personnel. All officers gave their informed consent prior to participating and ethics approval was received prior to testing.
2.2. Materials and procedures
The vision-in-action (VIA) system (Vickers, 1996, 2007) was used to record the coupled gaze and motor behaviors of the officers. The Mobile Eye is a light (76 g) monocular eye-tracking system that uses corneal reflection to measure eye-line-of-gaze with respect to the field of view (accuracy of ±1° visual angle and precision of 0.5°). Three frames of VIA data (A–C) are shown in Fig. 1, as recorded during the final two seconds. Image 1 of each frame (A–C) was recorded by the cameras on the eye tracker worn by the officer. The small circle shows the location of the officer’s gaze and the larger circle indicates nor- mal pupil recognition. Image 2 was recorded by an external camera that simultaneously captured the officer’s shooting movements. Audio of the assailant and officer’s shots was determined using a central microphone connected to a Shure SCM 268 mixer. Precise synchronization of images 1 and 2 and the audio output occurred post-data collection using Final Cut Pro (Apple Corporation).