Gaze Control and Tactical Decision-Making Under Stress in Active-Duty Police Officers During a Live Use-of-Force Response

Abstract

Police officers during dynamic and stressful encounters are required to make rapid decisions that rely on effective decision-making, experience, and intuition. Tactical decision-making is influenced by the officer’s capability to recognize critical visual information and estimation of threat. The purpose of the current study is to investigate how visual search patterns using cluster analysis and factors that differentiate expertise (e.g., years of service, tactical training, related experiences) influence tactical decision-making in active-duty police officers (44 active-duty police officers) during high stress, high threat, realistic use of force scenario following a car accident and to examine the relationships between visual search patterns and physiological response (heart rate). A cluster analysis of visual search variables (fixation duration, fixation location difference score, and number of fixations) produced an Efficient Scan and an Inefficient Scan group. Specifically, the Efficient Scan group demonstrated longer total fixation duration and differences in area of interests (AOI) fixation duration compared to the Inefficient Scan group. Despite both groups exhibiting a rise in physiological stress response (HR) throughout the high-stress scenario, the Efficient Scan group had a history of tactical training, improved return fire performance, had higher sleep time total, and demonstrated increased processing efficiency and effective attentional control, due to having a background of increased tactical training.

Keywords:

• police
• use of force
• training
• use of force
• visual control

Tactical decision-making is the process of reassessing the environment and actors to control the situation (Boulton & Cole, 2016) and is an essential element for effective perception that is influenced by an officer’s capability to recognize critical visual information, evaluate threat, and regulate emotions (Fallon et al., 2014). During dynamic and stressful encounters, officers are required to make rapid decisions that rely on effective decision-making processes, experience, tacit knowledge, and intuition (Penney et al., 2022). These encounters can be potentially uncontrollable, novel, and often involve time pressure that can lead to dire consequences, including injury or death (Andersen et al., 2016; Cohen et al., 1998). Combined, these factors influence an officer’s propensity to formulate efficient and proper tactical decisions, prepare, initiate, and execute a response behavior.

Emotional experiences, such as those experienced by officers in dynamic, dangerous encounters will often cycle through a process that includes exposure to emotional stimuli, attention to the stimulus, appraisal, and finally a behavioral and physiological response (Beatty & Janelle, 2020). Exposure to these emotional stimuli leads to the activation of brain regions that are associated with affective, perceptual, cognitive, and motor processes. Further, the emotional response can be dominated by highly arousing stimuli which can result in an increase in sympathetic nervous system (SNS) response, including elevated skin conductance, increased pupil dilation, elevated heart rate, and modified visual function.

Considerable work has demonstrated the impact of emotional regulation and heightened SNS in sport (e.g., Korobeynikov et al., 2016; Sessa et al., 2018). However, only recently have researchers demonstrated the impact of heightened SNS response on police performance (e.g., Andersen & Gustafsberg, 2016; Baldwin et al., 2021). In Baldwin et al., 122 active-duty police officers performed a realistic scenario that included lethal force to induce stress. The investigators found that the scenario caused elevated heart rates and visual changes including reporting tunnel vision. The average performance rating (which reflected a composite score from 3 measures: Deadly Force Judgment and Decision-Making, Tactical Social Interaction, and Crisis Intervention Team surveys) for the officers was 59%, with 27% of the participants committing at least one lethal force mistake, with more mistakes occurring when stress was higher. No effect was found when comparing the level of standard police training with cardiovascular stress reactivity or other symptoms caused by stress. This may indicate that current training techniques do not improve an officer’s ability to operate under stress. Similar results were found by Baldwin et al. (2019) who collected autonomic responses and GPS positioning of 64 police officers during duty and were able to correlate an officer’s heart rate to specific times/stressors. The researchers found that the physiological stress of the officers is swayed specifically by the phase of the call and incident factors and not officers’ years of experience.

An officer’s heart rate is often elevated due to stress during dynamic encounters; however, this may not have a direct impact on performance as both high and low performance outcomes are associated with elevated cardiac levels. Nieuwenhuys and Oudejans (2011) compared the shooting behaviors of officers under either high anxiety (targets shoot back) or low anxiety conditions (targets do not shoot back). Each participant took a pre- and post-test, with four 1-h training sessions in between (under high or low anxiety conditions), and a retention test 4 months post-training. Although the two groups had similar shot accuracy for the pretest, the post-test results showed the high anxiety group had better accuracy than the low anxiety group, and this continued in retention. These findings indicate that specific high-anxiety training can improve performance in police officers in both the short- and long-term and can improve performance outcomes over standard police training.

The Integrative Model of Stress, Attention, and Human Performance (IMSAHP) accounts for the effects of stress on performance. It also explains the influence of emotional modifications on attentional control, processing efficiency, and how emotions influence subsequent visuomotor performance. According to IMSAHP, when a stimulus is perceived as a threat, the stimulus-driven attentional system disproportionally taxes cognitive resources, increases physiological responses, reduces performance, modifies gaze characteristics, and results in heightened distractibility from relevant cues (Vine et al., 2016). These modified gaze characteristics result in increased fixation quantities, decreases in fixation durations, and increases in the number of saccades (e.g., Bradley et al., 2011; Przybyło et al., 2019). Increases in emotional and task demands can also disrupt cognitive processing efficiency and result in narrowed attention, increases in saccades, and heightened distractibility during which there is a delay in decision-making and motor responses (Murray & Janelle, 2003). For example, firearms training programs lead to relatively high accuracy during qualifying tests (above 90%), but low accuracy during highly emotional in-field encounters (35% accuracy; Donner & Popovich, 2019). Performance effectiveness is the quality of the performance, whereas processing efficiency is the relationship of performance effectiveness and the amount of cognitive effort invested in the task (Murray & Janelle, 2003). Generally, highly emotional, and arousing states result in a decrease in performance effectiveness and increase in processing efficiency demands (Murray & Janelle, 2007). The consequence is heightened distractibility and focus away from relevant cues for performance success.

In contrast, when stressful tasks or environments are perceived as a challenge or as an opportunity to demonstrate competence and achieve success, performance effectiveness and processing efficiency improves. Considerable evidence demonstrates that effective visuomotor strategies through relevant training optimizes emotional regulation and allocation of attention to performance-relevant stimuli (e.g., Vine & Wilson, 2011, Ferri et al., 2016). Researchers have investigated optimal visuomotor strategies utilized by experts in both self-paced (e.g., Janelle et al., 2000; Vine et al., 2011; Ziv & Lidor, 2015), and externally paced tasks (e.g., Hunfalvay & Murray, 2018). These studies consistently conclude that skilled behaviors are marked by specific visuomotor strategies, one of which is quiet eye duration (Helsen et al., 2000; Rodrigues et al., 2002; Vickers et al., 2019; Vincze et al., 2022). Quiet eye represents the final fixation or tracking gaze at a task-relevant location prior to the initiation of the final phase of the movement (Vickers et al., 2019). Causer et al. (2010) found that expert shotgun shooters enter a quiet eye state earlier and remain in it for longer durations compared to sub-elite shooters. It was also found that when experts missed their shots, their quiet eye onset was later and duration was shorter than when they hit the target.

Extensive research has been conducted focusing on the differences between expert and novices within a multitude of performance environments (e.g. de Oliveira et al., 2008; Nagano et al., 2004; Rienhoff et al., 2012; Vickers & Williams, 2007). Multiple studies have specifically looked at the phenomenon of visual search within the realm of performance settings. However, most of these studies have been conducted in laboratory environments (e.g., Moore et al., 2012; Vine et al., 2011), while only few have been conducted in real world or outside lab simulated settings (Vine & Wilson, 2010; Wilson & Pearcy, 2009).

Evidence suggests that the context of testing, training, and expert performance in policing matters. In the absence of stressors, differences between those with and without training may be minimized. For example, when tested for accuracy using a Grey Man target with only a time duress, Lewinski et al. (2015) found minimal differences between police recruits, who passed their firearms qualifications and were about to finish their formal police training, and novices who had never fired a handgun in their life. In contrast, Vickers and Lewinski (2012) found performance differences between elite officers with extensive experience and training in firearms incidents, and rookie officers as they faced a potentially lethal encounter. The officers had to decide if the person they were observing was a threat while the person was pulling an item out of their waist band. If the item was a weapon, the person was deemed a threat and the officer had to deploy their handgun. Alternatively, if the person pulled out a cellphone, the officer would not deploy their handgun. Overall, the elite officers shot more accurately (74%) than the rookie officers (53.8%) and made very few decision errors in the cell phone condition. The elite officers had a longer quiet eye duration on the cue (i.e., assailant’s weapon/cellphone) or where the cue was going to appear prior to firing and an increased number of fixations on the assailant’s weapon and preattack actions with their limbs and hands. Vickers and Lewinski’s work highlighted the need for more sophisticated firearms training, game knowledge, and practice in the role of optimal gaze control when under extreme pressure.

Tactical decision-making is also influenced by the officer’s capability to recognize critical visual information and estimation of threat. When an individual is in a stressful situation there is evidence that the stress leads to behavioral errors and high-order cognitive processes. A recent review by Anderson et al. (2019) supports this notion in policing and further indicates there is reduced drop in performance under stress even when an officer has relevant training. Based on these recent studies showing the effects of the test and training environment, it is clear that researchers must examine police officer performance in situations that mimic the stress induced in realistic, threatening situations.

It is presently unknown what defines visual expertise in officers (e.g., years of service, tactical training, related experiences) when they encounter a threat or potential threat. The purpose of the current study is to investigate how visual search patterns using cluster analysis and factors that differentiate expertise (e.g., years of service, tactical training, related experiences) influence tactical decision making in active-duty police officers (44 active-duty police officers) during a high stress, high threat, realistic use of force scenario. The role of physiological arousal on visual search during a high stress, high threat, realistic use of force scenario is also unclear. Rather than use an arbitrary definition of expertise, the purpose of this research was to differentiate groups via visual search patterns using cluster analysis, to examine scenario-related gaze characteristics including fixation location, fixation duration (ms), and saccades, and to determine factors that differentiate these groups in a high stress, high threat, realistic use of force scenario. Consistent with previous studies (Lebeau et al., 2016; Mann et al., 2007), highly skilled officers defined by cluster analysis will demonstrate fewer fixations of longer duration on higher threat locations than lower skilled officers who are expected to exhibit more fixations for shorter duration and more fixations on non-threat locations. That is highly skilled officers will exhibit more efficient visual search patterns (longer durations and fewer fixations) which will be associated with better scenario performance.

Methods

Participants

Forty-four active-duty police officers (M age = 32.86 ± 7.2 years) with experience ranging between 7 wk − 23 years (Mean =7.05 ± 6.16 years) participated in the study.

Procedure

Upon arrival, participants completed the informed consent forms and were explained the purpose of the study (Figure 1). Prior to the commencement of the scenario, all participants completed a safety check of all their duty equipment. All lethal weapons were removed from the participants and secured at the location. Officers received a neon yellow Velcro tag to wear on their uniform indicating visually they have been properly safety-checked. Only the two officers who were the test participants were evaluated within each trial; the remaining police trainers were confederates who were actors in the scenario. Prior to starting the scenario, each officer was outfitted with an eye tracker, heart rate monitor, and their vest with simulated weapon and simulated taser. The scenario ran for approximately 15 min. Each scenario included two police officer participants, and four additional pre-scripted police trainers (e.g., actors). Of the four actors, one acted in the capacity as an off-duty patrol officer who was involved in a crash, one was a passenger in the assailant’s car, one was the driver (assailant) who caused the crash and later attacked the officers, and one was another motorist who stopped to assist the passenger in the assailant’s car. The study design enabled us to create nine time points of critical events as we increased the police officers’ stress response within this scenario. These nine points are identified below within the description of the scenario.

FIGURE 1. Methodological infographic: data collection and data analysis.

This scenario revolved around a citizen/assailant’s traffic crash into a police car. The assailant’s car was occupied by the actor who played the role of the assailant and another occupant (or passenger) actor. The study participants (two officers) on cue responded at the request of the on-scene officer asking for backup. The study participants entered their squad car (Time Point 1) and were given the instruction of ‘green’ (Time Point 2) to start the scenario. Following the officers completed a loop driving course while receiving ‘Code 2’ (non-emergency) response from a dispatch noting the confrontation (Time point 3). Shortly after the Code 2, the study participants received a ‘Code 3’(emergency) response from dispatch (Time Point 4) and then were required to complete a slalom course that simulated fast driving and sharp turns while being informed that the suspect was becoming agitated, verbally threatening to the on-scene off-duty officer and had a warrant out for his arrest. The officers pass a victim (Time Point 5) and then enter the staged part of the scenario by driving around a corner (Time Point 6). As the study participants arrived, they responded as their training and police experience dictated, with the goal to defuse the agitated driver and resolve the conflict at the crash scene. Once the officers exited their car (Time Point 7), the witness role-player became involved by attempting to communicate with the primary responding participant officer and the hostile driver. With both officers (study participants) involved in de-escalation attempts toward the hostile driver, he then became increasingly agitated and eventually acquired and discharged a firearm at the officers (Time Point 8). The study participants were given an opportunity to react as their training dictated and respond appropriately by returning fire or taking cover. Following completion of the scenario (Time Point 9), the study participants completed the questionnaire (survey).

Measures

Eye movements were tracked using two infrared eye trackers: Tobii 3 Eye Tracking Glasses (50 hz; Stockholm, Sweden), and Pupil-Labs Invisible eye tracking glasses (200 hz; Berlin, Germany). The Tobii 3 has an accuracy of .6 degrees of visual angle and the Pupil Invisible accuracy is .5 degrees of visual angle. ECG physiological data were acquired using a Biopac MP 150 (Goleta, CA, USA). Participants had their skin shaved (when needed), abraded, and cleaned of oils and sweat before three Ag/AgCl ECG electrodes were placed in an inverted triangle over the right side of their chest. The electrodes were secured in place with tape. Video was obtained through the eye tracking glasses giving a point of view (P.O.V.) of the officers, the body cameras on the officers, which was a mounted Hero8 GoPro (San Mateo, CA., USA) on the officer’s chest, and through many high-resolution scene cameras that were positioned to record places of interest. The videos were examined for behavioral outcomes such as where the officers were visually attending, the type and speed of their responses to the changing dynamics in the scenario, and how quickly the situation was resolved, and then were used to compare these behaviors to the officers’ questionnaire responses as a manipulation check.

In addition, performance was indicated by the response to time to return fire and whether the officer returned fire. We deemed this as an appropriate measure as it related to their perception of the scenario and the appropriateness of their response. That is the ones who were more prepared to return fire were better at processing the threat and situation.

Survey/Questionnaire

The demographic survey which was completed after the scenario finished consisted of 30 questions that included personal variables (e.g., gender, age, height, weight, fitness, disease status, sleep quality), education level, and variables related to law enforcement including duration in law enforcement, current rank, current assignment, tactical training, military training, informal training, duration of current shift, etc. The questions related to law enforcement were open-ended to capture the breadth of possibilities from our participant pool.

Data Analysis

All eye tracking data was processed through iMotions Biometric Research Platform (v 9.3; iMotions (93), 2022). iMotions is multi-modal software suite that allows for comparison across eye trackers from two different manufacturers plus it allows for the collection of HR data within the same platform. Visual search variables included the number of fixations (defined as ≥100 ms), fixation durations, saccades, and search rate or the number of fixations in a defined period of time. Area of interests (AOIs) were established for four different areas: threat, potential threat, witnesses, and the suspect’s truck. The truck was included as it represented a distractor, limited visual information, and functioned as an irrelevant cue.

Hierarchical and nonhierarchical cluster analyses were conducted using a two-step process to improve stability in the cluster solution (Hair et al., 2010). Using standardized scores, the observed variables (fixation duration, fixation location difference score and number of fixations) were entered into the cluster analysis. The first stage involved a hierarchical cluster analysis using Ward’s linkage method with squared Euclidean distance measure to determine the number of clusters in the data. Ward’s method is an agglomerative clustering method based on sum-of-squares criterion and produces groups that minimize within-group dispersion (Hair et al., 2010). The second stage involved a k-means (nonhierarchical) cluster analysis by specifying the most appropriate cluster solution from stage 1.

After identifying the visual profiles, we performed a Group (2) × Time (9) repeated measures ANOVAs on the time points identified in the methods above and on heart rate data, and separate Group (2) × AOI (4: Threat, Potential Threat, Witness, and Truck) repeated measures ANOVAs using the dependent visual control variables (fixation duration, fixation location, and number of fixations). Significant main effects and interactions for the ANOVAs were followed-up with appropriate post hoc test. In addition, we examined the relationship between heart rate change (entering the car to arriving on the scene) to initial arrival on the search rate and saccadic activity. Saccadic variables included number of saccades, average saccade duration, average amplitude, and peak saccadic acceleration. Amplitude represents the distance traveled by a saccade during an eye movement and peak saccadic acceleration is the maximum velocity during the duration of a saccade (Brunyé et al., 2019).

Next, we conducted two multiple regression analyses with one to predict the total number of fixations that officers used and to assess the relative input of various visual targets to the overall search rate. The predictors were fixation durations on the threat, potential threat, witnesses, and truck, while the criterion variable was overall search rate. The second was to examine the total number of fixations that officers used and to assess the relative input of various visual targets to the time to duration in law enforcement. The predictors were fixation durations on the threat, potential threat, witnesses, and truck, while the criterion variable was duration in law enforcement. For all analyses, test of assumptions was completed and if violated then appropriate non-parametric analysis would be used. Significant multivariate effects (p < .05) were followed up with post hoc comparisons between cluster groupings variables using a t-test with Bonferroni adjustments as appropriate and chi-square test for proportional group data comparisons. Lastly, partial eta squared (η²_p) was used to determine effect sizes.

Results

Hierarchical Cluster Analysis

For the hierarchical cluster analysis, the agglomeration schedule coefficient and the dendrogram classified either two or three clusters as two possible solutions. A two-cluster solution (an Efficient Scan Group and an Inefficient Scan Group) was deemed the best fit according to empirical considerations (specific patterns of the observed variables) and how interpretable the cluster solution was. Next, a k-means cluster analysis was conducted on the standardized visual control variables for the two-cluster solution. The nonhierarchical solution provided support for the hierarchical analysis. To provide a descriptive indication of the strength of our cluster solution, we conducted a MANOVA on the multivariate effect of cluster membership. The MANOVA revealed a significant multivariate effect on cluster membership, Wilks’ Lambda = 0.486, F(2, 26) =6.69, p <.01, n_p² = 0.514, thus indicating reasonable support for our cluster solution. Clusters significantly differed on saccades (p < .001), fixation duration (p < .001), and average saccadic duration (p < .001) which further supports our two-cluster solution.

Profile of Each Cluster

Experience, wakefulness, and performance profile of each cluster (Table 1).

TABLE 1. Summarizes the experience and wakefulness profiles of the efficient and inefficient scan groups.

Factor	Efficient scan group	Inefficient scan group
Experience
Duration in law enforcement (months)	80.92 (69.56)	104.07 (77.72)
Tactical training^a	72%	53%
Military training^a	28%	10%
Lethal force experience	21.43%	20.70%
Detective	14.29%	24.14%
Officer	85.71%	72.41%
Officer in training	0.00%	3.45%
No college	21.43%	13.79%
</=2 Years of college	42.89%	51.72%
Undergraduate degree^a	21.43%	34.48%
Masters degree^a	14.29%	0.00%
Wakefulness (SD)
Time on shift before testing	3.90 (3.01) hours	4.42 (4.71) hours
Time awake before testing	7.63 (3.2) hours	8.79 (5.13) hours
Last sleep length^b	6.01 (1.27) hours	6.90 (1.58) hours
How rested they felt (1–5 scale; M & SD)	3.29 (0.87)	3.52 (1.27)

^a p<.05; ^bindicates not all participants answered this question.

In terms of performance, +85.5% of the Efficient scan group returned fire, whereas only 50% of the Inefficient Scan group returned fire (Figure 2). For those officers who returned fire, there was no significant difference in response time to fire (p = .787; Efficient scan group M = 1.18 sec, SD = 0.69; Inefficient Scan group M = 1.02 sec, SD = 0.75)

FIGURE 2. Percentage of officers who returned fire by group. Overall 85% of the efficient scan group returned fire whereas 50% of the Inefficient Scan group did.

Heart Rate Change

There was not a significant difference in heart rate change between the groups (p = .754) and there was not a significant interaction of Group × Time (p = .289). However, heart rate significantly increased over time as the officers approached and entered scenario F(8, 208) = 72.14; p < .001, η²_p = 0.742 (Figure 3). Tukey’s post hoc test revealed that is heart rate significantly increased at each point and from time point to time point from Code 2 to the end of the scenario (Time Points 4–9). The only time points not significantly different were Enter Car and Enter Track (Time Points 1 and 3). Furthermore, a modest relationship between Heart Rate Change and Search Rate (r = 0.43, p < .05) following Code 3 driving occurred.

FIGURE 3. Heart rate change across meaningful time points; heart range from the start point where the officers were kitted up, to the end of the scenario was 86–181 bpm (*p <.05).

Visual Response Data

Descriptive statistics for the areas of interest, vision variables including fixation locations and fixation duration for the two clusters were evaluated by separate repeated measure analysis of variance. The ANOVA for fixation location was significant, F(3, 21) = 6.168, p < .05, η²_p = 0.227. Tukey post hoc analysis for fixation location indicated significant differences between four locations: threat, potential threat, witness, and truck AOI locations. Overall, threat and potential threat had the highest number of fixations compared to witness and truck AOI locations. In addition, Group differences for total fixation duration and AOI fixation duration were significant (F(3, 126) = 10.439, p < .01, η _p ² = 0.199; F(3, 126) = 9.108, p < .01, η_p² = 0.178; respectively) and a significant interaction effect between cluster and AOI duration (F(3, 126) = 4.24, p = .009, η_p² = 0.168). A post-hoc analysis using a simple effects model revealed a higher fixation duration by the Efficient scan group on the threat, potential threat, and witness with lower duration for the truck (Figure 4).

FIGURE 4. Total fixation duration (SD) by area of interest by group (*p < .01).

In addition, we examined saccadic activity (i.e., visual search rates) including the number of saccades, average duration, average amplitude, and peak saccadic acceleration. All these measures indicated differences in visual search rates and scanning activity between the Efficient Scan Group and the Inefficient Scan Group. Specifically, the Efficient Scan Group had fewer saccades with longer saccade durations and higher amplitudes with lower peak accelerations (Table 2).

TABLE 2. Univariate test statistics for Saccadic activity variables.

Saccadic activity	F	Sig.	ηp^2	ESG Mean (SD)	ISG Mean (SD)
Number of saccades	12.91	<.001	0.259	47.85 (22.63	99.92 (51.24)
Average duration	8.023	<.01	0.178	90.5 ms (29.26	62.60 ms (29.63
Average amplitude	12.38	<.001	0.251	8.00° (2.47)	5.55° (1.84)
Peak acceleration	5.823	=.021	0.136	7469 (9354.93	15720.02 (10691.63)

ESG: efficient scan group; ISG: inefficient scan group.

Multiple regression was used to predict the total number of fixations that officers used and to assess the relative input of various visual targets to the overall search rate. Fixation durations on the threat, potential threat, witnesses, and truck were used as predictors. The model was significant, as indicated by F(5, 15) = 3.48, p = .044. Overall, 64% (R² = 0.635) of the variance in total number of fixations was explained by this model. The results revealed that fixation duration, t(−3.248) = −3.163, p < .01 was a significant predictor of the total number of fixations. The negative value of the slope weight indicates that as officers were more fixated on the threat, their overall search rate decreased (Table 3).

TABLE 3. Estimated results for model coefficients: Beta (B), Standard Error (S.E.), T-stat (T), and levels of significance in the regression model.

	B	S.E.	T	Sig.
Response time to weapon discharge (sec)	−1.416	2.124	−0.667	0.52
Fixation duration of threat	−13.782	4.243	−3.248	0.009
Fixation duration of potential threat	−4.399	5.393	−0.816	0.434
Fixation duration of witness	−24.508	12.728	−1.926	0.083
Fixation duration of truck	3.537	3.447	1.026	0.329

Multiple regression was used to predict the total number of fixations that officers used and to assess the relative input of various visual targets to the duration of law enforcement. Fixation durations on the threat, potential threat, witnesses, and truck were used as predictors. The model was not significant, R² = 0.091, F(9, 30) = 0.332, p = .957.

Discussion

In this study, we sought to investigate factors such as years of service, history of tactical training, related experiences, factors that influence behavior outcomes, and physiological arousal on visual search behavior, tactical decision making, and performance. To challenge their emotional control, cognitive response, and motor behavior, a complex task was developed involving all stages of a response to a traffic incident that included a both high stress and high threat in a realistic use of force scenario. In that scenario, officers were submitted to progressive complexity and danger, from the first instructions from dispatch to an escalating confrontation that resulted in the discharge of a firearm against the officers and other actors on site.

The Integrative Model of Stress, Attention, and Human Performance was used as a framework to assess the emotionally-driven changes to attentional control, information processing efficiency, and visuomotor performance (Vine et al., 2016). In our study, both groups experienced significant arousal when presented with the escalating scene, evidenced by increased heart rate. According to IMSAHP, it would be expected for both groups to experience reduced task performance, modified gaze characteristics, and reduced focus on relevant cues as physiological arousal occurred concomitantly with a disproportionate taxation of their cognitive resources. In other words, as expected in situations in which bottom-up sensory processing becomes dominant. However, the performance of the officers with relevant training had more efficient gaze behavior and resulted in better performance outcomes (e.g., being in a position to return fire). We found visual search variables in the more Efficient Scan Group to be representative of a more planned approach. This group had fewer saccades with longer saccade durations and higher amplitudes with lower peak accelerations. In addition, the Efficient Scan Group comprised of officers having a higher amount of tactical training who were able to exert better control of their visual searches and attentional focus, despite the stress-inducing environment, suggesting a relationship between relevant training, visuomotor performance, and processing efficiency. This can be seen by the fact that out of the Efficient Scan group 72% of the officers had tactical training (28% had military training), whereas in the Inefficient Scan group only 53% of the officers had tactical training. A qualitative assessment of the tactical training demonstrated that although varied the Efficient-scan group had more training relevant to skills needed in this use of force scenario. For example, the Efficient Scan group had combat deployments within the military deployment, advanced firearms training, street crime advanced training, firearms instructor, swat school, scenario training, and defense tactics instructor. These results indicate that officers with relevant training exhibited lower scan rates and a greater focus on the threat, potential threat, and the witness, respectively, while the less trained officers spent more time cycling through visual cues and focusing on the truck, an irrelevant cue. This supports the findings of Nieuwenhuys and Oudejans (2011) that found officers who trained in high-anxiety environments performed better on high-anxiety tests than officers who trained low-anxiety environments, regardless of police experience. The officers that did the high-anxiety training scored similarly on their post-test and retention test, inferring that the training has a lasting positive impact. It is important to note that in our study, having more real-world police experience did not correlate with having a lower scan rate, which aligns with past research (Baldwin et al., 2021; Nieuwenhuys & Oudejans, 2011).

Our expectation for this study was that better performances would occur with more efficient visual search patterns (fewer fixations and longer durations) and a robust history of tactical training. The data supports this expectation as efficient search patterns commonly resulted in better performance. This expectation was further supported by the fact that more efficient visual search pattern was a greater indicator of better performance compared to the officer’s duration of professional experience. The increased efficiency can be seen in Figure 4 where the Efficient Scan group spent most of their time looking at the threat, with the potential threat being looked at second most, and had a return fire rate of 85%. The Inefficient Scan group (lower efficiency) looked at the truck for the majority of the time, with the threat being looked at second most, and had a return fire rate of only 50%. The result of increased training bettering performance is supported by previous research (Clark et al., 2020; Vine et al., 2011).

An increase in performance due to efficient visual search patterns, regardless of experience, supports past findings of visual stability resulting in a regulation of emotions (Beatty & Janelle, 2020). Furthermore, substantial evidence demonstrates that appropriate attentional allocation can improve emotional regulation. Attentional focus on relevant cues can be an effective regulatory strategy. The Efficient Scan group, although demonstrating significant arousal response (and similar to the Inefficient Scan group) was better at managing visual motor control and regulating emotional responses. Deploying attention and selecting relevant visual cues attentionally may serve as a regulator for the impact of physiological response and mitigate emotional regulatory processes. For example, officers in the Efficient Scan group had considerably more controlled actions and were better apt to engage suspect. Fewer fixations for longer durations represents greater processing efficiency and aligns with previous research (e.g., Murray & Janelle, 2003, Murray & Janelle, 2007, Wilson, 2008). Bell (2004) proposed an inextricable link between cognitive and emotional processing, with both processes drawing from the same pool of resources. Furthermore, increased saccadic activity results in saccadic suppression or loss of information processing during an eye movement and can blur vision (Schweitzer & Rolfs, 2020). Within sport related literature highly skilled performers are better at acquiring perceptual cues which leads to improved response accuracy and lowered response time compared with novices (Lebeau et al., 2016; Mann et al., 2007). Experts are able to extract more task-relevant information and, as we noted here, tend to have fewer fixations, but of much longer duration (Kredel et al., 2017).

Other factors that may have impacted the findings are that the Efficient Scan group was on shift 31.2 min less on average before testing compared to the Inefficient Scan group. Also, the Efficient Scan group was awake for 1.16 h less than the Inefficient Scan group before testing on average. Previous research has found a link between more total time awake and worse sleep quality leading to worse decisions by police officers, however, the officers were affected more during non-shooting tasks than shooting tasks (Blake & Cumella, 2015). These findings allude to the importance some factors of overall officer quality of life inside and outside of work may play in their ability to respond properly in disproportionally stressful conditions. Precisely, sleeping schedules and sleep quality may have affected, to some degree, their ability to visually scan, process identifying information (i.e., relevant cues), and make timely decisions based on their assessment and training recollection.

Overall, expertise levels were associated with better performance in a high stress, realistic scenario involving an active shooter. However, tactical expertise was the main determinant of outcomes while the mere amount of time working as a police officer was not. Specifically, officers that had undergone a greater amount of tactical training were able to identify the assailant faster and maintain their focus on him for longer periods of time while the officers that had less previous tactical training spent more time switching their visual focus between cues on site and spent less time focusing on the visually compelling behavior of the assailant. As a result, experts were able to appropriately detect the evolving threat and return fire 85% of the time, while those with less tactical expertise only managed to detect the evolving threat and only responded appropriately with gunfire 50% of the time. It is possible that some of the performance decrease in participants with less tactical expertise was due to being awake, on duty for more hours, but this assumption requires further research to investigate it specifically.

Conclusion

Although both groups had a significant arousal response, additional tactical training in their career led to increased processing efficiency and effective attentional control. Selective attention represents a mechanism through which officers filtered the abundant, and sometimes complex sensory stimuli within their environments at any given moment. Due to the fact that attentional resources are limited, appropriate professional training is necessary to help officers capture salient stimuli across and within their different operating environments that can then direct their relevant and appropriate future choices and behaviors.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This project was supported by Force Science® LTD.