Emotional intelligence training improves stress regulation and performance in high-stress occupations

Abstract

In this randomized controlled study (N = 66) in an elite military setting, we compared stress regulation and performance outcomes between soldiers who received emotional intelligence (EI) training (n = 35) and those who received non-EI control training (n = 31). The EI-trained group underwent 15 h of instruction focused on recognizing, understanding, and regulating emotions. When exposed to intense simulated combat stressors, EI-trained participants showed significantly lower biological stress levels compared to controls and superior performance across multiple domains: shooting accuracy (94.1% vs. 51.6%, p < .001), memory retention for mission-critical details during stress serials (M = 5.15 vs. M = 3.10 items recalled, p < .001), increased speed and accuracy in complex mathematical calculations under pressure (56% vs. 19% correct, p < .001), and greater pain tolerance during cold water immersion (trainees persisted 72% longer, M = 50.94 vs. M = 29.08 min, p < .001). These findings highlight the potential of EI training to enhance stress regulation and cognitive and behavioral performance under pressure, offering valuable insights for improving employee well-being and avoiding burnout in high-stress environments.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36216-8.

Introduction

Work-related stress is a pervasive and growing concern worldwide, with far-reaching implications for individual well-being, organizational productivity, and societal health. The World Health Organization¹ estimates that workplace stress and related mental health conditions cost the global economy approximately US $ 1 trillion per year in lost productivity, identifying it as one of the most significant occupational health challenges of the 21 st century. The effects extend beyond individual health to influence broader economic and social systems. Recent findings by Gallup² reveal that 44% of employees globally report high levels of work-related stress, with the figure rising to 52% in the United States and Canada. Chronic stress contributes to numerous adverse outcomes, including burnout, anxiety, depression, cardiovascular diseases, and decreased job satisfaction³. For organizations, stress manifests as reduced productivity, increased absenteeism, higher turnover, and significant financial costs.

These challenges are particularly acute in high-risk occupations, such as the military, law enforcement, and emergency services, where employees routinely face life-threatening scenarios and high-stakes decision-making. Research has shown that individuals in these professions are disproportionately affected by stress-related conditions. For example, a study conducted by Beyond Blue⁴ on Australian police and first responders found that 21% reported high psychological distress compared to just 8% of the general population. Additionally, 39% of these workers had been diagnosed with a mental health condition at some point in their lives, nearly double the national average of 20%. Rates of trauma exposure are also significantly higher, with 51% reporting profoundly distressing experiences and 10% meeting the criteria for probable PTSD − well above the general population rate of 4%.

Despite these alarming statistics, many individuals in high-risk roles are reluctant to seek help due to stigma and fears about potential career consequences. Concerns about being perceived as weak or vulnerable further exacerbate the underreporting of stress-related challenges^5,6. The organizational consequences of unaddressed stress are substantial. In addition to the personal toll on employees, stress drives higher costs associated with recruitment and training, workers’ compensation claims, and health insurance premiums⁷. Addressing these challenges requires effective, evidence-based interventions to mitigate stress and enhance performance.

Emotional intelligence

Emotional intelligence (EI) training has emerged as a promising avenue for equipping individuals with skills to better manage stress. EI, as conceptualized by Mayer and Salovey⁸, refers to the ability to perceive, use, understand, and regulate emotions in oneself and others. This ability is particularly relevant in high-risk occupations, where employees must remain composed, focused, and effective under pressure.

Recent systematic reviews have demonstrated that emotionally intelligent individuals can buffer the effects of acute stress⁹. They exhibit superior emotion-regulation skills, enabling them to modulate their physiological and psychological responses to stressors. For example, they are more adept at reframing negative situations, reducing the intensity of emotional reactions, and employing adaptive coping strategies such as problem-solving and cognitive reappraisal^10,11. These emotion-regulation abilities are essential for maintaining composure and decision-making capacity in high-pressure environments.

In addition to enhancing individual stress regulation, EI contributes to stronger social support networks, which serve as buffers against workplace stress. Research demonstrates that high-EI individuals are better able to establish and maintain meaningful interpersonal relationships and to seek and offer support during times of stress^12,13. Social support is a well-documented protective factor, providing emotional, informational, and practical resources that mitigate the negative impact of stress.

EI also plays a critical role in improving cognitive and behavioral performance under pressure. High-EI individuals can maintain focus, prioritize tasks, and recover quickly from setbacks. For instance, Sanchez-Gomez and Breso¹⁴ found that EI was positively associated with memory retention and problem-solving abilities in stressful environments. Similarly, Fallon et al.¹⁵ reported that individuals with high EI demonstrated greater decision-making accuracy and flexibility in tactical scenarios. These findings underscore the potential of EI as a key factor in managing stress and enhancing performance in demanding occupational settings.

Defining and conceptualizing EI

Existing research offers three main models of EI, which Ashkanasy and Daus¹⁶ refer to as Streams 1, 2, and 3. Stream 1 models define EI as an ability to “carry out accurate reasoning about emotions and the ability to use emotions and emotional knowledge to enhance thought”¹⁷, p. 511] implying that such ability could increase with age and training. Specifically, such models define and measure EI using the Mayer and Salovey four-branch model¹⁸. Stream 2 models are also based on the Mayer and Salovey four-branch model but rely on self-report or peer-reports to assess EI^10,19. Stream 3 (or trait models) consider EI to be a form of emotion-related disposition, hierarchically located at the lower levels of personality²⁰. Trait EI, measured via self-report inventories (e.g., the TEIQue-SF²¹; the EQi^22), tend to measure typical behaviors in a range of situations rather than ability.

Ashkanasy and Daus¹⁶ argue that most criticism directed at EI inappropriately conflates the three models, emphasizing that most problems occur in the Stream 3 (trait-based) models. Although such EI definitions appear reasonable, intuitive, and are typically less costly and more efficient to administer²³, considerable ambiguity and redundancy remain in the way they are conceptualized²⁴. Ashkanasy and Daus¹⁶ argue that Stream 3 EI models do not provide sufficient incremental validity beyond traditional models of personality²⁵ and social/organizational behavior^26,27 and often share considerable conceptual overlap with personality attributes.

Researchers increasingly concur²⁸ that Mayer and Salovey’s¹⁸ ability model (a Stream 1 model) reflects a more conceptually sound construct of EI than mixed ability and personality-based models^13,16,29,30. A meta-analytic review by Hodzic et al.³¹ further supported this view, demonstrating that ability-based EI models show stronger validity and utility, particularly in predicting emotion-related outcomes beyond traditional personality traits. We therefore used this model to develop our EI training intervention, using a widely adopted ability-based measure of EI, namely the MSCEIT³².

Methodological shortcomings in EI training research

Despite growing evidence supporting the theoretical and practical benefits of EI, research on EI training interventions is still in its infancy and marked by significant methodological shortcomings. Many studies rely solely on self-report measures of EI and stress, which are prone to social desirability biases and subjective inaccuracies⁵. While these measures provide valuable insights into participants’ perceptions, they do not adequately capture the physiological dimensions of stress. Objective biological markers, such as salivary cortisol, are better suited to evaluating physiological stress responses. Cortisol is a widely used indicator of hypothalamic-pituitary-adrenal axis activation, which regulates the body’s stress response^33,34.

Moreover, existing research on EI training often lacks rigorous experimental designs. A meta-analytic review by Hodzic et al.³¹ found that most EI training studies employed non-experimental or quasi-experimental designs, limiting the ability to draw causal inferences. Few studies have used randomized controlled trials (RCTs), which are considered the gold standard for evaluating intervention efficacy. In an RCT, participants are randomly assigned to either receive an intervention or serve in a control group, allowing researchers to establish causal relationships while minimizing selection bias and controlling for confounding variables⁷. Additionally, many studies are conducted in laboratory settings or with student samples, raising concerns about ecological validity and applicability to real-world high-stress environments³⁵.

Another limitation is the short duration of most EI training programs, which may not allow sufficient time for participants to internalize and practice new emotion-regulation skills. Techniques such as resonant frequency breathing, commonly taught in EI training, require consistent practice over weeks to become effective³⁶. Longitudinal studies are therefore needed to assess the sustained impact of EI training on stress regulation and performance outcomes.

To address these shortcomings, we designed an RCT to evaluate the impact of ability-based EI training in a real-world occupational context. In our study, Australian Special Forces (ASF) soldiers were randomly assigned to either receive EI training (intervention group) or standard training (active control group). We measured outcomes using objective biomarkers (salivary cortisol) and performance indicators while participants underwent high-stress training exercises. We focused on this high-risk occupational group because they routinely face time-sensitive, critical stress events where decision-making can impact lives. The RCT design allowed us to establish causal relationships by comparing how EI-trained participants managed stress and performed on important cognitive and behavioral outcomes during challenging activities in a real work environment compared to those who received standard training, while also addressing limitations associated with prior self-report measures⁴.

Hypothesis development

Impact of EI training on stress

We offer three theoretical mechanisms through which EI training can help to improve stress management: (1) self-regulation, (2) problem-solving, and (3) social support.

Self-regulation

Individuals with high EI skills exhibit enhanced efficiency and accuracy in processing emotional information (such as facial expressions, tone of voice, body language, as well as interoceptive awareness) compared to their lower-EI counterparts, as evidenced by studies^37–38. EI-trained individuals, therefore, understand the importance of recognizing and responding to their own and others’ emotional cues³⁹. Joseph and Newman⁴⁰ suggest this acute sensitivity can act as a radar, alerting individuals to shifts such as anxiety or distress. This enhanced perceptual ability fosters a deeper comprehension of emotional fluctuations and cues when behavioral adaptations are warranted. Through structured exercises and feedback, individuals can develop their awareness of both internal emotional states and external environmental cues, enabling them to better gauge whether their emotional responses are proportionate to situations⁴¹. Research suggests that the ability to accurately assess one’s stress levels may help prevent burnout. For instance, Sanchez-Gomez and Breso¹⁴ found that individuals with higher EI showed lower burnout levels and better work performance, though the correlational nature of their study prevents conclusions about causation. Those who demonstrate more refined emotional perception tend to exhibit greater impulse control⁴⁰, which provides a foundation for effective self-regulation strategies⁴².

Problem-solving

Individuals with high emotional intelligence (EI) are more adept at problem-solving, largely because they not only perceive but actively gather relevant and emotionally pertinent information (which helps them grasp the nuances of social and emotional contexts) ³. This in-depth understanding is crucial for making well-informed decisions in situations that involve complex emotional dynamics. MacCann⁴³ adds that high-EI individuals employ a range of effective emotion-regulation strategies, including actively modifying their environment to a more emotionally advantageous one and choosing positive reappraisal strategies. They also tend to avoid less effective coping methods such as maladaptive avoidance, which involves shying away from emotional difficulties³⁹. Furthermore, their ability to flexibly regulate emotions is linked to improved problem-solving skills⁴⁴, demonstrating a versatile approach to handling emotional challenges that enhances their problem-solving effectiveness.

Social support

Meta-analytic research demonstrates that accessing and maintaining social support networks serves as an effective buffer against workplace stress⁴⁵. Research shows that individuals with higher emotional intelligence tend to develop and maintain higher-quality social relationships⁸ and demonstrate greater skill in communicating their emotional needs⁴⁶. Moreover, Mikolajczak et al.⁴⁷ found that those with higher EI scores were more effective at both providing and receiving social support during stressful situations. Building on this evidence, we propose that EI training can enhance workers’ capacity to manage high-risk workplace stress by developing their skills in self-regulation, problem-solving, and utilizing social support networks.

In this study, we aimed to address the problems with previous research, where genuine reactions to real-life high-stress occupation events appear to have been largely unexplored. We chose to investigate stress indicators within a military training environment, a setting known for its rigorous demands on specialist combat soldiers to disregard both physical pain and emotional distress, as outlined by Driskell et al.¹⁵. In sum, we set out to examine pre- and post-event cortisol levels surrounding a stress event, as well as in-the-moment cortisol levels across a set of real stress activities occurring in a high-stress workplace context (military training). Given that EI-trained soldiers understand the importance of recognizing, understanding, and regulating their emotions in high-stress situations, as well as seeking and maintaining social interactions from other soldiers, we hypothesized that:

Hypothesis 1

Compared to non-EI-trained soldiers, EI-trained soldiers will better regulate stress when exposed to stress activities, as demonstrated by lower biological indicators of stress before, during, and after stress activities.

Hypothesis 2

Since EI skills improve cognitive performance and memory under stress, EI-trained soldiers will perform better than non-EI-trained soldiers on (a) cognitive, (b) memory, and (c) behavioral activities during stress activities.

Method

Participants

We conducted our study using a sample of Australian Special Forces (ASF; hereafter “SF” when referring to Special Forces) soldiers. Convenience sampling was used to recruit volunteers from the pool of soldiers already selected to the Commando Reinforcement Cycle. Explanations of the study were communicated broadly to participants by a senior commando who commanded the SF Regiment Human Performance Wing and who had no influence or authority regarding the success of candidates over their training continuum. Because all participants met the same high standards for cognitive ability and fitness during selection, the two groups were likely equivalent in baseline performance capacity (even though specific pre-tests of those skills were not administered). Recruitment, participation conditions, and the research design followed Australian Defence Human Research Ethics Committee guidelines and were assessed on three occasions by a 12-person expert panel. Informed consent was obtained from all participants prior to their involvement in the study. If soldiers chose not to participate, their senior commanding officers were not informed. Soldiers undergoing Commando Reinforcement training have already completed the Commando Selection Course and have been deemed suitable against the required attributes to commence the 18-month reinforcement cycle.  Seventy percent of soldiers were existing Defence Force members from the Army, Navy, and Air Force, while 30% were from the SF Direct Recruitment Scheme , which draws candidates from the general civilian population. All applicants must pass a SF Entry Test, requiring above-average IQ, demonstrated advanced proficiency in mechanical aptitude, speed and accuracy in information assimilation, and cognitive flexibility. They must also demonstrate a high level of fitness and the ability to endure hardship for prolonged periods without sleep, food, or water.

The research was approved by the Australian Defence Human Research Ethics Committee (Protocol # 813 − 15), and the University of Queensland Human Research Ethics Committee. All methods were performed in accordance with relevant guidelines and regulations. All participants provided additional informed consent for the publication of any potentially identifying information in this open-access publication, in accordance with ethical guidelines for online publishing.

A Commando’s role is to conduct large-scale offensive, support, and recovery operations and to solve complex problems in support of Australia’s national interests − tasks that are beyond the scope and capability of regular Australian Defence Force units. ASF soldiers are required not only to survive but to thrive in environments of extreme stress, ambiguity, complexity, and confusion. Commando training is unique in comparison to that of conventional Army personnel, with the former emphasizing technical excellence in weaponry combined with heightened cultural awareness and political sensitivity for operations in a variety of highly stressful environments with soft-skill requirements. These soft-skill sets include stress management, adaptive coping, interpersonal dexterity, and emotional agility and regulation (which fall under the category of EI skills⁴⁸).

Seventy-eight male ASF soldiers voluntarily participated in the study, of whom 66 completed the survey, cortisol collection , as well as stress activities. Commando selection was open to female candidates, but none passed selection. Their average age was 26.24 years (SD = 3.46) and they had been enlisted on average for 5.76 years (SD = 3.41). Participant characteristics are shown in Table 1.

Table 1.

Independent groups T-test results for Pre-Treatment Differences.

Variables	Group	Mean	SD	t	df	p
Age	Training	25.66	3.88	−1.35	64	0.18
	Control	26.81	2.90
Years of enlistment	Training	5.39	2.84	− 0.70	64	0.48
	Control	5.98	3.86
Combat exposure (yrs)	Training	0.20	0.41	−1.69	64	0.10
	Control	0.39	0.50
EI	Training	104.54	14.67	1.12	64	0.27
	Control	100.44	14.93
Baseline Cortisol (ng/mL)	Training	7.95	7.11	− 0.22	64	0.83
	Control	8.34	7.08

Procedure

To test the effectiveness of the EI training program as a pre-emptive approach to managing ASF soldiers’ stress levels, we conducted an experimental pre-post-test design¹ within the context of an established ASF training continuum. We randomly assigned the participating soldiers, using a computerized number generator, to either a treatment (N = 35) or a control (N = 31) group. The differences in numbers occurred due to natural attrition during the reinforcement cycle and were not attributable to our study. These differences arose because some soldiers did not meet specific requirements, which are not publicly disclosed and therefore cannot be reported. All control group participants subsequently completed the EI training program.

After assignment, all 66 soldiers participated in a training exercise, with the treatment group receiving the EI training and the control group receiving technical training and participating in non-EI-related advanced human performance/recovery modalities. We measured participants’ baseline stress levels using salivary cortisol before randomly allocating them to the treatment and control groups. We then took an additional measure immediately before the start of the stress activity (T1AS), which we regard as the anticipatory stress response, to distinguish it from subsequent stress measures. Stress was further measured during (T2, one minute after the completion of the stress activity), post activity (T3, 15 minutes after the completion of each stress activity), and at a final time point after participants completed all the stress activities. Figure 1 illustrates the research procedure.

Fig. 1.

Diagram depicting the research procedure.

After we randomly allocated the participating soldiers to the experimental or the control group, all soldiers completed an online survey that included the EI measure, using computers in a supervised room. The following day, all participants took saliva tests to measure their baseline cortisol levels. The treatment group received three full 8-hour days of EI training; while the control group spent equivalent amount of time engaged in technical (non-EI) training. Six weeks after the treatment and control training sessions, all soldiers participated in two stress activities designed to simulate stressful tasks typically endured by SF soldiers, and thirty of them also participated in a third stress activity. Only 30 participants completed the third stress activity because attrition occurred during the Commando Selection Course; this attrition was not related to the study in any way. All participating soldiers completed each stress activity in their group (treatment or control) at the same time and followed the same daily routine before each stress activity, which reduces the potential impact of factors such as waking time, physical activities or diet on cortisol levels.

The 6-week interval between training and stress activities was required since one of the emotion-regulation techniques taught required sufficient practice to develop into unconscious competence and procedural muscle memory. Candidates were instructed to practice an emotion-regulation intervention technique – resonant frequency breathing – for 10 minutes a day, twice a day. The device used to train breathing was the MyCalmBeat, with its software loaded on computers within the training facility. Training sessions were logged under each soldier’s name so they knew compliance could be checked if necessary. Soldiers were also able to practice in their rooms with a breathing app called Kardia. Techniques taught to the control group were accessible daily and included instructions to practice for the same amount of time as part of their routine during the same period.

We administered the cognitive and memory performance tests immediately prior to the first stress activity and after the second stress activity. During the stress activities, all soldiers wore full assault order, which included 20 kg (44lb) of gear, armor, and assault weapons. Four weeks after soldiers conducted the three stress activities, they received a half-day booster training session to review core EI principles, discuss barriers to applying EI skills in high-stakes environments, and participate in scenario-based application exercises. One month later, soldiers received a second half-day booster session and then completed the behavioral performance test of pain tolerance and the collection of post‑intervention cortisol.

Treatment group training

The first author designed a training module for this study, grounded in EI principles of perceiving, using, understanding, and managing emotions intelligently, and drawing on the latest research in consultation with experienced senior combat Commandos and leading EI researchers. The training program consists of four key components: emotional perception; emotional understanding, emotional facilitation and emotional regulation (see Appendix A). Emotional perception (other) training focuses on facial expression recognition, tone-of-voice detection, non-verbal cue observation, and emotional aperture detection, helping participants accurately identify emotional states in others. To improve emotional perception (self), participants engaged in interoceptive activities to identify their earliest signs of stress, categorizing them as physical, mental, or emotional signals. They also participated in the “Name to Tame” exercise to help them label and process emotions effectively, fostering greater emotional clarity. Emotional understanding (others) training involves analyzing scenarios and real-life combat images, requiring participants to demonstrate emotional contemplation and understanding. Emotional understanding (self) training encourages participants to reflect on instances where they had encountered unexpected to unexplained emotions and to engage in continuous self-reflection. Emotional facilitation training includes exercises such as emotional laddering and emotion-based decision-making discussions, which train participants to leverage emotions for adaptive problem-solving and stress tolerance. Finally, emotion-regulation training incorporates resonant breathing, stress regulation techniques including the propriety protocol TKS™, and mood utilization strategies, equipping participants with the ability to downregulate or upregulate emotions effectively in high-pressure situations.

With these training components and activities, we designed the EI training to develop optimal strategies to regulate emotion and personal endocrine profiles in ways that foster mental clarity in critical stress-moments, promote constructive attitudes and a more positive mindset³⁶ toward stress events, and enhance psychological and physiological performance during prolonged high-intensity periods.

Each training session was conducted by the first author in collaboration with senior ASF soldiers who had extensive combat experience. The soldier shared narratives, recounting real-life situations, events, and anecdotes. These stories vividly illustrate instances where the application of specific EI skills played a crucial role in achieving success or ensuring survival. Conversely, the senior soldier also highlighted cautionary tales, demonstrating how the absence or neglect of these EI skills led to severe consequences, perilous situations, or even fatal outcomes. Throughout the training, participants learned about the science of the stress-response system, how to manage physiological reactivity, the biopsychosocial factors that influence emotional stability, how to read emotional data in the environment to reduce stress-inducing uncertainty, how to increase attentional focus and broaden situational awareness, and how to foster team trust and cohesion. They also completed a series of physical exercises designed to reduce their experience of anxiety prior to undertaking each stressful activity and to increase their physiological alacrity, consistent with the principles of EI. Across the training sessions, the soldiers were invited to provide feedback on how they practically and successfully implemented the strategies they learned during daily training, during missions, and outside of work. See Appendix A for illustrative examples of the EI-focused training modules that were administered to the treatment group.

Control group training

Control group training focused on physical and language-skill development and was equal in training time to that of the treatment group. The training involved altitude training in a hypoxic environment, functional lifting, and RapidRote^® Learning language-app training, a language-learning tool designed for military purposes. Practice sessions were logged in the software dashboard so compliance could be checked. Additionally, the control group engaged in recovery pre- and post-movement blood-flow protocols and vibration therapy using Niagara therapy equipment. Vibration therapy involved participants passively lying or sitting on vibration pads or chairs from www.niagramassage. These devices are designed to aid physical recovery, promote blood flow around the body, soothe aching joints and muscles, and reduce muscle tension. This non-clinical, alternative therapy is intended to enhance pre-movement activation and accelerate post-exercise recovery by increasing blood flow, oxygenation and nutrient delivery to muscles, aiding in the repair of micro-tears resulting from aerobic and resistance training.

Stress activities

The stress activities used in this study are established SF training protocols specifically designed to induce high levels of physical and psychological stress, activating the hypothalamic-pituitary-adrenal axis and leading to cortisol secretion³³. Psychological responses in this intense environment typically involve feelings of lack of control, fear, and anxiety, which contribute to increased cortisol levels^49,50. When trainees perceive situations as threatening or feel uncertain about outcomes, their stress response intensifies, further elevating cortisol secretion^11,51. These negative emotions further amplify the stress response, reinforcing the physiological impact of stress.

Stress activity 1: Stroop-shooting-activity

Soldiers engaged in reaction-time and target no-shoot (friend) vs. shoot (foe) interference tasks, including a pre-shoot cognitive math task under stress (Appendix B). Soldiers ran up and down three flights of stairs (equivalent to three stories) to simulate assault heart rates, followed by a cognitive math question. They then engaged in a shoot-no-shoot (friend or foe) Stroop-style target under time pressure (i.e., 4 -second exposure) at a 15-meter shooting range while surrounded by simulated enemy gunfire. The number of correct (foe) targets hit for each soldier was recorded. The soldiers were briefed using the following instructions:

• You will receive a single exposure of 4 s where you are to engage targets that are green in color.

• You are to fire two rounds only and only one round per target.

• You will not see the targets prior to the 4-second exposure.

There were four Stroop targets: a green rectangle, a red rectangle, the word red in green font, and the word green in red font. Of four possible targets, only two were green. Those who passed the test correctly fired one round into the word red in green font and one round into the green rectangle. Those who failed fired both rounds into the green square, fired both rounds into the word red in green font, fired one round into the green square and one into the word green in red font, or fired only one round.

Stress activity 2: Self-care-under-fire (SCUF) scenario

Under simulated enemy gunfire, soldiers ran three flights of stairs in full combat assault order to raise physiological stress responses and were then met with heavy simulated gunfire from a simulated enemy force. They were instructed to administer a tourniquet to stem bleeding from their lower limb, simulated with an artificial limb for safetys. Soldiers were blindfolded when instructed to apply the tourniquet. The outer package of the tourniquet was plastic and covered in a gel to simulate blood,  and making it difficult to open and administer. During this time, simulated enemy gunfire continued and a joint-terminal-attack-controller (JTAC) reported to a higher headquarters the current atmospherics for the ground force in contact with the enemy by radio. Information communicated included the size and disposition of the enemy force, coordination of the aeromedical evacuation asset (helicopter), and Military Grid Reference System coordinates (e.g., KL346 986) for helicopter landing-zone extraction . This stress activity concluded with a combat-information memory-recall test. Soldiers were assessed on their ability to accurately recall critical combat information, with details provided in the section on Cognitive, Memory and Behavioral Performance Tests.

Stress activity 3: Day-rappel

Soldiers ran up a 90-foot mock-helicopter tower and rappelled off under strict time constraints (required to complete the entire activity within 5 minutes), while physically strained from carrying their full combat load (approximately 20 kg of gear). Before rappelling, they were required to construct their own harness from tactical tape under time pressure, adding both a physical and cognitive challenge. The self-constructed harness was their only attachment point during the rappel, creating a situation that demanded both technical precision and emotional regulation in the face of height exposure and safety concerns (Appendix B).

Measures

Mayer-salovey-caruso emotional intelligence test (MSCEIT)

We asked the soldiers to complete a survey before the intervention. The survey included the MSCEIT and some demographic questions. The MSCEIT³² is a 141-item assessment that yields a total EI score, four branch scores, and two area scores for experiential EI (Branches 1 and 2 combined) and strategic EI (Branches 3 and 4 combined). Consistent with the view of EI as a cognitive ability, scoring of responses followed a correct/incorrect format similar to an ability-based IQ test, while requiring individuals to be attuned to social norms⁵².

Scoring of MSCEIT responses can be determined in two ways: (a) based on congruence with answers from emotion experts (i.e., expert scoring) or (b) based on the proportion of the sample that endorsed the same answer (i.e., consensus scoring)^13,52,53. Mayer et al.¹³ reported high agreement between the two scoring methods for correct answers (r =.91) and total test scores (r =.98). We administered the MSCEIT only during pre-intervention phase since our focus was on the EI training intervention, and hence we were concerned only with the participants’ ability-based EI prior to their engagement in the study. Caruso (personal communication, April 15, 2015)  noted that a second administration of the MSCEIT after an intervention is not recommended or accurate because of practice effects. The MSCEIT has demonstrated strong psychometric properties in numerous studies. The full test has shown high reliability (split-half reliability = .93-.91; test-retest reliability = .86) and good validity evidence, with discriminant validity from personality measures and convergent validity with other emotional intelligence measures¹³. The MSCEIT also demonstrates incremental validity in predicting important life outcomes beyond general intelligence and personality measures¹⁴.

Salivary cortisol

Given the emphasis on mental toughness in SF culture, self-report measures of stress are likely to be unreliable, as soldiers are conditioned to downplay or deny experiencing stress, even in objectively stressful situations. Following expert recommendations from ASF trainers, we employed objective physiological measures of stress (cortisol levels) rather than subjective self-reports. This approach provides a more accurate assessment of stress responses in this unique population and aligns with best practice in military stress research^20,54.

A physiological baseline measure of cortisol was taken at 0900 during a stress-free period to ascertain resting hormone levels. After EI training, we obtained cortisol measurements during the three stress activities to compare the stress levels between the control and training groups. The activities were conducted as follows: Self-Care Under Fire took place from 0900 to 1200, Stroop Shooting was conducted from 1400 to 1530, and Rappel occurred from 0900 to 0930. Samples were collected at consistent time points for both the treatment and control groups to ensure comparability: just before the activity (T1AS); one minute after completion of the activity (T2); and fifteen minutes after completion of the activity (T3). Two weeks after the last stress activity, we measured post-experiment cortisol.

We utilized a commercial product, IPRO Saliva Analysis (IPRO Interactive Ltd., Wallingford, UK), for cortisol measurements. This system uses mobile lateral-flow and saliva-analysis technologies to enable in-field testing without the need for laboratory facilities, therefore avoiding delay in obtaining results. While Ducker et al.⁵⁵ found that IPRO measurements exhibit moderate-to-large correlations (r =.53-.81) with laboratory-based ELISA immunoassays, they also noted that IPRO tends to underestimate salivary cortisol levels. However, this limitation was acceptable for our study as our primary focus was on relative changes in cortisol levels between conditions rather than absolute cortisol concentrations. The measurement range for the IPRO Cortisol test device is 2–40 nM (0.75–15.75 ng/mL). Soldiers were given an IPRO swab and instructed to leave the swab in their mouths until a blue line appeared (i.e., 2 to 4 min), indicating that sufficient saliva had been captured. They then placed the swab in a small reagent bottle and gently agitated for two minutes.

On the bottle’s label, we next instructed them to write their candidate number and the time and date. They then handed the bottle privately to a member of the research team and indicated whether they wanted the sample to be included. This procedure ensured that, if a soldier chosen not to participate, his sample could be destroyed without SOTEC instructors being aware.

After the saliva samples were collected, we kept them in a locked cooling refrigerator on base. The samples were later analyzed according to the IPRO testing manual. We recorded the results from the IPRO device into an electronic spreadsheet, and both the first author and a laboratory technician checked each entry to ensure accuracy of the data. The data file was kept in accordance with the Australian Defense Human Research Ethics Committee’s data-management protocols. The first author subsequently incinerated the saliva samples (as per Australian Defense Human Research Ethics Committee’s biological-sample handling protocols).

Cognitive, memory and behavioral performance tests

To examine the effectiveness of EI training on the participating soldiers’ performance, four performance tests were administered: one cognitive, one memory and two behavioral tests. The cognitive, memory, and behavioral performance tests used in this study are established assessment methods in military and stress research. The cognitive test was designed to assess working memory and cognitive function under stress, a critical skill for SF soldiers operating in high-pressure environments. Timed problem-solving tasks with social-evaluative pressure have been widely used to induce stress and evaluate cognitive performance in such conditions (see an example by Kirschbaum et al.⁵⁶). In our study, this test involved a math question asked prior to soldiers’ engagement in the Stroop shooting in Stress Activity 1. Senior Commandos presented a difficult math question (e.g., “If a dozen eggs cost 12 cents, how many can you buy for $1?”) in a hostile manner, creating time and social-evaluative pressures. Soldiers were required to answer the question within one minute. The accuracy of their answers was recorded by the first author. Mathematical calculation under pressure is a well-validated method for assessing working memory and cognitive function⁵⁶, with previous research demonstrating its sensitivity to stress-induced cognitive impairment²⁰.

For the memory test, soldiers were assessed on their ability to accurately recall critical combat information communicated to them in the JTAC report by radio during Stress Activity 2. The senior commando took the soldiers to a separate, closed area and asked eight questions regarding what they recalled hearing from the Joint Tactical Assault Controller report. The ability to recall critical combat information under stress is a key operational competency, as situational awareness and information retention are essential in high-risk military settings. The stress activity lasted five to six minutes, requiring soldiers to retain and process key operational details under pressure before being tested. The 8-item questionnaire, administered directly after the activity’s conclusion, aligns with standard approaches for evaluating situational awareness and memory retention in military environments. The memory-recall task has been used extensively in previous military studies to assess operational information retention under stress²⁰.

Behavioral tests include a shooting test and a pain-tolerance test. For the shooting test, the soldier’s shooting performance during Stress Activity 1 was recorded. Given that combat effectiveness depends on accuracy under stress, assessing shooting performance in a high-pressure scenario serves as a direct measure of operational task performance with high ecological validity. The pain-tolerance test examined how long soldiers could endure holding their arm in freezing water. The cold-water immersion test is a well-established method for assessing emotional regulation and physiological stress responses⁵⁷. We specifically designed the exercise to test the soldiers’ ability to regulate their emotions and tolerate discomfort, both of which are key components of psychological hardiness and mental fortitude in combat environments. The cold-water immersion test is a standardized stress-induction procedure with established reliability for measuring pain tolerance and emotion-regulation capacity^58,59.

Analyses

We did not exclude any participants with missing data from our final dataset (N = 66). However, if a participant missed a specific stress activity or performance test - often due to Army operational reasons - they were excluded from the analysis for that particular test. Outliers were identified as data points with z-scores exceeding ± 3. Across all variables, the prevalence of outliers ranged from 0% to less than 2%. Considering their minimal incidence and potential to offer valuable insights, these points were retained for analysis.

To test hypothesis 1, we employed a mixed ANOVA followed by post-hoc comparisons using paired t-tests with Bonferroni correction to compare cortisol levels at baseline, immediately before stress activities and at two time points after stress activities (1 minute and 15 minutes after, respectively). Mixed ANOVA enabled us to examine whether cortisol levels changed differently over time for the EI-trained group compared with the control group, while also assessing any baseline differences.  To further examine longer-term recovery patterns and potential differences in stress regulation between groups, we also used paired-sample t-tests to compare cortisol levels at baseline and two weeks after the last stress activity for the EI-trained group and the control group, respectively. To test Hypothesis 2, we ran a series of one-way ANOVA to compare differences in cognitive, memory, and behavioral performance between the EI-trained and control groups.

Results

Descriptive statistics

Tables 2 and 3 show the descriptive statistics and the correlations. This reveals that group membership (treatment vs. control) was significantly related to most cortisol measures (six of nine) assessed before, during and after stress activities as well as to all the cognitive, memory, and behavioral performance measures. Pre-intervention EI was not significantly relate to cortisol measures before, during and after stress activities, nor to any of the cognitive, memory, and behavioral performance measures.

Table 2.

Descriptive statistics and correlations for study variables.

Variable	n	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1.Groupa	66	0.47	0.50
2. EI	65	102.59	14.82	− 0.14
3. Baseline cortisol	66	8.13	7.05	0.03	− 0.18
4. Activity 1 cortisol T1	66	10.53	9.51	0.40**	− 0.15	0.22
5. Activity 1 cortisol T2	66	10.29	7.63	0.21	− 0.07	0.39**	0.23
6. Activity 1 cortisol T3	66	12.05	7.05	0.18	− 0.05	− 0.19	0.05	0.18
7. Activity 2 cortisol T1	66	12.02	8.18	0.25*	0.00	0.07	0.15	0.02	0.01
8. Activity 2 cortisol T2	66	14.98	8.76	0.42**	0.02	0.11	0.17	0.07	0.11	0.53**
9. Activity 2 cortisol T3	66	12.11	7.27	0.37**	− 0.08	− 0.19	0.14	0.18	0.59**	0.29*	0.35**
10. Activity 3 cortisol T1	32	13.77	9.38	0.75**	− 0.03	0.05	0.12	0.15	0.16	0.24	0.47**	0.22
11. Activity 3 cortisol T2	32	13.95	8.06	0.43*	− 0.10	− 0.13	− 0.23	− 0.19	0.25	0.18	0.15	0.32	0.39*
12. Activity 3 cortisol T3	32	19.58	9.38	0.30	− 0.04	0.04	0.13	0.22	0.39**	0.08	0.33	0.44*	0.46**	0.48**
13. Cognitive	65	0.38	0.49	− 0.37**	0.01	− 0.27*	− 0.46**	− 0.17	− 0.07	0.01	− 0.26*	− 0.14	− 0.36*	0.02	− 0.32
14. Memory	65	4.17	1.87	− 0.55**	0.08	0.46**	− 0.16	− 0.07	− 0.26*	− 0.05	− 0.1	− 0.43**	− 0.59**	− 0.03	− 0.29	0.05
15. Shooting	65	0.74	0.44	− 0.48**	0.11	0.16	− 0.13	− 0.01	− 0.06	− 0.01	− 0.05	− 0.12	− 0.41*	0.11	0.12	0.18	0.41**
16. Pain	58	41.14	17.03	− 0.64**	0.15	− 0.42**	− 0.34*	− 0.23	− 0.09	− 0.26*	− 0.35**	− 0.22	− 0.26	− 0.25	− 0.16	0.41**	0.123	0.32*
17. Post-cortisol	54	4.61	4.61	0.15	− 0.38**	0.12	0.18	0.00	− 0.13	0.07	− 0.07	− 0.06	0.29	0.11	0.17	0.09	− 0.26	0.03	0.17

^a 0 = treatment group and 1 = control group. ^*p <.05. ^**p <.01.

Table 3.

Correlations among the study variables for the intervention and control conditions.

Variable	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16
1. EI	−	− 0.15	0.01	− 0.05	− 0.01	0.03	0.20	0.14	− 0.04	− 0.04	0.08	− 0.04	− 0.12	0.06	− 0.03	− 0.32
2. Baseline cortisol	− 0.20	−	0.15	0.28	− 0.18	− 0.07	0.03	− 0.31	0.04	− 0.17	− 0.13	− 0.27	0.62**	0.24	− 0.43*	0.08
3. Activity 1 cortisol T1	− 0.30	0.39*	−	0.20	− 0.04	0.09	0.02	0.02	− 0.32	− 0.50	− 0.18	− 0.41*	0.14	0.11	− 0.05	0.08
4. Activity 1 cortisol T2	− 0.03	0.54**	0.10	−	0.34	− 0.09	− 0.10	0.21	− 0.15	− 0.39	0.06	− 0.30	0.02	0.10	− 0.16	− 0.05
5. Activity 1 cortisol T3	− 0.03	− 0.21	0.00	− 0.04	−	− 0.15	0.04	0.58**	− 0.34	0.17	0.42	− 0.18	0.09	0.01	− 0.23	− 0.30
6. Activity 2 cortisol T1	0.04	0.22	− 0.05	0.03	0.07	−	0.44*	0.26	0.19	0.05	− 0.16	0.33	0.07	0.21	0.12	0.07
7. Activity 2 cortisol T2	− 0.07	0.23	− 0.06	0.10	0.03	0.58**	−	0.22	0.14	0.03	0.06	− 0.11	0.35	0.25	− 0.05	− 0.21
8. Activity 2 cortisol T3	− 0.20	− 0.13	− 0.05	0.00	0.57**	0.18	0.25	−	− 0.29	0.23	0.50	0.06	− 0.14	0.18	− 0.04	− 0.17
9. Activity 3 cortisol T1	− 0.07	− 0.04	0.32	0.19	0.07	0.44	0.65**	0.10	−	0.20	0.27	0.14	− 0.44	− 0.32	0.44	0.46
10. Activity 3 cortisol T2	− 0.22	− 0.19	− 0.26	− 0.26	0.00	0.26	− 0.28	0.08	− 0.18	−	0.63*	0.18	0.19	0.35	− 0.14	0.11
11. Activity 3 cortisol T3	− 0.20	0.14	0.34	0.27	0.18	0.25	0.63**	0.19	0.74**	0.06	−	0.13	− 0.14	0.28	− 0.11	0.04
12. Cognitive	− 0.03	− 0.31	− 0.37*	0.06	0.11	− 0.09	− 0.17	− 0.06	− 0.48	0.32	− 0.44	−	− 0.34	− 0.02	0.29	0.34
13. Memory	0.15	0.50**	0.00	0.10	− 0.45**	0.16	− 0.10	− 0.46**	− 0.19	0.38	− 0.17	− 0.11	−	0.26	− 0.33	− 0.21
14. Shooting	0.05	0.12	0.00	0.11	0.07	− 0.08	− 0.04	− 0.15	0.21	0.20	0.22	0.03	0.11	−	0.08	0.06
15. Pain	0.22	− 0.65**	− 0.17	− 0.07	0.19	− 0.38*	− 0.25	0.08	0.11	0.09	0.26	0.17	− 0.41*	− 0.04	−	0.56**
16. Post-cortisol	− 0.52**	0.22	0.34	0.02	0.05	− 0.07	0.11	− 0.01	0.29	0.14	0.49	− 0.14	− 0.27	0.21	− 0.03	−

Notes. Correlations for participants randomly assigned to the intervention condition are shown below the diagonal line (N = 35) and correlations for participants randomly assigned to the control condition shown above the diagonal line (N = 31).

^* p <.05. ^**p <.01.

Pre-treatment differences between control and treatment groups

To assess pre-treatment differences between the control and treatment groups, baseline scores were compared across age, years of enlistment, combat exposure, EI, and cortisol. Results of independent-group t-tests appear in Table 1. No significant differences were found between the control and treatment groups at the start of the intervention.

Hypothesis testing

Stress activities physiological results

In Hypothesis 1, we proposed that EI-trained soldiers would be better able to regulate in-the-moment-stress before and after stress activities compared to non-EI trained soldiers, as demonstrated by similar baseline cortisol but lower cortisol levels measured one minute before each stress activity (T1AS); one minute after the completion of each stress activity (T2); and fifteen minutes after the completion of each stress activity (T3). Baseline cortisol levels did not significantly differ between the EI-trained and control groups across all stress activities (all p >.80), ensuring comparability prior to intervention exposure.

Stress activity 1: Stroop shooting task

The results of mixed ANOVA show a significant effect of time (baseline, T1AS, T2 and T3) and condition (EI-training vs. control) on cortisol levels, F (2.725, 174.394) = 2.93, p <.05, η_p² = .04. Further decomposition of this interaction revealed that, although there was no significant difference in cortisol levels between the EI-trained and control groups at baseline (mean difference = 0.38, p =.83), those in the EI-trained group experienced a significantly lower cortisol at T1AS than those in the control group, with a mean difference of 7.57 at 95% confidence interval [CI] [3.25, 11.90], F (1, 64) = 12.23, p <.01, η_p² = .16. At T2 and T3, while those in the EI-trained group also experienced lower cortisol than those in the control group (mean difference of 3.24 and 2.52, respectively), these differences were not significant, F (1, 64) = 3.07, p =.09 and F (1, 64) = 2.14, p =.15. A bar graph of these findings is presented in Fig. 2.

Fig. 2.

The effect of EI training (treatment versus control) on cortisol levels under stress activity 1: Stroop shooting task. Note. X axis shows levels of cortisol measured at baseline, 1 min before (T1), 1 min after (T2), and 15 min after the Stroop shooting task (T3). Error bars represent standard errors.

Stress activity 2: Self-care-under-fire task

The results of the mixed ANOVA show a significant effect of time (baseline, T1AS, T2 and T3) and condition (EI-training vs. control) on cortisol levels, F (3, 192) = 2.98, p <.05, η_p² = .04. Decomposition of this interaction reveals that, while there was no significant difference in cortisol levels between the EI-trained and control groups at baseline (mean difference = 0.38, p =.83), those in the EI-trained group experienced significantly lower level ofcortisol at T1AS than those in the control group, with a mean difference of 4.11 at 95% CI [0.18, 8.04], F (1, 64) = 4.36, p <.05, η_p² = .06. The same pattern was also found at T2, with a mean difference of 7.34 at 95% CI [3.40, 11.29], F (1, 64) = 13.81, p <.01, η_p² = 0.18. Finally, at T3, those in the EI-trained group continued to show significantly lower cortisol levels than those in the control group, with a mean difference of 5.30 at 95% CI [1.94, 8.66], F (1, 64) = 9.94, p <.01, η_p² = .13. A bar graph of these findings is presented in Fig. 3.

Fig. 3.

The effect of EI training (treatment versus control) on cortisol levels under stress activity 2: Self-care-under-fire task. Note. X axis shows levels of cortisol measured at baseline, 1 min before (T1), 1 min after (T2), and 15 min after the Self-Care-Under-Fire task (T3). Error bars represent standard errors.

Stress activity 3: Day-rappel task

A smaller group of participants (EI-trained group = 17 and control group = 15) also went through one additional round of stress activities. The pattern of results was similar to the above two primary stress activities. In Stress Activity 3, the results of the mixed ANOVA showed a significant effect of time (baseline, T1AS, T2 and T3) and condition (EI training vs. control) on cortisol levels, F (2.81, 84.33) = 5.89, p <.01, η_p² = .16. We found that those in the EI-trained group experienced significantly lower cortisol at T1AS than those in the control group, with a mean difference of 13.95 at 95% CI [9.42, 18.49], F (1, 30) = 39.47, p <.01, η_p² = .57. The same pattern was also found at T2, with a mean difference of 6.84 at 95% CI [1.49, 12.19], F (1, 30) = 6.81, p <.05, η_p² = .19. At T3, those in the EI-trained group also experienced lower level of cortisol than those in the control group with a mean of 16.94 and 22.56 respectively, however, this difference was not significant, F (1, 30) = 3.05, p =.09. A bar graph of these findings is presented in Fig. 4.

Fig. 4.

The effect of EI training (treatment versus control) on cortisol levels under stress activity 3: Day rappel task Note. X axis shows levels of cortisol measured at baseline, 1 min before (T1), 1 min after (T2), and 15 min after the Day Rappel task (T3). Error bars represent standard errors.

Post-intervention cortisol

We measured cortisol again two weeks after the last stress activity, although only for a smaller group of participants (EI-trained group = 26 and control group = 28). We again used a paired-sample t-test to compare cortisol levels at baseline and two weeks after the last stress activity for the EI-trained and control groups, respectively. Results indicated that cortisol decreased significantly from baseline (M = 7.74, SD = 7.15) to two weeks after the last stress activity (M = 3.92, SD = 2.98) for the EI trained group, t = 2.74, df = 25, p <.05, Cohen’s d = .76). For the control group, results of the paired-sample t-test showed no significant difference in cortisol at baseline (M = 8.29, SD = 7.25) and two weeks after the last stress activity (M = 5.26, SD = 5.71), t = 1.81, df = 27, p =.08).

Taken together, results for Stress Activities 1 to 3 and the two-weeks follow-up show that the EI-trained soldiers were better able to regulate their stress levels before, during and after exposure to stress activities than non-trained soldiers, as demonstrated by consistently lower cortisol levels.

Cognitive, memory, and behavioral performance results

In Hypothesis 2, we proposed that the EI-trained group would perform better on cognitive, memory, and behavioral performance than the control group. We tested the participating soldiers’ cognitive performance before they engaged in Stress Activity 2 (the Stroop shooting task) by asking them to complete a pre-shoot cognitive task under stress. In the task, they were required to answer a math question asked by a senior commando in a hostile manner under time and social-evaluative pressures (after running up and down three stories to simulate assault heart rates). ANOVA results show that 19 of 34 (56%) EI-trained soldiers were able to think under pressure and answered correctly, in comparison to 6 of 31 soldiers (19%) in the control group, F (1, 64) = 10.31, p <.001) (Fig. 5).

Fig. 5.

The effect of EI training (treatment versus control) on accuracy (maths test and shooting correct target). Note. Error bars represent standard errors.

For memory performance, we asked the soldiers to recall eight items vocalized by the Joint Tactical Assault Controller after they underwent Stress Activity 2. We did this to assess their ability to recover from stress and divert attention immediately to the next operational task. ANOVA results again showed a significant difference in accurate memory recall between the treatment (M = 5.15, SD = 1.52) and control (M = 3.10, SD = 1.62) groups, indicating that EI-trained soldiers recalled more items correctly, F(1, 64) = 27.71, p <.001 (Fig. 6).

Fig. 6.

The effect of EI training (treatment versus control) on memory recall. Note. Error bars represent standard errors.

For behavioral performance, we assessed shooting accuracy and pain tolerance. Soldier’s shooting performance during Stress Activity 1 was recorded. ANOVA results showed that 32 of 34 (94.1%) EI-trained soldiers hit the correct target, in comparison to 16 of 31 soldiers (51.6%) from the control group, F (1, 64) = 19.18, p <.001 (Fig. 5). For the pain-tolerance task, those in the treatment group lasted significantly longer on average (M = 50.94 min) than those in the control group (M = 29.08 min), F (1, 57) = 39.65, p <.001) (Fig. 7).

Fig. 7.

Mean number of minutes lasted in the pain tolerance test. Note. Error bars represent standard errors.

Discussion

In this study, we examined the efficacy of EI training as a pre-emptive approach to regulate physiological stress responses in a high-risk workplace. To accomplish this, we employed a randomized control experimental design in a natural field setting and objective measures of EI and stress. Stress levels were measured using objective physiological markers (cortisol) before, during and after each of three stressful activities during real military training. We found that at least one of these cortisol measures was lower for the EI-trained soldiers compared to non-EI-trained soldiers across all three l activities. We further found that the EI-trained group demonstrated better cognitive (including memory) and behavioral performance compared to the control group. Importantly, these results were obtained in a naturalistic military setting, using operationally relevant stress-induction tasks, thereby enhancing the ecological validity of our findings.

While our results showed lower cortisol levels in EI-trained soldiers, it is important to note that this represents a more regulated stress response rather than simply reduced cortisol output. The relationship between cortisol and performance follows an inverted U-shaped curve, where both insufficient and excessive cortisol can impair performance⁶⁰. This is supported by our behavioral and cognitive performance data. The EI-trained group’s cortisol patterns suggest they achieved a more adaptive response - maintaining sufficient physiological arousal to meet task demands while avoiding the performance-degrading effects of excessive stress responses. This balanced and metabolically efficient cortisol response aligns with our theoretical framework, particularly regarding self-regulation, where EI training appears to help individuals better calibrate their physiological stress responses to situational demands. The improved performance we observed across behavioral (94.1% vs. 51.6% shooting accuracy), cognitive (56% vs. 19% correct calculations), and endurance measures (72% longer pain tolerance) suggests that EI training helps achieve an optimal zone of stress arousal rather than simply minimizing cortisol output. This interpretation is consistent with our finding that EI-trained soldiers still showed elevated cortisol during stress activities, but at more moderate levels compared to the control group, indicating engagement without an overwhelming stress response. To better understand these differential effects across phases of stress exposure, we explored the underlying physiological mechanisms. The differential effects observed between anticipatory cortisol responses (T1AS) and task execution responses (T2, T3) may be explained by distinct physiological mechanisms regulating stress. Resonant breathing practice, included as part of the EI training, likely contributed to reductions in anticipatory stress via stimulation of the vagus nerve and enhanced parasympathetic nervous system activity^12,61. Breathing at a resonant frequency of between 5 and 7 breaths per minute synchronises respiratory and cardiovascular rhythms, increasing vagal tone and reducing sympathetic arousal⁶². This modulation affects brain regions involved in threat detection and anxiety regulation, leading to lower cortisol release in anticipation of stressors^34,63.

In contrast, smaller differences between groups during and after the Stroop task (T2, T3) may be attributable to the intense physical exertion involved (e.g., running stairs, shooting under time pressure). Physical stressors trigger cortisol release through direct activation of the hypothalamic-pituitary-adrenal axis³³, a mechanism less sensitive to psychological regulation techniques. Thus, while EI training and resonant breathing appear to modulate anticipatory (psychological) stress responses, cortisol reactivity during extreme physical tasks may primarily reflect physiological demands, overriding emotion regulation influences⁶⁴.

Our study advances understanding of the relationship between EI and stress by showing, in a real-life context, that EI training can be beneficial for managing stress levels and supporting performance for individuals engaged in high-risk occupations. This was found using an RCT design incorporating an objective measure of stress tracked over time. While these findings highlight the potential benefits of EI training in high-risk settings, they should be interpreted within the scope of the study’s design and participant characteristics.

Practical implications

This research highlights the critical importance of a pre-emptive approach to the physiological well-being and emotional stability of high-risk occupational employees. Considering the significant impact of stress-related illnesses, which consume up to 19% of global annual healthcare spending⁶⁵, a reparative, after-the-fact approach does not make strategic or economic sense when the cost of inaction is high. Even with ASF soldiers’ high levels of mental fortitude, characterized by high IQs, the ability to excel in arduous environments, and frequent combat training, our experimental study demonstrates that EI training better equips these soldiers to manage stress and perform at higher levels in comparison to soldiers that did not receive EI training.

Our study demonstrates the performance benefits of ability-based EI training and suggests potential well-being advantages for high-risk occupational employees, suggesting its utility as a critical component of formal training programs. Implementing such training may not only support employees’ ability to manage stress effectively but could also enhance their operational performance in no-fail mission scenarios, potentially extending the longevity of their careers in high-risk environments.

While this study focuses on military personnel, the findings have broader implications for civilian, organizational, and educational settings where individuals encounter high-pressure environments. As Shields et al.⁶⁶ demonstrate, our physiological stress response activates similarly whether facing physical danger or social threats like public humiliation in a hostile board meeting. In corporate settings, employees in leadership roles, emergency response teams, and high-stakes decision-making positions commonly experience chronic stress and performance pressure. Integrating EI training into executive development programs could help leaders manage workplace stress, navigate interpersonal challenges, and make more effective decisions under pressure. In educational settings, universities and professional programs could incorporate EI development into their curricula, particularly in fields such as medicine, law, and business, where emotional regulation is essential for success. Additionally, public safety and emergency response personnel, including paramedics, firefighters, and law enforcement officers, regularly face high-stakes, life-threatening situations where emotional regulation and reliable performance under pressure are critical. Implementing EI training in these fields could provide tools to regulate stress responses, reduce burnout, and improve effectiveness in crisis situations.

Theoretical implications

We identify two important theoretical implications that arise from the study.

First, our findings demonstrate that EI training is associated with objectively measurable outcomes in both stress response and performance. While all participants experienced high stress during the exercises, the EI-trained cohort showed lower physiological stress reactions than the control group and superior performance across multiple domains - behavioral, cognitive and memory tasks . However, since we did not measure EI post-training, future research should examine whether these performance differences are mediated by changes in EI abilities. Our research extends understanding of EI training effects on employee stress management and performance in real work settings, directly addressing Kotsou et al.’s [⁶⁷ p162] call for research “…to confirm that EI interventions improve work and academic performance.” These findings help bridge the gap between EI as a theoretical construct and its practical implementation in workplace settings.

Second, our study provides insights into the application of ability-based EI training. Our EI training was based on the Mayer and Salovey¹⁸ ability EI model, which has been commended for its conceptual and theoretical foundations^16,68. Our findings demonstrate that ability-based EI training can be effective in improving stress management and performance outcomes. This aligns with Hodzic et al.‘s³¹ meta-analytic findings showing positive effects of ability-based EI interventions. Similarly, Kotsou et al⁶⁷. emphasised the need for future EI training research to move beyond self-reports and assess performance outcomes in real-world contexts. While our study adds to the growing evidence supporting the utility of ability-based EI training, we did not directly compare different EI models and thus cannot make claims about their relative effectiveness.

Methodological implications

This research has two important methodological implications. First, by using cortisol as a physiological measure of stress, we sought to complement widely used self-report stress measures and offer a viable framework for conducting stress research in high-demand and high-stress environments. Cortisol provides an objective, biological marker of stress, which is particularly useful in cases where individuals may downplay, exaggerate, or be unaware of their stress levels⁴, particularly when workplace, role, gender, or cultural expectations influence self-reporting. Moreover, cortisol tracking over time helps distinguish between acute and chronic stress patterns and serves as a valuable tool for differentiating the nature of stress responses⁴.

Second, while conducting stress research, researchers typically ask participants to recall a historical event⁶⁹or participate in stress-inducing interventions that do not always reflect the types of stress experienced in their workplaces (e.g., Trier Social Stress Test⁴³). In our study, the stress activities following the EI training intervention involved real-time, realistic stressors that closely mirrored high-risk and high-stress tasks and contexts. These stress activities were an essential part of the ASF selection process and were designed to replicate real operational tasks, ensuring ecological validity.

Studying stress as experienced in-the-moment in response to realistic work activities—rather than in a recalled or proxy manner—offers researchers a way to obtain more reliable findings that can better generalize to populations beyond those under investigation. An additional benefit of using operational simulations is that, by engaging participants in life-like experiences with real consequences, they can put theory-based knowledge obtained through training into practice through in-the-moment stress activities. A large body of research (see⁷⁰ for a meta-analysis) supports simulation-based learning as superior to classroom-based learning. This is particularly important for kinesthetic individuals whose daily jobs involve outdoor physical activity, given such individuals might find theory-based classroom education difficult to sit through for extended periods.

Limitations and future research directions

Our study design has several strengths, including the use of an RCT experimental design in a naturalistic setting with realistic stressful activities and objective physiological measurement of stress. Nonetheless, we acknowledge several limitations that should be considered when interpreting and generalizing the findings.

Notably, all soldiers in our sample had passed rigorous medical and psychological screening as part of ASF selection. This pre-selection minimizes variability in health conditions or medication use that could confound cortisol responses; however, no additional screening for such factors was conducted in our study, which we acknowledge as a limitation.

One key limitation is the relatively homogeneous sample of healthy, fit, and intelligent young males used in the study, which raises the question of generalizability of the findings to other populations. Additionally, the sample was relatively small, which might have contributed to conservative results. Conducting large-scale training intervention studies in this population is particularly challenging due to the logistical complexities of using costly biological tests in a tightly controlled and top-secret environment.

To improve generalizability, future studies should aim to include larger and more diverse samples where feasible. Gender is a particularly interesting demographic to consider in future intervention samples, as research suggests that women have higher baseline EI and might differentially benefit from EI training⁷¹. Moreover, future research could examine EI training in occupations that experience more common everyday stressors. For example, employees in healthcare and service industries, where emotional labor is a significant job demand, may benefit even more from EI training. By extending this research to broader populations, future studies can provide a more comprehensive understanding of EI training’s effectiveness across diverse work environments.

Another limitation is that we did not measure EI post-training, which limits our ability to directly attribute the observed changes to improvements in EI. Future studies should incorporate pre- and post-training EI assessments to address this limitation. We further recommend that future EI-training studies test individuals on the branches of EI to identify aspects that would most benefit from training and implementation; and to enable tailoring of the training program to individual needs. Future research could also consider including pre- and post-experiment qualitative interviews to assess coping mechanisms and thought processes used, and how these change in response to the training intervention. This would provide valuable information on whether high-EI individuals are more impervious to stress or are better equipped at down-regulating negative emotions after experiencing it, as well as how EI training is used by, and may differentially impact, high and low-EI individuals. Additionally, we recommend future research include an active control group to examine and isolate the unique effects of EI training from other forms of stress-regulation interventions.

Finally, our reliance on cortisol as a single biological marker of stress may not capture the full complexity of physiological stress responses. Future research could incorporate additional physiological measures (e.g., heart rate variability, heart rate, blood pressure) via wearable devices such as the WHOOP (WHOOP Inc., Boston, MA) to provide a more continuous (24/7) and comprehensive understanding of stress reactivity. Moreover, while objective biological markers provide valuable physiological data, we acknowledge that subjective interpretations of stress play a crucial role in shaping both physiological and psychological responses to stress. Research has demonstrated that factors such as stress mindsets³⁶, anticipatory stress responses^72-111, and perceived psychosocial stress⁴⁹ can differentially predict whether stress responses are adaptive or maladaptive. Therefore, a comprehensive approach to stress research should integrate both objective biological markers and subjective experiences of stress to provide a more holistic understanding of stress processes and outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

The authors contributed equally to this work.

Data availability

The de-identified datasets generated and analysed during the current study are not publicly available due to security and privacy restrictions associated with Australian Defence Force personnel and training. De-identified data are available from the corresponding author on reasonable request, subject to approval by the Australian Defence Human Research Ethics Committee and relevant Defence authorities.

Declarations

Competing interests

The authors declare no competing interests. J.B. King is the founder and director of BioPsychAnalytics Pty Ltd, a consultancy providing behavioural science services to government, defence, and corporate sectors. This business was established after the completion of this research and did not influence the study design, data collection, analysis, or interpretation of results.