Mitigating the Health Impairment Vicious Cycle of Air Traffic Controllers Using Intra-Functional Flexibility: A Mediation-Moderated Model

Abstract

Air traffic controllers (ATCOs) make a significant contribution to ensuring flight safety, making this profession a highly stressful job globally. Job demands–resources (JDR) theory proposes a health impairment process stemming from job demand (complexity) to mental workload, which causes job stress, resulting in compromised flight safety. This vicious cycle is evident among ATCOs and is recognized as an unsustainable management practice. To curb this process, we propose intra-functional flexibility as a conditional factor. Intra-functional flexibility refers to the flexibility in the reallocation and coordination of resources among team members to help in urgent times. This is a relatively new concept and is yet to be empirically tested in the ATCO context. ATCOs work in a dynamic environment filled with sudden surges of urgent jobs to be handled within short time limits. Intra-functional flexibility allows standby crews to be called to ease these tensions when needed. To ascertain the role of intra-functional flexibility in mitigating health impairment among ATCOs, a questionnaire was administered to 324 ATCOs distributed across Saudi Arabia. Partial Least Squares Structural Equation Modeling (PLS-SEM) analysis exhibited two critical findings: First, the study revealed the prevalence of a vicious cycle of health impairment among Saudi ATCOs, whereby job complexity leads to increased mental workload, resulting in elevated levels of job stress. Secondly, the presence of intra-functional flexibility weakened this vicious cycle by mitigating the influence exerted by mental workload on job stress. That is, the mediation-moderated model proposed in this study provides empirical evidence supporting the applicability of intra-functional flexibility in mitigating the dire suffering of ATCOs. This study discusses limitations and future research directions in the end.

1. Introduction

1.1. Background

It is generally accepted that the air traffic control (ATC) environment is highly stressful because of the high-stakes implications of the services that it offers [1]. In such a situation, job complexity and mental workload are inevitable in the ATC function, due to the nature of the work. These roles come with dire repercussions, where a slight human error can cause the loss of life and high damage to assets (e.g., aircraft) [2]. Indeed, ATC has been classified as a job involving the highest mental workload [3] in the aviation industry, as it entails a high level of complexity and dynamic processing in maintaining safe flight operations [4].

Advanced technology is used heavily in the ATC context [5] in many ways. For instance, ATC providers have introduced automated simulators to improve ATCOs’ skills and get them used to extra workload or abnormal events in flight movements [6]. Furthermore, some regions (Eurocontrol) have implemented air traffic flow management systems to control the capacity of a specific airspace or sector [7]. Also, ref. [8] introduced a novel technique to estimate ATCOs’ workload that can be calibrated with the existing system or future advanced automated systems. Regardless of the systems, workload remains a challenging factor to reduce [4]. This phenomenon presents a need to identify ways to mitigate the impact of job complexity and mental workload (as specific job demands) on job stress, so as to ensure safe flight operations and the well-being of ATCOs.

1.2. Research Gap

In the case of ATC, job complexity is one of the leading causes of stress. ATC’s job complexity is defined as the difficulties associated with controlling a traffic situation [9] consisting of various general activities [1] that collectively cause stress. For instance, it was reported that job complexity serves as a significant indicator of air traffic controllers’ (ATCOs’) workload [10], implying that the mental burden of work could play the role of a mediator in the complexity–workload–stress sequential process [4].

This sequential relationship aligns with the process of health impairment outlined in the JDR model, where job demands (job complexity) lead to mental workload [1] and, subsequently, to job stress. Also, the JDR model suggests that job resources (intra-functional flexibility) can reduce the stress level by allowing for dynamic reassignment and utilization of internal resources. However, this concept remains underexplored in the ATC context. Accordingly, it is proposed that job complexity predicts job stress through the mediated association (mental workload) and a moderated association (intra-functional flexibility).

1.3. Study Objectives and Contributions

There are a few objectives that we aimed to achieve in this research, as follows:

• To extend the JDR model to the Saudi ATC operational context.
• To explore the mediating role of mental workload in the health impairment process in the job complexity–job stress relationship.
• To examine how IFF moderates the mental workload–job stress relationship.

Accordingly, this research contributes to the literature in three essential dimensions: Firstly, it expands the job demands–resources (JDR) model by assessing its relevance to a specific group (ATCOs) within a particular country (Saudi Arabia). Secondly, it offers evidence supporting the universal applicability of the health impairment process (physical demand—mental demand—response) by evaluating the mechanism of job complexity to mental workload to job stress. Thirdly, it identifies whether rarely used job resources (intra-functional) serve as a conditioning factor in mitigating the negative consequences of mental workload on job stress. More specifically, this research identifies both a mediator and a conditional variable to investigate whether the vicious cycle of job complexity to mental workload to job stress can be weakened.

2. Theoretical Underpinning

2.1. The of Job Demand–Resources (JDR) Model

The JDR framework is extensively applied in human resource settings [11]. It is an occupational stress model that argues that job strain is a response to the existing mismatch between what is required from the workforce and the available resources offered to manage these jobs. Job demands correspond to such characteristics related to a job in need of a steady mental or emotional skillset or efforts like workload and job complexity, along with job insecurity [12].

The JDR theoretical framework has a few essential presumptions. Firstly, job aspects could be either job-related demands or resources. Secondly, it presents double-layered processes; job demands develop a health impairment process, while job resources develop a motivational process. Simply put, job demands are related to psychological or physical costs; conversely, job resources become functional when reaching the targets of the work in mitigating job demands and related costs, as well as reinforcing a person’s learning and growth. Thirdly, job-related resources safeguard the influence exerted by job demands [13]. At a later stage, personal resources were added to the model, which can be employed as a potential moderator either in the health impairment or motivation path [13].

2.2. JDR Model Applications

The JDR model is also applied in various contexts, such as healthcare and education, but is limited in the aviation context [14]. The application of this theory in this industry has been recent and infrequent. Within aviation, ref. [15] reported that job crafting impacted the service recovery, working along with job-related meetings among flight attendants. Moreover, the relationship of job crafting with intention to quit is mediated by work-related engagement. Similarly, within the flight attendant context, ref. [16] indicated a direct positive association of job crafting with organizational citizen behavior, as well as an indirect association through augmenting job-related resources and contesting job-oriented demands. Concerning the ATC scenario, ref. [17] used the JDR theory to explain the mechanisms of job burnout. It was reported that job stress is linked positively with Polish ATCOs’ burnout, both directly and indirectly, via personal resources (i.e., general self-efficacy).

Exploring further, this study employs the JDR theory to propose that job complexity (as job demand) influences mental workload (psychological demand) and job stress positively. Also, it is proposed that mental workload mediates the relationship between job complexity and job stress with respect to the health impairment process. Lastly, it is proposed that intra-functional flexibility (as a job resource) moderates the mediated relationship of the health impairment process.

JDR theory proposes that appropriate job resources could mitigate job stress. Ref. [18] reported that intra-functional flexibility was part of the job resources that reduced stress levels and improved sales performance, providing flexibility in resource reallocation and coordination among groups and the ways in which such resources are structured and utilized. It has two dimensions: resource flexibility and configurational flexibility [18]. The resource flexibility dimension is defined as “the resources available for use by the organization” [19], while the configurational flexibility dimension is the extent to which individuals are empowered to decide how these resources are utilized.

The essence of intra-functional flexibility is to own flexible resources that can be deployed and given autonomy in terms of how they are utilized and configured. Intra-functional flexibility could play a buffering role concerning the relationships between job demands (job complexity and mental workload) and response (job stress). That is, when the workload is overwhelming, the unit is given the flexibility to mobilize staff for support and ease ATCOs’ stress. However, there is limited research utilizing this notion of intra-functional flexibility as a conditional factor. Accordingly, we propose that job complexity predicts job stress through a mediation (mental workload) and a moderated (intra-functional flexibility) relationship.

The JDR model was selected due to its capacity to conceptualize the development of the health impairment process. For instance, it explains the impact exerted by high job demands (e.g., job complexity along with cognitive overload) on health-related impairments (e.g., job stress and burnout) in the ATC context. Knowing that ATCOs work in risk-sensitive tasks, under time pressure, and in environments of sustained vigilance, the pathway of health impairment generates a robust theoretical framework to quantify such stressors and their psychological outcomes. This match between the operational characteristics of ATC work and the structure of the theoretical framework facilitated a smooth implementation to assess the association (see Figure 1).

3. Hypothesis Development

Researchers have implied that job complexity is more likely to bring a high degree of pressure and demands [20]. Logically, a job with high complexity would usually require additional or sustained physical, psychological, and emotional efforts. In the aviation context, ref. [21] indicated that job demands lead to job stress among flight attendants. Parallel to the JDR theory, these excessive demands arising from complex jobs lead to significant stress on employees [22]. With the continuous growth of the civil aviation field, an increasing demand for air services has led to increased job volume and complexity, leading to increased job stress among air control managers and administrators [23]. Consistent with the JDR theory, the following hypotheses are put forward:

H1. 

Job complexity influences mental workload positively.

H2. 

Job complexity influences job stress positively.

It is acknowledged that high work demands drain resources such as time, energy, and emotions, contributing to mental workload [24]. The mental workload has been reported as the primary factor contributing to employees’ job stress in various industries [25]. ATCOs often experience the highest mental workload associated with the aviation field [3], due to the distinctive nature of an ATCO’s job [26], which requires them to apply concurrent critical thinking and make quick, accurate decisions. The intense temporal pressure further compounds the emotional turmoil [27] among ATCOs. In addition, their job function requires a high mental workload to prioritize multiple tasks, while at the same time managing both their psychological and cognitive resources, as well as working around the clock. JDR theory posits that when job demand exceeds an employee’s capacity, individuals may experience a high mental workload due to the additional effort required to complete the job [28], potentially leading to stress. Therefore, the following hypothesis is put forward:

H3. 

Mental workload has a positive association with job stress.

Scholars have established that ATCOs’ job complexity remains a compelling indicator of their mental workload [1,4]. At the same time, the mental workload has been consistently found to impact job stress [25]. Combining the above, mental workload acts as a mediator in the complexity–workload–stress mechanism. This means that mental workload represents an intermediate psychological mechanism explaining the impact created by job complexity, impacting job stress. This indicates that job complexity does not directly influence job stress but, rather, increases the mental workload, which, in turn, increases job stress. This is a clear reflection of the health impairment process as theorized by the job demands and resources framework. This sequential process is supported by the job demands–resources (JDR) theoretical framework: job strain is a response to the gap between what is required from the employees and the available resources to manage these jobs [12]. In the ATCO context, job demands (e.g., job complexity) indeed lead to psychological demand, which translates to mental workload and results in responses manifested as job stress, representing a health impairment process of JDR. Underpinned by JDR, the following mediation hypothesis is formulated:

H4. 

Mental workload mediates the job complexity–job stress association.

IFF as a job resource is related to the organization’s goal accomplishments and could be used as a basis for buffering job demands [13]. 1IFF is a relatively new concept; it was first introduced in the sales occupation and found to reduce job stress and improve sales performance [18]. It was shown that the higher the IFF, the lower the unhealthy effect of stress on job performance, hence indicating that IFF is an effective supportive mechanism to generate positive consequences [18]. With regard to the ATC context, ATCOs work in a dynamic environment that exposes them to sudden increases in unpredictable ATC demands. With limited time to respond quickly and safely to such demands [29], IFF may be helpful for such situations. This includes giving ATC operations and supervisors the authority to mobilize ATCOs on duty, set up standby crews, or call ATCOs to duty to fill urgent job gaps as needed. In other words, IFF allows for resolving conflicts and coordinating with other functions in a way that makes the job less stressful. Thus, it is likely that intra-functional flexibility is a boundary condition that weakens the health impairment process of complexity–workload–stress.

Thus, the following hypothesis is proposed: Moderation is defined as the impact of IFF that modifies the direction or strength of the mental workload–job stress relationship. Accordingly, the positive effect of mental workload on job stress is weakened by IFF through the ATC units in high-demand situations. Indeed, this is a reflection of the buffering hypothesis of JDR, stating that job resources reduce the unhealthy influence exerted by high demands, affecting stress-induced outcomes.

H5. 

The mediated job complexity–job stress association via mental workload is moderated by IFF, such that at a high level of IFF, the positive association between mental workload and job stress is weaker.

4. Research Methodology

4.1. Target Population and Sampling

To test the model, a cross-sectional survey approach was utilized by analyzing responses from ATCOs operating in Saudi Arabia. Being strategically located in the center of the Middle East, Saudi ATCOs provide services to a massive volume of flights flying in and through the Saudi airspace, which connect Asia with Africa, and Southeast Asia to Western Europe. Saudi ATCOs were selected as the study context due to the job complexity of the ATC function and high service demands for Saudi Arabian ATC. The ATC services are provided by the Saudi Air Navigation Services (SANS) company, responsible for managing multiple ATC units (13) located at different airports with variations in ATCO numbers. The population size was 594. Due to the scattered distribution of the population, as well as the variations in ATCOs’ characteristics and the ATC service provided, a proportional sampling technique was employed [30]. Following the proportional sampling procedure, 450 ATCOs from 13 units were requested to fill in the questionnaire survey. Ref. [31] claimed that any sample size larger than 200 can be considered sufficient for quantitative studies. Finally, 324 completed surveys were returned, providing a response percentage of around 72%.

4.2. Measurement Instruments

The measurement instrument was an online self-completed questionnaire. The language used in the questionnaire was English, as Saudi ATCOs are proficient in English, the language of formal ATC work worldwide. This questionnaire had two sections: The initial section sought to gather demographic information regarding the ATCOs. In contrast, the subsequent segment solicited responses about their attitudes and opinions towards the researched constructs, which were adopted from earlier research. The mental workload was measured employing the CarMen-Q scale developed by [32]. The scale includes performance (5 items), emotional (7 items), cognitive (10 items), and temporal (7 items) dimensions; 0.8 is the scale reliability. The job complexity measure, adapted from [33], consists of 4 items, with a reliability of 0.72. A scale was developed by [34], consisting of four items—with a Cronbach’s alpha of 0.87—was applied for measuring job stress. The intra-functional flexibility scale has two dimensions: (1) resource flexibility (a 3-item scale from [18]; Cronbach’s alpha 0.9) and (2) configurational flexibility (a 3-item scale from [35]; Cronbach’s alpha 0.81). During the pre-test, a panel of aviation experts added two items to the configuration flexibility, namely, “planning ahead” and “coordination,” due to their importance in the ATC context. Consequently, configuration flexibility now comprised a total of five items.

4.3. Pre-Test and Pilot Testing

Pre-testing and pilot testing are considered to be important requirements in questionnaire preparation to eliminate any mistakes and identify weak points in the questionnaire [36]. Hence, a pre-test was carried out on six aviation experts in the field. These experts possess an extensive understanding of ATC functions. Upon receiving their input, the questionnaire was revised accordingly for clarity. Specifically, the pre-test panel suggested minor changes and the addition of two items to the configurational flexibility dimension, as mentioned above. Subsequently, a pilot study was conducted on 30 air traffic controllers. This number was selected based on the suggestion of [37] that 25 to 40 should be enough for pilot testing. According to the results of the pilot study, criterion-related validity and construct validity were established. Reliability was also met for all of the constructs, as the Cronbach’s alpha for all of the scales was more than 0.70, exhibiting that there was internal consistency among the items of each construct [38].

4.4. Data Analysis Strategy

Data cleaning and descriptive analysis were performed using SPSS 28. Subsequently, this study applied the PLS-SEM approach to assess the theoretical framework, due to the following reasons: (1) The theoretical framework was considered to be a reflective–formative measurement model, which PLS-SEM is more proficient in analyzing to lessen I and type II errors [39]. (2) The primary goal of this investigation was causal-prediction-oriented [40]; hence, the PLS approach was a good choice of statistical tool. (3) The PLS-SEM method is considered to be a great structural model estimation method when dealing with a complex model, such as dealing with the mediation-moderated relationship that we propose in this study [41].

5. Findings

5.1. Common Method Bias and Demographics

To ensure that there was no single-source bias or common method bias (CMV), both procedural and statistical remedies were used. Two different types of Likert scales were applied for measuring the constructs: a four-point Likert scale ranging from zero (never) to three (always), and a five-point Likert scale ranging from one (strongly disagree) to five (strongly agree). The intent was to lower the bias of the standard method as a procedural remedy. Afterwards, Harman’s single-factor technique was applied as a statistical remedy. The results of the analysis showed that the variance explained by the first factor was 18.06% (below the threshold limit of 50%). It can therefore be concluded that CMV was not a serious issue in the current data [42].

As evident from

Table 1. Demographic Profile.

Description		Frequency	Percent
Gender	Male	307	94.8
	Female	17	5.20
Age	25 years old and below	28	8.60
	26–30 years old	113	34.90
	31–35 years old	82	25.30
	36–40 years old	34	10.50
	41–45 years old	35	10.80
	46 years old and above	32	9.90
Marital Status	Single	88	27.20
	Married	228	70.40
	Divorce	8	2.50
Education	Diploma	155	47.80
	Bachelor’s degree	149	46.00
	Master’s degree or higher	20	6.20
Working Experience	Less than 5 years	99	30.60
	5–10 years	115	35.50
	11–15 years	29	9.00
	More than 15 years	81	25.00
Job Position	Tower	132	40.70
	Approach	97	29.90
	Area	95	29.30
Unit	Jeddah	124	38.27
	Riyadh	80	24.70
	Dammam	32	9.90
	Madinah	10	3.09
	Abha	16	4.90
	Hail	9	2.80
	Alhasa	4	1.20
	Jazan	13	4.00
	Qasim	12	3.70
	Tabouk	7	2.20
	Taif	11	3.40
	Yanbu	6	1.90
	Najran	0	0
	Total	324	100.00

, the dominant gender among ATCOs was male (94.8%), and about 70% of them were aged between 26 and 35 years. Most of them were married (70%) and had worked for about 5–10 years in ATC operations. All possessed diplomas or higher, and they mainly worked in towers (40.7%) and came from the Jeddah ATC unit (38.27%).

5.2. Descriptive Statistics

Under the descriptive analysis, three important assessments were conducted—namely, analysis of the mean, standard deviation, and normality—to understand the ways in which the participants gave responses to each construct. The means of the mental workload dimensions ranged from 1.150 to 2.584, based on a 0 (never) to 3 (always) scale, with performance demand being the highest. Job complexity, intra-functional flexibility, and job stress demonstrated means ranging from 2.343 to 4.169, using a 5-point Likert scale of 1 (strongly disagree) to 5 (strongly agree). Finally, the normality results illustrated that all constructs achieved a satisfactory normality range (i.e., −3 to +3), as the scores of skewness ranged between −1.421 and 1.007, while the scores of kurtosis ranged between −1.014 and 2.191 [43] (see

Table 2. Results of descriptive statistics.

Construct	Mean	Std. Deviation	Skewness	Kurtosis
(1) Mental Workload	1.898
(a) Cognitive Demand	1.806	0.608	0.386	−0.832
(b) Temporal Demand	2.053	0.588	−0.765	0.260
(c) Performance Demand	2.584	0.526	−1.421	1.630
(d) Emotional Demand	1.150	0.969	0.768	−0.501
(2) Job Complexity	2.969	1.161	0.144	−1.014
(3) Intra-Functional Flexibility	4.015
(a) Configurational Flexibility	4.169	0.730	−1.218	2.191
(b) Resources Flexibility	3.861	0.818	−0.748	0.580
(4) Job Stress	2. 343	1.314	1.007	−0.321

). Consequently, the results showed that the current dataset had a normal distribution.

5.3. Measurement Model Assessment

Ref. [44] recommended the assessment of the construct reliability, convergent validity, and discriminant validity, which are considered to be the psychometric characteristics of the constructs, to examine the measurement model. The reliability was checked for composite reliability, along with Cronbach’s alpha (α). The reliability values (α) for all of the constructs ranged between 0.804 and 0.985 (see

Table 3. Results of the measurement model.

	Mean	SD	Outer Loading	CA	CR	AVE
Cognitive Demand (CD)				0.929	0.94	0.613
CD1: My work involves the processing of complex information.	2.102	0.706	0.692
CD2: My job requires thinking and choosing between different alternatives.	2.278	0.655	0.783
CD3: I have to make difficult decisions.	2.225	0.64	0.706
CD4: My job requires handling a lot of knowledge.	2.052	0.699	0.869
CD5: My job requires dealing with information that is perceived with difficulty.	1.522	1.007	0.765
CD6: I have to deal with information that is not easily understood.	1.272	0.899	0.803
CD7: My job requires a lot of information.	1.867	0.76	0.773
CD8; My job requires memorizing a high amount of data.	1.528	0.84	0.861
CD9: My work is mentally intense.	2.074	0.589	0.731
CD10: I have to do a great search and information gathering to carry out my tasks.	1.142	0.935	0.826
Configurational Flexibility (CF)				0.899	0.925	0.714
CF1: Please rate your negotiation skills. (e.g., offering a direct route with lower altitude).	3.88	0.95	0.764
CF2: Please rate your persuasiveness and assertiveness skills.	4.009	0.908	0.84
CF3: Please rate your communication skill (e.g., complying with the standard phraseology).	4.216	0.811	0.851
CF4: Please rate your coordination skill.	4.383	0.872	0.897
CF5: Please rate your planning ahead skill.	4.355	0.782	0.866
Emotional Demand (ED)				0.985	0.987	0.917
ED1: I have trouble forgetting the problems of my job.	1.154	1.028	0.908
ED2: My work makes me nervous.	1.102	1.065	0.939
ED3: My work is affecting my personal relationships (family, friends…).	0.966	1.078	0.942
ED4: I feel very tired, physically fatigued.	1.198	0.964	0.982
ED5: My work affects me a lot emotionally.	1.201	0.988	0.968
ED6: When I finish my workday, I feel a lot of physical exhaustion.	1.201	0.959	0.978
ED7: My work is affecting my health.	1.225	1.001	0.983
Job Complexity (JC)				0.828	0.879	0.648
JC1: I do tasks that are extraordinary and difficult.	3.549	1.401	0.579
JC2: I have to make complicated decisions in my work.	2.87	1.528	0.893
JC3: I receive so many tasks that I cannot handle them orderly.	2.478	1.413	0.874
JC4: I receive tasks with many dependencies and interactions.	2.978	1.357	0.833
Job Stress (JS)				0.973	0.98	0.925
JS1: My job is extremely stressful.	2.38	1.419	0.931
JS2: Very few stressful things happen to me at work®.	2.269	1.396	0.986
JS3: I feel a great deal of stress because of my job.	2.392	1.297	0.961
JS4: I almost never feel stressed because of my work®.	2.333	1.347	0.969
Performance Demand (PD)				0.940	0.954	0.805
PD1: My job requires maintaining a high level of attention.	2.565	0.603	0.871
PD2: My job requires no mistakes.	2.417	0.6	0.835
PD3: I have to give very precise responses	2.611	0.585	0.952
PD4: My mistakes can have serious consequences.	2.645	0.578	0.915
PD5: My job involves a lot of responsibility.	2.682	0.562	0.908
Resource Flexibility (RF)				0.819	0.892	0.733
RF1: Our operation department is able to shift resources from one crew activity to another if needed.	3.753	0.889	0.833
RF2: Our operation department, if under achieving versus plan, is capable of changing the way it deploys its resources within the operation department in order to put things back on track.	3.938	0.914	0.877
RF3: Our operation department is capable of redeploying its own resources if essential for fulfilling their strategic and/or operational requirements.	3.892	1.053	0.858
Temporal Demand (TD)				0.804	0.859	0.55
TD1: I have to work constantly; I cannot take breaks beyond strict regulations.	D
TD2: The pace of work is excessive, difficult to reach even by an experienced worker.	1.815	0.851	0.649
TD3: I often work with annoying interruptions.	1.352	0.959	0.683
TD4: I cannot stop my work when I need it.	2.444	0.741	0.765
TD5: The pace of work is imposed on me.	2.25	0.704	0.818
TD6: The accomplishment of my tasks demands a lot of speed.	2.401	0.732	0.781
TD7: It is normal for me to accumulate the tasks.	D

D—item deleted due to low loading (<0.40), SD—standard deviation; CA—Cronbach’s alpha; CR—composite reliability; AVE—average variance extracted; ^®—reverse-coded.

), fulfilling the suggested limit of 0.70 [45]. The composite reliability values ranged from 0.859 to 0.987, exceeding the minimum threshold value of 0.70, which is considered to be adequate as recommended by [46] guideline. Two metrics were then used to assess the adequacy of the convergent validity, i.e., average variance extracted (AVE), and factor loading. The items’ loading satisfied the recommended value of 4.0 or above [47]. Furthermore, the AVE for all constructs indicated satisfactory values between 0.550 and 0.925 (above 0.50), except for temporal demand. Two items of temporal demand were deleted (i.e., TD1 = 0.435; and TD7 = 0.469) due to the weak loading required to pass the minimum score of 0.50 for AVE [48].

The heterotrait–monotrait ratio of correlations (HTMT) was measured to confirm the discriminant validity of the reflective constructs. The results (see

Table 4. Results of HTMT.

	CD	CF	ED	JC	JS	PD	RF	TD
CD
CF	0.251
ED	0.531	0.139
JC	0.106	0.084	0.086
JS	0.349	0.355	0.256	0.299
PD	0.142	0.151	0.212	0.083	0.143
RF	0.101	0.374	0.032	0.094	0.129	0.166
TD	0.477	0.19	0.451	0.13	0.31	0.804	0.165

) show that the HTMT values of all reflective constructs were lower than the conservative threshold limit of 0.85, indicating acceptable discriminant validity [49]. Accordingly, the constructs were genuinely distinct from each other.

5.4. Higher-Order Construct Assessment

Higher-order constructs were adopted to capture the multidimensional nature of key theoretical variables, particularly those that cannot be adequately represented by a single observed dimension. For example, constructs such as intra-functional flexibility and mental workload often comprise subdimensions (e.g., resource flexibility and configurational flexibility in IFF; or temporal, physical, performance, and cognitive demands in workload) that must be captured to preserve the construct’s theoretical integrity. Such an approach enables researchers to estimate multidimensional constructs by combining their first-order reflective or formative dimensions into a more parsimonious and theoretically cohesive second-order construct.

Two constructs are conceived as reflective–formative higher-order constructs (HOCs), which are composed of various lower-order constructs (LOCs) [39]. Mental workload consists of four LOCs: cognitive, performance, temporal, and emotional [32]; and intra-functional flexibility has two LOCs: configurational and resources [18]. As recommended by [39], this study employed a two-stage technique for assessing all of the HOCs. In the first step, an evaluation of all LOCs was performed following the usual process of the reflective measurement model. At the subsequent level, the HOCs’ evaluation was performed employing the usual method featured by the formative measurement model.

Initially, the evaluation of the collinearity of the formatively measured constructs was conducted by employing the variance inflation factor (VIF). The VIF values for all LOCs were between 1.125 and 2.765 (see

Table 5. Results of higher-order constructs.

HOC	LOC	Outer Weight	Standard Deviation	VIF	t-Statistics	p-Values
MW	CD	0.627	0.022	1.852	21.473	0
	TD	0.392	0.017	2.764	10.541	0
	ED	0.126	0.019	1.486	26.705	0
	PD	0.099	0.028	2.034	4.585	0
IFF	RF	0.039	0.037	1.125	8.256	0
	CF	0.986	0.036	1.125	23.783	0

), i.e., below 3.0 [44]), confirming no collinearity issues. Subsequently, an assessment of the outer weights and significance of each LOC was conducted utilizing the bootstrapping technique, with 5000 resamples. This indicated that all mental workload (MW) LOCs (i.e., performance, cognitive, emotional, and temporal) and intra-functional flexibility (IFF) (i.e., configurational, resources) were statistically significant at p < 0.01.

5.5. Structural Model and Hypothesis Testing

Table 6. Results of structural model.

	Std Beta	Standard Error	t-Statistics	p-Values	VIF	F2	R2	Q2
Direct Relationships
JC -> JS	0.251	0.046	5.493	0.000	1.018	0.084 (S)
JC -> MW	0.129	0.053	2.448	0.007	1.000	0.017 (T)	0.017	NA
MW -> JS	0.288	0.053	5.428	0.000	1.078	0.105 (S)	0.266	0.239
Mediation Relationship					VAR
JC -> MW -> JS	0.037	0.018	2.085	0.037	34% (PM)

JC: job complexity; MW: mental workload; JS: job stress; F²: T—trivial, S—small; NA—not applicable.

presents the VIF values for all of the exogenous constructs (i.e., mental workload, job complexity) on job stress, ranging between 1.018 and 1.078 (below the threshold limit of 3), meaning that there is no collinearity problem [44]. Secondly, the bootstrapping technique results show that job complexity positively impacted mental workload significantly (H1: β = 0.129, t = 2.448, p < 0.05); thus, H1 is supported. Also, job complexity positively impacted job stress significantly (H2: β = 0.251, t = 5.493, p < 0.001); thus, H2 is supported. Next, mental workload also positively impacted job stress significantly (H3: β = 0.288, t = 5.428, p < 0.001); thus, H3 is supported. For the mediation hypothesis, it was found that mental workload mediates the job complexity–job stress relationship (H4: β = 0.037, t = 2.085, p < 0.05); accordingly, H4 is supported.

Thirdly, the explanatory power (R²) for the endogenous construct was evaluated. R² values can be substantial, moderate, or weak, at values above 0.67, 0.33, and 0.19, respectively [50]). The results showed that job complexity and mental workload accounted for 26.6% (R²: 0.266) of the variance in job stress. In a nutshell, all of the exogenous variables exercise moderate explanatory power for job stress. Fourthly, ref. [51] indicated that the effect size (F²) analysis can be classified into three groups: 0.02, 0.15, and 0.35, reflecting small, medium, and large effect sizes, respectively. When the path with job stress served as an endogenous construct, two constructs, job complexity (0.084) and mental workload (0.105), exhibited a small effect size.

Lastly, the framework’s predictive relevance for Q² was confirmed by employing the blindfolding technique. The results showed that the Q² value for job stress (0.239) was greater than zero [52]. Subsequently, we used the PLS prediction technique to test the prediction relevance of the key endogenous construct, JS [53]. As evident from

Table 7. Assessment of the PLS predictions.

		PLS-SEM	LM	PLS-SEM—LM
Item	Q2 Predict	MAE	MAE	MAE	Decision
JS1	0.185	1.045	1.051	−0.006	Moderate Predictive Power
JS2	0.154	1.047	1.034	0.013
JS3	0.141	0.944	0.945	−0.001
JS4	0.132	1.006	1.008	−0.002

Note: MAE—mean absolute error; LM—linear model; PLS-SEM—proposed PLS model.

, three items from the JS construct had MAE values that had lower prediction errors in the PLS-SEM model as opposed to the linear model (LM), excluding JS2, suggesting that the essential target endogenous construct of JS had medium predictive power [53]. Hence, the findings suggest that the moderating effects of IFF will likely influence the overall relationship of JC–MW–JS among ATCOs. This improves the representation and predictability of the study’s findings from the holdout sample to the target population.

Consequently, the current study employed a new moderated mediation approach through PLS-SEM [41] to investigate whether intra-functional flexibility (IFF) strengthened or weakened the mediating impact exerted by mental workload between job complexity and job stress (i.e., hypothesis H5). As

Table 8. Results of the moderation–mediation relationship.

Conditional Mediation Modeling	Std Beta	SE	t-Value	p-Value	LB	UB
Index of Moderated Mediation	−0.015	0.007	−2.026	0.021	−0.028	−0.004
IFF at −1 SD (Low)	0.051	0.022	2.305	0.011	0.018	0.091
IFF at Mean (Medium)	0.037	0.017	2.191	0.014	0.013	0.069
IFF at +1 SD (High)	0.023	0.014	1.717	0.043	0.006	0.053

presents, the p-value was less than 0.05 and the confidence interval excluded zero (95% confidence interval: −0.028 to −0.04), supporting the index of moderated mediation. Subsequently, the mediated influences at various stages of the moderation from intra-functional flexibility were analyzed. It was found that at low IFF (0.051), the path coefficients for the conditional mediation effect of mental workload were stronger, while at high IFF (0.023), they were found to be relatively weaker (see Table 8). Along with this, the IFF effects for high, medium, and low levels of the moderator are significant. Overall, the results endorse hypothesis H5.

6. Discussion

6.1. Direct Association

Based on the data analysis, job complexity has a positive association with mental workload (H1 was supported), which is the same as shown by previous studies in the ATC context [1,4,23,26]. That is, job complexity is indeed an important predictor of mental workload in the ATC context [2]. As expected, job complexity has a positive association with job stress (H2 was supported), which is in line with JDR theory, and this supports the argument presented by [22], which indicated that demands generated from job complexity increase job stress significantly. Outside the ATC context, the literature reports conflicting findings about the impact created by job complexity on job stress. Although ref. [54] indicated a positive effect of job complexity on job stress, ref. [55] reported a curvilinear U-shaped relationship. This research provides support for a positive linear association in the ATC context, indicating that any increase in job complexity will lead to proportionately increased stress. It is not a curvilinear U-shaped relationship with a negative relationship initially, and then it progresses to a positive one after a certain level of job complexity. That is, the job complexity level in ATC is so high that it has already passed the stage of not causing stress; every slight increase in complexity leads to stress.

Our findings also indicate that mental workload influences job stress positively (H3 was supported), consistent with empirical studies [25,56]. Indeed, high mental demands for work will drain precious resources such as time, energy, and emotions, compromise well-being, and create stress [57]. Concerning the ATC scenario, this finding is consistent with the claims of [58], who reported that it is important to monitor mental workload in the ATC context because the nature of an ATCO’s job is mentally demanding, as it requires “parallel” thinking and quick decision-making [2]. As explained, ATCOs often face emotional tasks that emanate from the need for accurate and critical decision-making under intense temporal pressure [27].

6.2. Mediation Association

Furthermore, the results reveal that mental workload mediates the job complexity–job stress association (H4 was supported). This finding aligns with the health impairment process, stating that physical job demands (job complexity) result in mental job demands (mental workload) and lead to the development of systematic unhealthy outcomes (job stress) within organizations [12]. Also, Ref. [20] reported that complex jobs usually need additional physical, psychological, and emotional efforts from individuals, thus causing stress. Indeed, the vicious cycle of health impairment is evident in the ATCO context in Saudi Arabia.

6.3. Moderation Association

Lastly, our findings show support for the moderating hypothesis. It was hypothesized that intra-functional flexibility weakens the health impairment mediation process of the job complexity to mental workload to job stress sequential process. In a condition of high intra-functional flexibility, the mental workload–job stress association was weaker (H5 was supported). Ref. [59] reported that the positive connection of job complexity with job stress could be mitigated through emotional stability and openness to experience. It was found that IFF was the mitigator in the ATC context. For ATC units with higher IFF, the increased job complexity and mental workload led to increased job stress, but this was found to be less prevalent in units with lower IFF. These findings are consistent with the JDR theory, which acknowledges that job resources (intra-functional flexibility) weaken the health-impairment-driven process. For instance, job resources moderate the link of job demands with exhaustion [28]. Intra-functional flexibility was reported to moderate the connection of role stress with sales job performance [18]. Indeed, the same moderating function of intra-functional flexibility is evident in the ATC context.

In the ATC context, intra-functional flexibility (IFF) encompasses several key aspects, including sufficient break times for ATCOs, additional support for resolving unexpected issues, assistance with complex tasks through double-checking potential conflicts between aircraft, coordination with other units, and decomposing sectors as necessary. This helps to manage the complexity and mental workload, resulting in lower job stress among ATCOs. The essence of IFF is ensuring that ATCOs possess flexible resources that can be deployed, utilized, and configured to weaken the influence of physical and mental job demands on job stress.

7. Implications

7.1. Theoretical Implications

This research adds value to the literature in three dimensions: Firstly, it broadens the JDR model by testing its applicability to air traffic controllers (ATCOs) specifically, as well as a specific country (Saudi Arabia). Secondly, it provides evidence about the universality of the health impairment process (job demand–psychological demand–response) by evaluating the sequential process of job complexity to mental workload to job stress. Thirdly, it examines whether rarely used job resources (intra-functional) serve as a conditioning factor in mitigating the negative mental workload–job stress association. In other words, the outcomes identified a mediator and a conditional factor to investigate whether the vicious cycle of job complexity to mental workload to job stress can be weakened.

Furthermore, this study makes an additional contribution to the organizational behavior literature. Initially, it extends the JDR model toward the Saudi Arabian ATC setting. Indeed, the JDR model is applicable in ATC and the Saudi context, where job complexity demand positively predicts mental workload, and mental workload positively predicts job stress among ATCOs in Saudi Arabia. Mental workload is a significant positive predictor of job stress, indicating that the mental workload dimensions of performance expectations, temporal pressures, cognitive demands, and emotional loads are reasons for increasing job stress levels among ATCOs. Secondly, evidence of the applicability of the JDR health impairment process is provided here. This research addresses the gaps in the limited empirical studies evaluating the health impairment mechanism in the ATC context. Ref. [56] demonstrated that mental workload increases job-related stress in ATC in the Philippines, whereas [1] identified job complexity as a key indicator of ATCOs’ workload. These studies implied mediation but did not test the mechanism empirically. The findings empirically showed that the health impairment mediation mechanism indeed manifested in the ATC context. Thirdly, this study contributes to the ATC literature by introducing intra-functional flexibility as a moderator in the JDR model. It was empirically found that a condition undermines the unhealthy influence exerted by job demand (mental workload), affecting the response (job stress). In careers with high job demands (physical or mental demands), organizations can implement intra-functional flexibility to alleviate stress. Also, it is noteworthy that the configurational flexibility scale was improved [18] by adding two items suggested by the pre-test expert panel to measure ATC intra-functional flexibility better.

7.2. Practical Implications

The findings of this study offer substantial insights into ATC management. Firstly, the significant influences exerted by mental workload on job stress suggest the urgency of reducing the mental workload within the ATC context. According to the analysis, among the mental workload dimensions, the highest mean was the performance demand (mean = 2.584), followed by temporal demand (mean = 2.053) and cognitive demand (mean = 1.806), and the lowest was the emotional demand (mean = 1.150). This indicates that the ATCOs were most burdened by the performance demand. Performance demand refers to how successful the ATCO is in resolving the conflicts between flights safely [32]. The management is strongly advised to implement concentrated on-the-job training for ATCOs, which aims to intensify ATCOs’ capabilities (e.g., planning, resolving conflicts, etc.) and increase their cognitive abilities to handle performance demands more effectively [60]. The on-the-job training should include exercises with more flights than in real-life scenarios, enabling ATCOs to increase their awareness and plan for conflicting situations promptly. This would improve their skills in response to temporal demand [61].

Furthermore, the use of ATC simulators is highly encouraged due to their ability to imitate real-life flight movements [62]. Additionally, ref. [63] suggested implementing an automation system to enhance situational awareness while reducing mental workload. The management of ATC may consider restructuring the aerospace industry optimally [9]. Secondly, ATC management is suggested to revisit the authority matrix to grant power to ATC supervisors to implement intra-functional flexibility, where supervisors are given additional manpower support when needed (e.g., standby crews) through proper, pre-defined procedures. Correspondingly, a variety of in-house skillset training is needed to support the implementation of intra-functional flexibility so that ATC can be mobilized to support a broad range of jobs when the need arises.

8. Limitations and Future Studies

Despite providing considerable insights to both theory and practitioners in the ATC context, this study has some limitations that future research could work on. Firstly, this research focuses solely on the ATC context in Saudi Arabia, making it complex to generalize the results to the context of other countries. Future studies are suggested to replicate the model in other Middle Eastern and North African (MENA) countries—especially countries in the Middle East region (e.g., Bahrain, UAE, Oman, Jordan, Egypt), as they share similar geographical airspace—to provide more generalized findings that can be adopted widely in the ATC context in the region. Secondly, this study proposes three predictors of job stress. Other variables related to ATC job stress could be researched in future investigations, such as the variations in terms of infrastructure, operating procedures, working hours, volume of flights, and airspace size. Thirdly, as the outcomes recorded a moderate level of mental workload due to a massive drop in ATC service demand caused by the COVID-19 pandemic while collecting data, future researchers are advised to replicate this study when ATC demand returns to the pre-pandemic level. Lastly, this study did not employ any control variables, e.g., age or gender. Future studies are advised to introduce control variables to verify whether there are differences or variations.