Examining the relationships between time pressure, challenge and hindrance technostressors and work outcomes in online labour markets

Online labour markets (OLMs) have emerged as an established and thriving work practice in recent years (Williams et al., 2019). OLMs are websites that facilitate interaction between individuals seeking to purchase or sell cognitive work performed remotely and are often referred to as “digital gig platforms”, e.g. Amazon Mechanical Turk (MTurk) ^[1] and Upwork (Horton, 2010; Stephany et al., 2020). Over 162 million workers are employed through OLMs (Manyika et al., 2016), with 13% considering platform work as their full-time job (Anderson et al., 2021). Despite exponential growth and opportunities, the working conditions of OLMs remain a matter of debate. The precarious nature of work, characterised by a lack of social protections, makes workers more vulnerable to exploitation (Fairwork, 2023).

While platforms promote temporal flexibility (i.e. the ability to adjust work patterns to accommodate personal needs and preferences), the precarious nature of work in reality limits this flexibility (Lascău et al., 2022). Alongside economic precarity (Bajwa et al., 2018) and financial precarity (Berg, 2016), we see that temporal precarity, i.e. “losing the capacity to claim and protect private time” exists in these platforms, as well as concern due to uncertainty and insecurity related to the pace and scheduling of work (Lascău et al., 2022; Ugaz, 2022, p. 971). Time plays a crucial role in OLMs where “time”, “timing” and “tempo” constitute the basis of work structure and related experience (Tietze and Musson, 2002; Zheng and Wu, 2022). Workers on OLMs tend to operate with a persistent need to do more in less time, often described as a “rat race” (Bajwa et al., 2018; Jauch, 2020), which can be a source of frustration. To date, limited research has examined the social implications of platform work from a temporal perspective, particularly experiences of time pressure and technostress. As almost every task faced by OLM workers is technology-mediated, the potential to experience both challenge (good) and hindrance (bad) technostress exists. The consequences of both forms of technostress in OLMs are an overlooked research theme (Umair et al., 2023). Hence, this study proposes that time pressure serves as a key environmental antecedent to challenge and hindrance technostress, ultimately shaping work outcomes in OLMs. Specifically, we aim to address the following research question.

Time pressure holds significant importance in all organisations, but its implications are particularly pronounced in OLMs. Workers are only paid after the completion of a task, often with no guarantee of acceptance once the task has been successfully completed (Halliday, 2021). Stress due to time pressure and technology use tends to escalate as workers receive no formal support with limited control over their work processes, unlike in traditional settings (Arnoldi et al., 2021). One of the defining features of OLMs is the use of algorithmic control to manage work operations such as task execution, coordination and a rating system, which creates a sense of constant surveillance (Duggan et al., 2019). In particular, OLMs create a demand for speed and acceleration (Zheng and Wu, 2022), fostering work intensification and increased competition, influencing worker well-being. In terms of OLMs, information systems (IS) literature has primarily focused on algorithmic management, reputation systems, bidding mechanisms and production of speed (Kokkodis, 2023; Möhlmann et al., 2020; Zheng and Wu, 2022).

Technostress in diverse contexts has shown that the condition is tied to the nature of work and that new forms of work create new sources of technostress (Cram et al., 2022). Therefore, we focus on the unique context of OLMs. From an IS perspective, there is limited knowledge about how time pressure relates to technostressors. To investigate this issue, we use the traditional transactional theory of stress, which frames stress as a dynamic process between an individual and their environment, whose interaction can create either a threat or a challenge (Tarafdar et al., 2019). This overall transactional process involves the assessment of antecedents, technostressors and outcomes. While early literature focused on technology characteristics as a source of technostress (Ayyagari et al., 2011), recent studies highlight job characteristics (Suh and Lee, 2017; Umair et al., 2023), obligation (Raluca and Hanne, 2020), emotions (Califf and Springer, 2022) and mindfulness (Pflügner et al., 2021) as technostress creators. Building on this, our study extends the existing technostress research from a temporal perspective. In doing so, this research introduces time pressure as a key environmental antecedent in OLMs. In terms of technostressors, this study incorporates both challenge technostressors (creating positive impact) and traditionally developed hindrance technostressors (creating negative impact). This aligns with the recent scholarly calls to further explore challenge technostressors to gain a more nuanced understanding of the broader technostress phenomenon (Tarafdar et al., 2019; Cram et al., 2022). Therefore, this study empirically shows that time pressure predicts hindrance technostress and also challenge technostress, which has positive connotations. Furthermore, this study investigates unique outcomes of technostress, including time distortion and workers' perceptions of meaningful work in OLMs.

Overall, the study theorises and validates the role of temporal issues as antecedents and outcomes of the technostress process, a perspective that has not been previously considered in prior technostress studies. The research contributes to the debate on working conditions of OLMs by highlighting the role of time pressure in shaping workers' experiences and outcomes. By focusing on time pressure and technostress as key factors, the research offers new insights into how platform work can be improved. The findings of the study help OLM workers and other stakeholders, such as platform owners, requesters and regulators, develop strategies to mitigate technostress, making a practical contribution.

Technostress has emerged as an important research area in IS (Cram et al., 2022; Salo et al., 2019). The term was introduced by Brod (1984) and defined as stress experienced due to the use of IT (Tarafdar et al., 2019). Traditionally, the technostress phenomenon is grounded in the transactional theory of stress, which suggests that stress occurs when a person perceives that the demands from their environment exceed their available resources, thus posing a threat to their well-being (Lazarus and Folkman, 1987). In IS literature, technostress is presented as a multi-step process consisting of environmental antecedents, technostressors, appraisal, coping and outcomes (Califf et al., 2020).

Environmental antecedents are workplace conditions that have the potential to create a stressful situation, e.g. technology demands, role demands, task demands, job conditions or policies (Califf et al., 2020). Ayyagari et al. (2011) identified technology characteristics as primary environmental antecedents. However, more recent work on the topic has highlighted other characteristics as environmental antecedents, such as organisational climate, job characteristics, IT mindfulness and emotions, among others (Fischer and Riedl, 2022; Suh and Lee, 2017; Umair et al., 2023). In the context of OLMs, algorithmic control (Cram et al., 2022), IT complexity and feedback (Umair et al., 2023) have emerged as key environmental antecedents.

Technostressors are stimuli or demands appraised by individuals as useful or stressful (Califf et al., 2020). Tarafdar (2007) identified five key technostressors in an organisational context, namely overload, invasion, insecurity, uncertainty and complexity. Ayyagari et al. (2011) expanded these technostressors and included role ambiguity and work-home conflict. In a non-organisational context, life comparison discrepancy, online discussion conflicts and privacy and security uncontrollability have been identified as pertinent technostressors (Salo et al., 2019). It is important to note that technostressors vary according to the context in which an individual works, and an individual can appraise the same stressor differently (Califf et al., 2020).

Once individuals experience technostressors, they evaluate the situation and its significance for their well-being through a secondary appraisal and coping mechanism. The technostress process appears linear, but it is complex and recursive, with an individual moving back and forth through the stages several times (Cram et al., 2022; Lazarus and Folkman, 1987; Tarafdar et al., 2019). Thus, it is common practice in research to focus on specific elements of the transaction-based model of stress to simplify analysis (Cram et al., 2022). Technostressors are further associated with adverse outcomes. The short-term effects of technostressors lead to exhaustion and strain, while the long-term effects include decreased well-being and burnout (Lazarus and Folkman, 1987).

Early IS research investigating the phenomenon of technostress has focused mostly on negative technostressors in the workplace, with unintended consequences for workers and their well-being (Benlian, 2020). However, more recent research highlights that not all technostressors are detrimental to the individual; some have the potential to stimulate and encourage individuals positively (Tarafdar et al., 2019, 2024). Drawing on the transactional theory of stress, Cavanaugh et al. (2000) developed the challenge–hindrance stressor model, which sets the basis for understanding challenge and hindrance technostressors. Challenge technostressors are defined as work demands or situations that, although stressful, offer potential benefit for individuals' goal attainment. For example, task complexity provides an opportunity to upgrade skills and competence. Hindrance technostressors are defined as work demands or situations that do not offer any potential benefit but rather restrict or impede an individual's goal attainment. For example, invasion implies monitoring of employees, which can generate fears of job insecurity (Tarafdar et al., 2019). In the technostress context, researchers have recently focused on technostressors that create challenges (Benlian, 2020; Califf et al., 2020; Tarafdar et al., 2024). However, there is a lack of empirical evidence, indicating potential for further exploration in this area (Cram et al., 2022).

This study particularly focuses on time pressure as an environmental antecedent that can initiate both challenge and hindrance technostressors in OLMs. Time pressure is defined as “subjective perceptions that the time to complete one's work is not enough or that in order to complete one's work one needs to work faster than usual” (Stiglbauer, 2018, p. 64). Time pressure induces stress and creates a need to cope with the limited time (Dóci et al., 2020). Time pressure is widely recognised as a significant job demand in work settings (Bakker and Demerouti, 2007; Bunjak et al., 2023; Schilbach et al., 2023). Previous research on organisational stress shows that time pressure can cause strain (Doef et al., 2000; Karasek, 1979; Silla and Gamero, 2014). Time pressure consists of both quantitative demands (e.g. aspects of work related to speed and amount of work) and qualitative demands (e.g. aspects of work related to effort, such as task complexity, responsibility and hassles), and both types of demands co-exist (Ohly and Fritz, 1998; Schilbach et al., 2023). With the rise of new, flexible forms of work, technology enables work to engulf personal life, leading to a pervasive sense of time pressure in contemporary work environments (Dóci et al., 2020).

Time pressure is central to the concept of technostress, as job characteristics that combine high levels of job responsibility with time pressure may propel individuals to either turn toward or away from IS (Tarafdar et al., 2019; Umair et al., 2023). Yener et al. (2021) found that time management strategies play a moderating role in minimising technostress. In this study, we conceptualise time pressure playing a dual role, acting as a potential catalyst for both positive and negative effects of technostress in OLMs. Previous organisational stress literature confirms that time pressure has an ambivalent nature (LePine, 2022; Schilbach et al., 2023). When there is no adequate time pressure, it can lead to a lack of interest in the task and feelings of monotony (Schmitt et al., 2015) or too much time pressure at work can enhance the threat stressor overload, resulting in adverse effects on performance (Galluch et al., 2015). On the other hand, the presence of time pressure may also stimulate experiences related to challenges such as mastery, engagement, success and personal growth (Benlian, 2020).

OLMs represent a unique case for exploring the relationship between time pressure and technostress, as the economic risks and the responsibility for skill development shift toward workers on these platforms. This creates a time-sensitive work environment, where workers with different levels of experience possess different bargaining power and are expected to have divergent outcomes (Wood et al., 2019). Additionally, the impact of time pressure on technostressors can be influenced by workers' socio-economic conditions, particularly precarity and dependency on platform income (Bajwa et al., 2018; Keith et al., 2019). For workers in financially vulnerable positions, time pressure can intensify technostress. For example, workers may use the platform more often, must meet tight deadlines and strive to maintain high ratings.

Workers have the flexibility to choose tasks, working hours and place of work. However, this flexibility comes at the cost of longer working hours and significant insecurity, resulting in high physical and mental stress (Anwar and Graham, 2020; Standing, 2018). Therefore, dependency on the platform limits workers' ability to decline work even if they want to (Umair et al., 2023). Work in OLMs resembles Taylorism, as work is fragmented into short tasks and workers' performance is measured by speed and output. The pressure is further intensified by the use of algorithmic control and continuous surveillance (Duggan et al., 2019). In such contexts, under time pressure, workers engage with platforms differently, leading to varied outcomes that can range from a temporary challenge to a persistent source of psychological strain.

The early literature on technostress has investigated various well-being outcomes. Researchers unanimously agree that technostressors can influence job satisfaction, turnover intention, organisational commitment, performance and productivity (Maier et al., 2015; Ragu-Nathan et al., 2008; Salo et al., 2022; Tarafdar et al., 2007). However, this study explores unique outcomes related to well-being in OLMs since worker well-being in OLMs is a key metric of platform health and sustainability (Ashford et al., 2018). The study explores time distortion, which is defined as “time being out of tune with one's overall experience” (Blom et al., 2021, p. 2), such that individuals perceive stimuli as lasting for varied times as compared to the actual time (Hedger et al., 2017). Time distortion is selected as an outcome measure because it is identified as the “hallmark” characteristic of addictive behaviours, encompassing excessive technology use (Turel and Cavagnaro, 2019). This choice also aligns with time perception theories that illustrate time distortion as a factor that alters the cognitions and behaviours associated with mental simulations, resulting in temporal misperceptions (Tausen, 2022; Turel et al., 2018). Thus, in their subjective experience, people can perceive time distortion either as time being accelerated or decelerated or as time being stagnant (Blom et al., 2021). Time distortion has been explored earlier in online gaming, web navigation, social media addiction and e-shopping contexts that highlight the symptoms can manifest in the form of mood changes, sleep deprivation, limited social functioning and reduced work performance (Pelet et al., 2017; Turel et al., 2018; Turel and Cavagnaro, 2019). Platforms can create compelling emotional and cognitive behaviours that individuals may value and attempt to replicate (Pelet et al., 2017). Time distortion is considered relevant in the OLMs context because literature has shown that excessive time spent in terms of IT use is an important risk factor or sign of Internet addiction and technostress (Salo et al., 2019; Yener et al., 2021). Similarly, platforms offer micropayments, and to earn a decent income, workers perform more and more tasks. Time distortion can be influenced by various factors, including the nature of the tasks, the structure of the online environment or the specific demands and pressures associated with completing tasks on platforms (Pelet et al., 2017). Among knowledge workers, time distortion has been found to result in work inefficiency (Schéele et al., 2019).

Our study also explores how technostressors affect meaningful work as an outcome. Meaningful work is defined as a positive work-related psychological state that reflects the extent to which workers feel their work is useful, valuable and worthwhile (Landells and Albrecht, 2019). Given the negative perceptions of organisational environment in OLMs, including manipulation, undermining employment relationships and low payment (Arnoldi et al., 2021). Research on OLMs has sparked debate over whether work on these platforms is meaningful. In these highly competitive environments, workers often invest significant, irrecoverable effort up front. Due to the absence of formal obligations, enforcing work effort can be a challenge (Ashford et al., 2018; Liu et al., 2023). Moreover, opportunistic behaviours occasionally emerge, such as requesters exploiting or refusing to pay (Ye and Kankanhalli, 2017). These behaviours reduce the effort workers exert on a task. Since a positive work environment not only impacts workers performance but the quality of work as well, an important consideration is how technostressors shape perception of meaningful work in OLMs.

This section conceptualises the research model and establishes relationships among constructs through hypotheses development. As depicted in Figure 1, time pressure is associated with both challenge and hindrance technostressors, which influence various work outcomes in the OLM context.

OLMs present a dynamic multitasking environment in which effective task management is crucial. Besides the time spent executing the task, workers must also spend time searching, organising and managing new tasks, which is often unpaid. Time pressure is a critical factor in OLMs, with each task having a specific duration before expiration. Workers must adapt their multitasking strategy for each task, available on a first-come, first-served basis, with limited time (Lascău et al., 2022). This intensifies pressure on workers to complete as many tasks as quickly as possible. Even after task submission, the final decision often rests with the requester, introducing an element of unpredictability and the possibility of unfair outcomes for the efforts of workers (Umair et al., 2023). Submissions within mere seconds may raise concerns about the potential involvement of automated bots, suggesting a lack of attention to task details. Submitting too late also results in restrictions or task rejection (Pyo and Maxfield, 2021). Along with these stress-inducing scenarios, workers require an array of technologies to streamline work, especially when they need to employ various software tools and scripts to find high-paying tasks (Lehdonvirta, 2018). As OLM workers are completely dependent on IT to perform their work and earn a living, a large portion of the work stress they experience, both good and bad, is likely attributable to this dependence and manifests as technostress.

Time pressure is highly context-dependent, as perceptions of time pressure are shaped by environmental conditions (Kleiner, 2014; Marpuing et al., 2015; Schilbach et al., 2023). Time pressure can influence a situation by focusing attention on specific informational cues. While time pressure may narrow focus, it does not cause individuals to disregard seemingly peripheral cues under intense stress (Chong et al., 2011). The outcome depends on whether the perceived time pressure is motivational. Individuals with limited attentional resources may still actively engage in task-related cognitive activities if motivated to persist. Csikszentmihalyi (1990) describes this in terms of flow experience when the tasks have the right amount of challenge and time allows to perform even tough tasks effortlessly. As challenge technostressors are linked to demanding yet attainable goals, the interplay of these technostressors with time pressure may encourage people to employ more effective strategies to achieve their objectives.

The fundamental concept of the transactional theory of stress revolves around the appraisal process where individuals initially evaluate the environment to appraise technostressors. This evaluation, in turn, shapes their decision to manage technostressors (Lazarus and Folkman, 1987; Tarafdar et al., 2019). Specifically, individuals who perceive technostressors as potentially beneficial tend to adopt proactive strategies or a problem-solving coping style (Pirkkalainen et al., 2019). In this study, we hypothesise that time pressure can have a positive impact on challenge technostressors for OLM workers. Due to their reliance on platform income, these workers perceive time pressure as an opportunity to enhance their skills and performance. Thus, even when facing time constraints, technostressors are expected to motivate workers to effectively handle tools and complete workloads, resulting in a form of “good” stress (Benlian, 2020). For example, workers in OLMs utilise software tools that integrate with platforms to find high-paying human intelligence tasks (HITs) and interact with community forums for upcoming opportunities. While executing their tasks, OLM workers implement strategies such as using dual screens or adopting scripts (e.g. automating the search for work, employing scraping/filtering techniques on MTurk to identify specific HITs by ID and automatically accepting HITs based on specific criteria like minimum pay (Williams et al., 2019). These approaches enable workers to manage their time efficiently, complete tasks in batches, and meet challenging deadlines (Lascău et al., 2022). The use of such tools to manage their time allows workers to view time pressure not merely as a hindrance but as a challenge that can be effectively overcome. Hence, the study proposes that.

Likewise, for OLM workers, time pressure will positively influence hindrance technostressors. OLMs create an environment of speed as workers rely heavily on technology, making it difficult to meet ongoing work requirements. Platform characteristics such as algorithmic control lead to work overload (requiring more effort than a worker can handle) and complexity (struggling to interpret the expectations embedded through algorithms), which amplify technostress (Cram et al., 2022). The ongoing pressure to maintain an approval rating and inconsistent task completion timings further elevate technostress. When an OLM worker is under time pressure to complete a task, and their IT is not performing to their expectations (e.g. slow Internet speeds), their perception of stress due to IT (i.e. technostress) may increase. Workers who perceive technostressors as potentially menacing tend to adopt a passive coping strategy, such as avoidance or discontinuance (Pirkkalainen et al., 2019). Under intense situations, these technostressors act as hindrances that induce “bad” stress by impeding progress towards goals, often due to issues like restricted access to system resources or system failures (Benlian, 2020).

In OLMs, workers not only have to complete their immediate tasks but also need to organise their work and manage upcoming tasks, which can be overwhelming. Moreover, OLMs promote a culture of being on call as the tools notify workers of new tasks at unpredictable times of the day (Lascău et al., 2022). The tools and work practices used for finding HITs can contribute to interruption overload, i.e. a type of distraction stemming from an excessive number and delivery of notifications or alerts (Williams et al., 2019). The irregularity around task availability can intensify cognitive load through frequent task switching, restrict workers from taking breaks and make it difficult to disengage from work to earn a decent income. Workers also face difficulty scheduling their time because of the way requesters set the task. For instance, workers face uncertainty in task execution due to missing codes or broken HIT links. In some cases, a task is misleading, requiring workers to submit a screenshot or video recording at the very end to exploit workers and collect their sensitive information (Xia et al., 2017). Such tasks can be intrusive, leaving workers with limited options, especially after they have already invested considerable time. Additionally, certain tasks may necessitate additional features, such as managing the visual aspects of the HIT browser or engaging in discussions and sharing HITs within the community (Williams et al., 2019). This additional burden, coupled with unclear information and various obstacles, intensifies the experience of hindrance technostressors. Therefore.

The perception of a fixed time duration varies based on the context in which it is embedded. The way individuals assess a stressor can provide valuable insights into the emotions, behaviour and decisions related to an event (Tausen, 2022). Considering time perceptions, Csikszentmihalyi's (1990) flow theory is of direct relevance, which highlights how individuals get absorbed in an activity. When an activity is inherently enjoyable and not so challenging as to cause frustration, individuals are more likely to become fully engaged. Engagement is particularly high when an individual participates in an activity driven by intrinsic motivation, such that the experience feels rewarding and fosters a desire to repeat it. This state of flow typically pushes an individual to continue the activity where loss of self-consciousness and a distorted sense of time emerge (Guo et al., 2016).

A sense of personal connection to the future event plays a key role in how individuals motivate and control themselves. In particular, the perceived length of an event is interconnected with assessments of monetary compensation, and this connection is further tied to the impacts of temporal distortions (Tausen, 2022). The structure of OLMs and its design can shape the perceptions of time, such that the perceived duration of tasks feels altered. Such processing of time can influence cognitions, appraised stressors and task engagement of workers (Yin et al., 2018). The investment of time and effort may fluctuate for workers who are financially dependent on platforms and motivated to maximise their income. This is particularly evident when tasks strike a balance between being challenging and aligning with their skill set (Csikszentmihalyi, 1990). Especially, when individuals feel that their investments of time and energy in ITs can be rewarded. In OLMs, the immersive and rewarding experiences may lead workers to lose track of time. For instance, commencing the day with a seamless flow of work, such as securing highly rewarding tasks or participating in activities aligned with one's interests, the use of complex yet interesting IT tools can make the experience enjoyable. Additionally, receiving prompt responses from requesters in the middle of a task, without the stress induced by task-related challenges, can contribute to positive engagement. In these instances, workers can become fully engrossed in the task, especially when it comes with clear goals. When experiencing such challenge technostressors, workers may find themselves absorbed and inclined to repeat such experiences, creating a subjective perception that time has passed swiftly. Hence, the proposed hypothesis is.

Work meaningfulness refers to individuals considering the importance and value of their work based on their subjective experience (Rosso et al., 2010). Individuals derive the meaning and value of their work based on cues within their work environment (Landells and Albrecht, 2019). In the context of OLMs, task design, user equipment and platform interface play pivotal roles in shaping work experiences (Williams et al., 2019). The interplay between challenging stressors, combined with feedback from requesters, emerges as a critical factor in enhancing the enjoyment derived from the work. This process cultivates a profound feeling of importance among workers, fostering the belief that their contributions truly matter. OLM workers act as owners of their own careers and engage in building their careers through learning self-management and regulation (Xiongtao et al., 2021). In OLMs, such self-regulation behaviours can help workers manage uncertainty and cope with tough setbacks (Ashford et al., 2018). For example, MTurk workers may experience joy and satisfaction from contributing to thought-provoking and creative tasks. Similarly, from participating in survey research that aims to contribute to a broader goal, bringing value to others and themselves. In essence, when tasks and platforms adeptly address workers' inquiries while ensuring fair compensation, the work experience becomes more meaningful. Hence.

Stressful and threatening stimuli can induce time distortion effects, impacting individuals' perception of the duration of these stimuli, leading to either a sense of rapid passage or an experience of prolonged slowness (Hedger et al., 2017). OLMs present a paradoxical work environment, offering flexibility and control while simultaneously constraining task availability through algorithms. When workers encounter monotonous and repetitive tasks, they may perceive time as slowing down. This phenomenon arises from the lack of mental stimulation, creating an illusion of slow-paced time (Hedger et al., 2017; Tausen, 2022). Indeed, cognitive time tends to progress more slowly in stressful situations (e.g. deprivation due to limited resources or monotonous tasks), compared to less stressful situations (Turel et al., 2018). Likewise, when looking back on the past, the context in which a previous event unfolded can also impact the way time is perceived (Tausen, 2022). For instance, a worker who has experienced numerous work rejections may have a different time perception based on those past experiences. In MTurk, the inclusion of attention check items is a common strategy. However, certain tasks are organised in a way that can be frustrating, as requesters often include too many attention check items without a clear description. This lack of clarity can make the worker lose interest in the task and make the time appear to be dragging for the task. Therefore, poorly designed tasks, complex tasks, extensive reading or ambiguous instructions can make time feel like it is dragging on. Even if the actual duration is not excessively long, the perceived effort and lack of clear progress can distort the subjective experience of time. Similarly, the implementation of algorithms restricting task availability may contribute to a perceived stretching of time. The sporadic nature of task availability can lead to periods of waiting, searching for suitable HITs or refreshing the dashboard. These idle or low-engagement periods can make time feel very slow, especially if workers are trying to maintain a consistent income stream. In such instances, workers must exert more effort and actively engage in searching tasks, particularly the ones paying more to counteract the perception of time being wasted. Therefore, hindrance stressors such as role ambiguity and work hassles can lead to a distorted sense of time. Consequently, we expect that.

Research conducted in organisational contexts suggests that individuals' time and resources are valuable, and that workers strive to acquire and protect them. For workers who cannot perceive the meaning in their work, it may be regarded as a form of resource loss (Hobfoll, 2001). The absence of formal obligations within OLMs poses challenges, as gig workers must acquire necessary resources themselves to obtain recognition (Ashford et al., 2018; Petriglieri et al., 2018). Research on the motivation of gig workers has indicated that, beyond financial incentives, gig workers are driven by the desire to enhance and preserve their sense of self (Keith et al., 2019). The interpersonal interaction between gig workers and requesters during service provision plays a crucial role in shaping the meaning of their work. When gig workers encounter negative stressors, such as unfair rejections, it raises doubts about their work, thereby impeding their ability to derive meaning from it (Xiongtao et al., 2021). Characteristics of OLMs, such as fragmented tasks disconnected from the broader work, rigid reviews, reputation mechanisms and the constant need to be online, can influence meaningful work (Nemkova et al., 2019). Therefore.

The data were collected through a survey via MTurk by following the guidelines for conducting behavioural research (Jia et al., 2017; Mason and Suri, 2012). A sample was recruited to perform a survey, which was posted as a human intelligence task (HIT) on MTurk with a clear description. Workers could click on a web link that redirected them to the online survey hosted on the Qualtrics platform. The HIT was made available to workers who have completed a varying number of HITs, from 500 to 10,000 HITs, to increase the chance of participation. The HIT approval rate was set at 98%; generally, setting the HIT approval rate at 95% can produce high-quality data (Hunt and Scheetz, 2019).

The MTurk platform has frequently been used for data collection over the years. However, data quality is a concern when collecting data from the platform. Therefore, appropriate measures were taken to minimise the risk of low-quality data following the guidelines for data collection on MTurk (Aguinis et al., 2021). First, the reCAPTCHA test was used at the start of the survey to identify whether the activity on the screen was produced by a human or a computer program such as a bot. Second, instructed-response items were used to address the issue of spam individuals. Two attention-check items with an obvious correct response were embedded at two different points in the survey. A total of 467 responses were received in Qualtrics. However, after eliminating incomplete responses, respondents who failed the attention check and multiple submission attempts, only 354 responses were suitable for data analysis.

Before conducting the final survey, a preliminary pilot test was executed. Additionally, experienced MTurk workers and two researchers critically reviewed the survey. The feedback and suggestions were incorporated into the final survey. The data collection phase spanned two weeks, including weekends in October 2023. To ensure a diverse participant pool, the survey was administered in multiple batches at various times each day, including morning, evening and night. This strategic timing approach aimed to account for potential variations in the online sample, as highlighted by previous research (Arechar et al., 2017). To determine the compensation rate for the HIT, careful consideration was given to the estimated time required to complete the task and the US minimum wage. Consequently, a compensation of US$2.50 was set for this 15-min HIT, aligning with fair remuneration practices.

The measurement items were adopted from well-established scales from both the IS and organisational psychology literature. The wording of the items was slightly modified to fit the MTurk platform context. For example, employers and tasks were referred to as requesters and HITs, respectively. Time pressure was measured by self-report, using a 5-item subscale of the Instrument for Stress Oriented Task Analysis (ISTA) (Irmer et al., 2019). Challenge and hindrance technostressors were measured by the items given by Benlian (2020). For time pressure and technostressors, a 7-point frequency scale was used ranging from 1 = “never” to 7 = “extremely often”. Time distortion was measured using two items given from Pelet et al. (2017). Meaningful work was measured by May et al. (2004). For time distortion and meaningful work, a 7-point Likert scale was used ranging from 1 = “strongly disagree” to 7 = “strongly agree”. The measures for all constructs are shown in detail in (Table 1).

We controlled for variables within the model that have been shown to potentially influence the hypothesised relationships. Therefore, individuals' gender, age, education level, MTurk platform experience and IT usage were controlled. Given the nature of self-report surveys, there is a concern regarding MTurk workers engaging in self-misrepresentation or displaying socially desirable responses (Kwak et al., 2021). One of the primary reasons for participation in a HIT is monetary compensation, so MTurk workers are more likely to provide socially desirable responses (Aguinis et al., 2021). Therefore, to mitigate this concern, social desirability was controlled and measured as a true and false scale given by Reynolds (1982). Also, during data collection, the authors ensured respondents' confidentiality and anonymity, thereby encouraging them to provide genuine responses. Overall, controlling these variables enhances the robustness of the study by accounting for potential confounding factors and ensuring the reliability of the findings.

A descriptive analysis was conducted to extract the demographic profile of MTurk workers. The sample included 62.43% men and 37.57% women. The majority of the workers who completed the survey were from the US (92.94%), followed by 7% of workers from India and other countries, including South Africa, Italy, the UK and Brazil. The workforce on these platforms is predominantly US-based, a trend consistently documented in the literature, although the representation of international workers has notably increased over time (Difallah et al., 2018). Our participant demographics indicate a relatively young to middle-aged population, with 29.4% aged between 25 and 30 years and 44.4% aged between 31 and 40 years. Furthermore, most participants are highly educated, with 68.4% holding a bachelor's degree and 17.5% a master's degree. Table 2 shows the demographic profile of participants.

Structural equation modelling (SEM) using partial least squares (PLS) was used to test the proposed research model with the help of SmartPLS software version 4.0 (Ringle et al., 2015). The SEM analysis was conducted following a two-stage analytical procedure (Hair et al., 2017). The first stage assessed the measurement model for reliability and validity. The second stage evaluated the structural model to test the research hypotheses.

Reliability, convergent validity and discriminant validity were tested to validate the measurement model. All constructs in the model were reflective; the procedures recommended by Hair et al. (2019) were followed. The first step in reflective measurement model assessment involves examining the indicator loadings, recommending a threshold value of 0.7. All outer loadings were above the 0.7 value, indicating that the construct explains more than 50% of the indicator's variance; thus, item reliability is acceptable. Table 3 show item loadings and descriptive statistics of each item. As the second step, composite reliability (CR) was used to measure reliability (CR > 0.7). The results indicate that the CR for all the variables is greater than 0.7. In addition, Cronbach's alpha α was used to check the internal consistency reliability (α > 0.7). The results indicate α above the threshold value for all the variables, indicating good internal consistency. The third step involved measuring the convergent validity of each construct measure. For this purpose, average variance extracted (AVE) was used to verify each construct, which in this case was greater than 0.5 for all variables. In sum, the model's convergent validity could be established. Table 4 shows construct reliability and validity.

The fourth step was to assess the discriminant validity, which is the extent to which a construct is empirically distinct from other constructs in the model. Traditionally used criteria, as proposed by Fornell and Larcker (1981), are to measure the square root of AVE of each latent variable, which should be higher than the correlations among the latent variables. In Table 5, Fornell–Larcker criterion off-diagonal values represent the correlation coefficients between potential constructs and the diagonal values represent the square root of the AVE value of each construct. Cross-loading factors are also used to test a reflective measurement model's discriminant validity. Table 6 reports the cross-loading factors among the measured items, indicating that each indicator loading is higher than its cross-loadings. In addition, the heterotrait–monotrait ratio (HTMT) of the correlations was utilised to assess further discriminant validity, which was within the recommended threshold of 0.90 (Henseler et al., 2015) reported in Table 7.

In summary, the results indicate that the data pass the reliability and validity evaluation. All the measured variables were tested for common method bias (CMB) since the data were self-reported and collected from a single source. To mitigate potential risks, various procedural measures were implemented in the development of the survey instrument (Podsakoff et al., 2003). These included keeping questions simple and concise, avoiding double-barrelled questions. When designing the questionnaire, the order of items was counter-balanced by keeping the dependent and independent variables separate and reiterating respondent anonymity along with the exclusive research purpose of this study. Statistically, the extent of CMB was assessed using two approaches. First, Harman's one-factor test, which was performed by including all items in principal components factor analysis. Evidence for CMB exists when one factor accounts for most of the covariance (i.e. explained variance >50%). The specified “one factor” explained 23% variance, and hence the data do not indicate evidence of CMB. Second, CMB was tested based on variance inflation factors (VIF) and a full collinearity test, which is considered an effective test for the identification of CMB, as suggested by Kock (2015). The results indicate values lower than 3; thus, the model is considered free of CMB.

To test the structural model, the path coefficients, the t-values of the variables and their statistical significance (p-values) were assessed using the bootstrap re-sampling method with 5,000 resamples. The results indicate that time pressure was significantly positively associated with both challenge and hindrance technostressors, supporting H1a (β = 0.74; p-value <0.001) and H1b (β = 0.70; p-value <0.001). In terms of outcomes, challenge technostressors were significantly positively associated with time distortion, supporting H2a (β = 0.24; p-value <0.01) and meaningful work H2b (β = 0.32; p-value <0.01). Hindrance technostressors were significantly positively associated with time distortion (β = 0.22; p-value <0.05), supporting hypothesis 3a. However, hindrance technostressors showed no significant relationship with meaningful work, so H3b was not supported. A summary of results along with path coefficients, t-values and p-values is provided in Table 8.

The commonly used measure to evaluate the structural model is the coefficient of determination R² value, which explains the model's predictive power (Hair et al., 2019). In terms of the model's predictive power (R²), in this model, time pressure explains 54% of the variance in challenge technostressors and 68% of the variance in hindrance technostressors. Both challenge and hindrance technostressors explain a variance of 27% in time distortion and 16% of the variance in meaningful work. We also examined Q²_prediction to assess predictive relevance for all endogenous constructs in our model. The Q² _prediction values for challenge and hindrance technostressors were 0.52 and 0.63, respectively. While Q² _prediction values for time distortion and meaningful work were 0.24 and 0.10, all greater than 0, thus validating the explanatory and predictive capabilities of the model.

The effects of six control variables were examined, including gender, age, education level, MTurk platform experience, IT usage and social desirability on the four endogenous constructs: challenge technostressors, hindrance technostressors, time distortion and meaningful work. Age and education level and MTurk platform experience had no significant effect on any of the endogenous constructs. Gender was significantly correlated to hindrance technostressors. IT usage was significantly correlated to meaningful work. Social desirability was significantly correlated with hindrance technostressors and time distortion.

The non-significant relationship between hindrance technostress and meaningful work was unexpected. Thus, we conducted a post-hoc analysis to try to explain this finding further. Drawing from the job demands-resources model (Bakker and Demerouti, 2007), one possible explanation is that the experience of challenge technostress buffers the negative effect of hindrance technostress on meaningful work. This interaction effect is depicted in Figure 2. A further analysis using the Hayes Process Macro in SPSS (see Table 9) reveals that the negative relationship between hindrance technostress and meaningful work is only significant when the experience of challenge technostress is also low (i.e. −1 SD). For OLM workers reporting higher challenge technostress (i.e. +1 SD or at least average), the negative relationship becomes insignificant. Thus, our post-hoc assumption is confirmed. Higher levels of challenge technostress buffer the negative effect of hindrance technostress on meaningful work.

This study examined antecedents and consequences of technostress from a time perspective in OLMs. Specifically, a research model was proposed to investigate how time pressure influences technostress. Time pressure, as an antecedent, influences technostress in both positive and negative ways. We explain this mechanism from two perspectives. First, time pressure is tied to contextual conditions, such as types of workers, use of platforms and motivations for engaging in OLMs. Depending on these circumstances, workers choose either push (external factors that influence one's action) or pull factors (internal factors that influence one's action) to accomplish their work (Keith et al., 2019). Second, time can play a key role in shaping worker's behaviour through three different ways: time as a resource (a precious asset), time as a structure (creating temporal order) and time as a process (connecting the past, present and future) (Blagoev et al., 2024). Workers try to acquire and apply these ways needed to sustain the behaviours (e.g. time, mood, activities) necessary for their goal. Research suggests that having a moderate level of time pressure in a job is important, as a complete absence of time pressure is associated with lower well-being and job dissatisfaction (Doef et al., 2000).

The findings indicate the impact of time pressure on both challenge and hindrance technostressors. Time pressure positively influences challenge technostressors, offering workers a chance to enhance skills and improve performance. For example, MTurk workers may learn a new script out of necessity to meet their daily earning goals, establish a work rhythm between active and passive tasks to optimise their work and apply past experience to make informed decisions. Time pressure also positively influences hindrance technostressors, making it a personal resource that requires proper allocation. For example, an MTurk worker may need to hurry to submit a task before it expires, automatic dashboard log-outs resulting in time loss, and overestimating their time by holding more tasks beyond their capacity, i.e. batch completion.

The results indicate that challenge technostressors can give rise to time distortion effects in OLMs. When individuals encounter challenges, their actions require minimal conscious control as they naturally focus their attention on specific stimuli (Hedger et al., 2017). This phenomenon can occur for workers when they engage in fulfilling tasks, becoming so engrossed that they lose track of time. Intrinsic motivation of workers plays a crucial role in this context. Those who are inherently motivated need no external rewards or goals to motivate them, since the task itself provides engagement and workers experience time distortion. Similarly, technostressors that enable workers to achieve their goals can produce a sense of joy and positive emotional experience, making their work feel more meaningful. On the contrary, when workers encounter poorly designed tasks that require extra effort to complete, they may perceive time as slowing down. This phenomenon arises from the lack of mental stimulation, creating an illusion of slow-paced time (Hedger et al., 2017). Individuals who are more extrinsically motivated by the reward of tasks may experience a distorted sense of time. In OLMs, workers feel a sense of wasted time, stress and missed opportunities to catch new tasks. However, hindrance technostressors had no impact on meaningful work. Unlike remote environments, where hindrance stress elicits emotional responses of frustration, which create perceptions of less meaningful work (Aleksić et al., 2023). In OLMs, workers may associate meaningful work with broader aspects of the platform. For example, even if they complete a poorly designed task, they may still value the impact it will have on their overall performance rating. Thus, the impact of hindrance technostressors on meaningful work is minimal due to the specific context and the flexibility offered by OLMs.

This study contributes to the literature in three significant ways, guided by Whetten's (1989) perspective on theory development. We articulate our theoretical contributions by addressing the how, why and what logic consistent with Whetten's position.

First, a contribution should demonstrate “how” the new insights affect the accepted relationships between the variables. Our research advances the IS domain by empirically showing that time pressure influences the challenge and hindrance technostressors as an input. The findings further highlight the effects of technostressors in creating time distortion effects as an output. This study helps recast contemporary technostress thinking to foreground the critical role of temporal issues, both as an input to and output of IT-induced stress. Therefore, this research presents time as a key facet, shedding light on the nuanced relationship between temporal issues and technostress, which is underappreciated in technostress literature. To date, previous research on technostress has mostly considered time from a methodological perspective (Venkatesh et al., 2021). Also, this research sheds light on the positive effects of technostress, which has gained attention more recently. While the focus here is on OLMs, where the breakdown of work is into micro tasks, the findings of this study are equally relevant to other IT-driven work environments.

Second, “why” theoretical contributions emerge when a study's findings challenge underlying principles and bring consensus to the field. The existing literature on time pressure presents inconsistent results, with arguments suggesting a positive, negative or inverted U-shaped relationship with outcomes (Schilbach et al., 2023; Schmitt et al., 2015; Widmer et al., 2012). Our findings provide some consensus on the understanding of time pressure, demonstrating that it can have both positive and negative effects. This research further enhances our understanding of time distortion effects stemming from positive stimuli, an area typically overshadowed by the prevalent emphasis on distortions caused by negative stimuli. Consequently, the findings in this study provide a foundation for deeper investigation into the role of positive stimuli such as challenging technostressors and time distortion.

Our post-hoc analysis adds to the technostress literature by demonstrating that the experience of challenge technostress buffers the negative effect of hindrance technostressors on meaningful work. In other words, we can say that hindrance technostressors are only negative if they are not matched by challenge technostress. Consistent with the Job-Demands Resources Model (Bakker and Demerouti, 2007), challenge technostressors can be considered a personal resource that protects against the unwelcome effects of job demands, such as hindrance technostressors. Existing technostress research has only considered the direct positive effects of “good” stress (Benlian, 2020; Califf et al., 2020; Tarafdar et al., 2024). Our study opens a new avenue of enquiry in technostress research by revealing an indirect benefit of challenge technostress, namely its ability to offset the negative effects of hindrance technostress. While hindrance technostressors may be unavoidable in many work contexts, our findings suggest that their deleterious consequences can be dampened if work IT systems are designed to facilitate challenge technostressors. In this way, our research challenges the underlying assumptions that hindrance technostress should be reduced. Our study suggests the focus should switch to enhancing challenge technostress.

Third, a “when” theoretical contribution explains when something new is learned about the pre-existing model or theory itself when it is applied to a new situation (Whetten, 1989). We use the transactional theory of stress as a theoretical lens to interpret technostress, which frames stress as a dynamic interaction between an individual and their environment. Our study focuses on OLMs, a unique environment where time and speed are critical. The efficient completion of tasks is tied to income, higher performance ratings and the establishment of trust with requesters. The competitive environment compels workers to appraise time pressure either as a challenge or a hindrance. Workers who take control of their time can manage uncertainty and cope well with setbacks (Ashford et al., 2018). In this way, time pressure allows workers to manoeuvre and seek opportunities to enhance their work practices. Conversely, time pressure poses a risk, as it can drain energy, given the inherent value of time as a crucial resource that must be utilised carefully. The work structure of OLMs further exacerbates technostress, as algorithms are heavily used in managing work. Such algorithms subject workers to continuous monitoring through detailed digital metrics and statistical analyses, linking pay to performance. This mode of control, commonly referred to as digital Taylorism, has profound implications (Duggan et al., 2019). It reinforces power asymmetry between platforms and workers, where those in more vulnerable positions struggle to voice concerns or influence decision-making, often leading to perceptions of unfair treatment and diminished agency.

Our study reveals that time distortion is a unique outcome of technostress. The structure of OLMs and their design can shape the perceptions of time, such that the perceived duration of tasks feels altered, which in turn can influence cognitions, appraised stressors and task engagement of workers (Yin et al., 2018). Workers who are financially dependent on platform tasks and motivated to maximise their income may invest time and effort, but the effort they put in may fluctuate. If a task aligns with workers' skill set and resources, they enjoy working on it for more extended periods, losing track of time and creating a distorted sense of time. The findings suggest that hindrance technostressors also contribute to the perception of distorted time within OLMs. In OLMs, the platform itself offers limited resources, prompting workers to draw on online forums and their own experiences (Petriglieri et al., 2018). Hindrance technostressors, such as algorithmic surveillance and poorly designed user interfaces, can induce a time distortion effect, creating the perception that time is passing either slowly or quickly.

Interestingly, hindrance technostressors demonstrated no significant association with meaningful work. These findings can be explained by two reasons. First, OLM workers often struggle with some aspects of their work, such as unfair rejections and limited control over their work. However, they tend to invest considerable time in accessing valuable resources through activities like participating in informal forums. Through these activities, they may derive a sense of meaning from their work, contributing to the enhancement of the broader worker community. Second, the heterogeneity of OLM workers across platforms plays a significant role in shaping their experiences of meaningful work. While early studies recognise the diverse motivations and circumstances that drive participation in OLMs, more recent research highlights how workers' personal circumstances influence their experiences within these platforms (Keith et al., 2019; Mayer et al., 2024). This is particularly important for workers who are excluded from traditional forms of employment due to specific needs, e.g. disability and caregiving responsibilities. For such workers, the financial compensation they receive for completing tasks helps reinforce the belief that their work is both valuable and meaningful to them, their families and society. In contrast, such reflections are often absent in productive but unpaid activities (Mayer et al., 2024). Therefore, despite the difficulties in these OLMs, workers view their work as meaningful. This study contributes to the ongoing discussion concerning the working conditions of OLMs by examining the influence of time pressure on workers' experiences. Overall, our study contributes to understanding how OLMs are perceived as a source of meaningful work, a narrative emerging in research by framing OLMs as long-term employment.

This study offers numerous practical implications for platforms, requesters, workers as well as national and international regulators of OLM work. Time pressure emerges as a crucial factor that can either positively or negatively impact technostressors, making it imperative for each stakeholder to play their role effectively.

Platforms should invest in measures that help to optimise time and effort for all. Several initiatives can be taken for this purpose: first, designing a user-friendly interface that offers more feature flexibility can save time. Second, setting a limit on the number of tasks a worker can complete in a single day. Third, enforcing a break after certain hours without any penalty. Given that technostress and time distortion are influenced by various factors, including an individual's biology, psychology and environment (Schéele et al., 2019; Tarafdar et al., 2019), platforms may benefit from designing specific training programs. For example, making it compulsory for new workers to learn interactive tools and scripts or offering a simulated environment where the first twenty tasks are completed without a HIT expiration timer or the risk of rejection. This training can offer ways to efficiently manage temporal resources, adapting to variations in work patterns to achieve more consistent, streamlined work practices. Platforms must focus on active feedback design. For example, providing a data-driven visualisation of their work patterns, such as a comparison between the worker's estimated time and the actual time spent on the tasks. These strategies can raise awareness among workers of their individual time distortion tendencies and equip them with skills to reflect on and mitigate these distortions and technostress. Additionally, platforms should establish policies that enforce accountability and ethical conduct. For instance, removing tasks that do not match their task descriptions and allocated time. By implementing such measures, platforms can foster a more conducive and ethical working environment, benefiting all stakeholders.

Requesters play a critical role in task allocation, and their practices in this regard can significantly shape the way workers perceive technostressors. It is crucial for requesters to set appropriate tasks and their time. Tasks should be designed responsibly, with sufficient elements of complexity and creativity rather than undue burdens. To address time-sensitive challenges, requesters should offer adequate resources, such as greater control, to enable review and edit submissions. Additionally, supporting workers by being accessible during task execution. Requesters can alleviate time pressure by cultivating a sense of safety through trust-based relationships with workers, rather than resorting to punitive measures. These strategies can maintain worker loyalty, foster a collaborative environment, minimise unethical behaviours and contribute to positive work experience. Platform workers can contribute significantly by being aware of exploitative practices and refusing to participate in them, even if the task offers high pay. They can support their peers by exposing such practices through community forums. Despite challenges, workers must establish their own boundaries regarding when to stop working and practice self-regulation to reduce time pressure. Finally, it is imperative that OLM regulators take a key role by conducting thorough checks on platforms that neglect fair work practices. Regulations concerning platform-based work should account for the unique labour market dynamics and institutional frameworks in each country. Efforts are required to narrow the gap by adopting standardised work practices and collaborative governance structures across administrative levels. Despite the challenges, legislation must be enacted to establish rules limiting working hours and mandating rest periods, including weekly breaks and paid annual leave.

While this study contributes valuable insights, it is essential to acknowledge its limitations, which offer avenues for future research. First, the research is cross-sectional, and the data are collected at a single point in time from one platform via a self-reported survey. Therefore, the generalisability of the findings to other platforms remains uncertain and should be approached with caution. Ideally, combining collaborative research methods with surveys via platforms that offer in-depth interview approaches would enhance the credibility of the results. As the constructs used in the research model are time-sensitive and subject to variability, it is important to acknowledge the potential for future changes in terms of work contexts and datasets. Future research should employ a longitudinal study design to provide better inquiry into the proposed relationships and inferences about causality by considering other OLM platforms such as Prolific, Upwork, and Fiverr. To complement the current findings, future research can employ traditional regression analysis for a more focused examination of the strength and direction of these specific relationships.

Second, appropriate measures were used, and data were collected at different intervals throughout the day. However, the results indicate that most participants are US-based workers, which may further limit the generalisability to all platform workers. As a result, potential researchers should collect data from workers in other countries to confirm whether the outcomes remain valid. Moreover, respondents who choose to participate in the study may introduce self-selection bias, undermining the ability to select a random sample from the target population. In MTurk, workers' decisions are based on personal and demographic characteristics, such as monetary incentives and employment status. To create more representative samples, researchers can collect data on MTurk using stratified sampling to capture the nuances about individuals who offer their services on these platforms.

Third, the study focuses on identifying time pressure as an antecedent of technostress and its related outcomes. The model is tailored to the “front end of the transactional theory,” emphasising a specific aspect of the multifaceted technostress process outlined by Cram et al. (2022). Future research should expand the model using secondary appraisal and coping processes, moderating variables (e.g. worker resilience) and other variables (time management behaviour) to further develop the technostress phenomenon. In addition, transactional theory is a relevant framework for studying technostress. However, to gain a multidisciplinary perspective, future researchers can integrate literature beyond IS, such as sociology or labour process theory.

Fourth, while this study draws attention to temporality issues in technostress, it focuses solely on the concept of time pressure. This is logical as it is a predominant issue in OLMs and conceptually suitable. However, time is an inherently complex, multi-faceted, subtle concept and is by nature socially embedded (Blagoev et al., 2024; Kunisch et al., 2021) in information systems (Kishore et al., 2024; Venkatesh et al., 2021). Along with the time pressure of tasks, technostress may be driven by the rhythm of tasks and interruptions caused by the OLM. Technostress can be influenced by a multitude of other temporal factors, such as temporal personalities or temporal perceptions. We encourage other researchers to examine these issues in an OLM technostress context.

Finally, this study did not include worker income as a control variable, which could potentially affect the relationships. Therefore, future work must take this variable into account. Moreover, this study has explored the role of time, which is an intricate concept in itself. Future studies could examine time-related variables and delve into more detailed components that play an essential role in shaping workers motivations and behaviours, e.g. the sensitivity of time to culture could be explored, allowing future research to test and refine the findings of this study in diverse cultural contexts and how culture shapes the technostress process. Moreover, examining the role of temporal personality and its potential in mitigating technostress could enhance our understanding of technostress.

OLMs are growing, and more workers will likely engage with these platforms in the future. The aim of this research is not to critique these environments, but rather to highlight the importance of designing them to support worker well-being. IT will remain an integral part of OLMs; however, platform design must take into account the psychological and cognitive demands placed on workers through its use, particularly those related to technostress. Our findings contribute to both the OLMs and technostress streams by offering new insights into how temporal aspects such as time pressure and time distortions interact with digital work environments to shape workers' experiences. Specifically, we demonstrate that time pressure can influence outcomes in different ways. While it can lead to negative forms of technostress that may undermine workers' sense of autonomy, it can also enhance focus and productivity, which allows meaningful engagement with their work. By foregrounding the temporal perspective of technostress, this study highlights the need for platform designers, users, policymakers and researchers to consider how time-related features such as deadlines, task availability, work pace and algorithms can be optimised to reduce cognitive strain and promote sustainable work practices.

The research has been approved by the ethical research committee of the University of Galway (Ref-19 Dec 30).