A longitudinal study on ICT workload in the extended stressor-detachment model: testing moderated mediation models for extended work availability and workplace telepressure

ABSTRACT Information and communication technologies (ICTs) are seen as essential tools for mastering knowledge work in the twenty-first century. However, ICTs do not solely improve workflows, but are experienced by employees as an additional demand described as ICT workload. In this study, we apply the stressor-detachment model in the context of ICT, and investigated relations of ICT workload to psychological detachment and emotional exhaustion. Furthermore, as extended work availability and workplace telepressure are organisational and individual factors associated with ICT use, we extend the stressor-detachment model by testing moderated mediational effects. Specifically, we investigated whether extended work availability and workplace telepressure strengthen the indirect effect of ICT workload on emotional exhaustion via psychological detachment. In a longitudinal study of 228 employees with three measurement points, ICT workload was positively related to emotional exhaustion over time. Additionally, we found tentative support for a mediation via psychological detachment, that is stronger for higher levels of workplace telepressure. In sum, our study shows that employees experiencing high levels of ICT workload have a higher risk of suffering from emotional exhaustion. An impaired ability to psychologically detach and high levels of workplace telepressure play a crucial role in this relationship.


Introduction
Digitised processes and mobile information and communication technologies (ICTs) have become an integral part of organisations and many occupations.Thus, technological changes modify work organisation and working conditions, causing an extension of the regular workday, compelling employees to work faster (Ragu-Nathan et al., 2008) and leading to increased workload (Chesley, 2010).Comparing the results of an employment survey on the work and health situation in Germany over three years, researchers showed that newly introduced computer programmes and process technologies are related to increased work intensity (Meyer et al., 2019).Furthermore, work-related use of e-mail and instant messengers is growing, and increased ICT use seems to be causing e-mail overload (Brown et al., 2014;Huang & Zhang, 2019).In sum, employees are likely to perceive an increase in the amount of work and, thereby, experience additional workload due to ICTs.
ICT workload is a specific form of the well-established job stressor workload.It is defined as the perceived additional amount of work attributed to technologies such as smartphones and the internet (Day et al., 2012).For instance, employees report working longer hours in and out of the office because of technology.In the present study, we build on previous research that has shown the detrimental effects of ICT workload on well-being in terms of psychological detachment and emotional exhaustion (Barber & Santuzzi, 2015;Day et al., 2012).Applying the stressor-detachment model (Sonnentag & Fritz, 2015) to the context of work-related changes due to ICT use, we propose that psychological detachment acts as a mediator for the relationship between ICT workload and emotional exhaustion.Additionally, based on boundary theory (Ashforth et al., 2000), we propose that the effects of ICT workload depend on organisational (extended work availability) and individual (workplace telepressure) factors, which amplify the relationship between ICT workload and psychological detachment, resulting in a moderated mediation effect.Figure 1 depicts our conceptual model.
Our study contributes to previous research in three ways.First, our study aims to test the stressor-detachment model (Sonnentag & Fritz, 2015) within the context of workrelated changes caused by ICT use.As ICTs are an indispensable part of today's work, it is important to replicate well-known stressor-strain relationships in this novel context.Thus, we focus on ICT workload as a specific and evolved form of workload to assess the effects of novel ICT-related stressors on employees' well-being.Second, our study thoroughly investigates how and under which circumstances ICT workload relates to emotional exhaustion.That is, we do not only put the stressor-detachment model into the context of ICT use, but also extend the model by analysing whether other ICT-related factors (e.g.extended work availability, workplace telepressure) act as moderators for the proposed indirect relationship between ICT workload and emotional exhaustion via psychological detachment.Third, previous research on the relationship between ICT workload and well-being outcomes is solely based on cross-sectional research designs (Barber & Santuzzi, 2015;Day et al., 2012).In contrast, our study has a full-panel design with three measurement points (all variables measured at each wave), and thus allows us to draw conclusions on how ICT workload is related to emotional exhaustion over time.By examining the interplay between these constructs, we contribute to the planning and implementation of future interventions to prevent the negative effects of ICT workload.

ICT workload and emotional exhaustion
ICTs change working life, for instance, employees report an increased amount of communication due to ICTs, requiring them to handle a large number of messages and maintain multiple communication channels, thus increasing their workload (Bordi et al., 2018).As incoming calls and messages cannot be foreseen, employees also experience more interruptions and unpredictability due to work-related ICT use (ter Hoeven et al., 2016).In sum, ICTs are likely to increase the amount of work that has to be managed within a restricted time span.In line with the effort-recovery model (Meijman & Mulder, 1998), we argue that the continued exposure to ICT workload depletes resources, resulting in strain reactions such as emotional exhaustion, a chronic state of physical and emotional depletion (Maslach et al., 2001).We assume that ICT workload increases emotional exhaustion over time because employees invest effort to manage their ICT workload (e.g. by working longer hours), suffer from the control-reducing aspects of ICTs (e.g.increased communication, unpredictability of work), and feel overwhelmed by the amount of work they must accomplish.Previous cross-sectional studies have shown that ICT workload is associated with higher levels of emotional exhaustion (Barber & Santuzzi, 2015;Day et al., 2012).Accordingly, we hypothesise: Hypothesis 1: ICT workload is positively related to emotional exhaustion over time.

Psychological detachment as a mediator for the relationship between ICT workload and emotional exhaustion
The stressor-detachment model (Sonnentag & Fritz, 2015) suggests that facing high levels of job stressors increases individuals' negative activation (e.g.promoting rumination about the job stressor), which in turn hinders the ability to mentally detach from work during non-work time.Eventually, a lack of psychological detachment increases strain reactions such as emotional exhaustion (Sonnentag & Fritz, 2015).We argue that when employees face high levels of ICT workload, their levels of activation are prolonged even after leaving the workplace.First, employees are aware of ICTs allowing them to work anytime and anywhere, which increases the likelihood to continue thinking about work-related issues and unfinished work, which in turn continues to drain resources.Second, employees experiencing ICT workload may also suffer from the unpredictability as to when and how they receive feedback and further tasks, promoting worry about the next workday during non-work time (ter Hoeven et al., 2016).Third, the technical possibilities increase the likelihood of working during non-work time to cope with the high amount of work.However, using ICTs such as smartphones for workrelated purposes after work undermines employees' psychological detachment (Derks et al., 2014;Park et al., 2011).In line with our assumptions, previous studies on workload have confirmed that psychological detachment mediates the relationship between workload and emotional exhaustion (Clauss et al., 2021;Sonnentag et al., 2010).Research on ICT workload showed a significant association with lower levels of psychological detachment (Barber & Santuzzi, 2015).Accordingly, we hypothesise: Hypothesis 2: Psychological detachment mediates the relationship between ICT workload and emotional exhaustion over time.Specifically, by reducing psychological detachment from work, ICT workload is positively related to emotional exhaustion.
Extended work availability and workplace telepressure as moderators Sonnentag and Fritz (2015) suggest that the relationships between the core variables within the stressor-detachment model are likely to vary between persons and situations.Within the context of work-related ICTs, we aim at extending the stressor-detachment model based on ideas from boundary theory (Ashforth et al., 2000).According to boundary theory, within a day, individuals frequently cross between their work and non-work domains and, consequently, psychologically (and when relevant physically) move between the attached roles.Those boundary-crossing activities vary interindividual.More specifically, personal and situational factors affect how individuals create, maintain and cross their boundaries (Ashforth et al., 2000).Thus, we suggest that ICT-related individual and organisational factors, that are connected to weak boundary management, aggravate the negative effect of ICT workload on psychological detachment.We propose that extended work availability as an organisational factor and workplace telepressure as an individual factor will amplify the relationship between ICT workload and psychological detachment.
Extended work availability is conceptualised as a condition in which employees are not officially working, but are available to supervisors, colleagues or customers and are, explicitly or implicitly, expected to respond to requests (Dettmers et al., 2016).Implying that employees remain constantly available during non-work time thanks to ICTs, it is an externally generated pressure (Day et al., 2012).Thus, extended work availability is perceived as the organisational consensus on how to correctly behave, influencing employees' boundary management during non-work time (Homans, 1958).Employees working under higher expectations to be available beyond working hours, might be expected to be available on at least one communication channel (e.g.instant messaging services) during their non-work time.This organisational setting influences their boundary management by reinforcing employees to create more permeable and flexible boundaries to engage in behaviours that enable them to meet the expectations (Ashforth et al., 2000), e.g. by having their work smartphone with them at all times.Empirical support stems from a study by Belkin et al. (2020), who showed that employees who are expected to monitor their e-mails after hours are more likely to spend time on emails during non-work time.Additionally, employees experiencing pressure to always be available for work-related purposes engage in behaviours that impair psychological detachment (Derks et al., 2014;Thörel et al., 2021).Beyond that, research has also shown that the mere requirement to be available leads employees to think more often about possible work-related interactions using ICTs, impairing recovery after work and psychological detachment (Belkin et al., 2020;Dettmers et al., 2016;Thörel et al., 2021).Applying these theoretical and empirical aspects to the stressor-detachment model (Sonnentag & Fritz, 2015) in the context of ICT, we propose a moderated mediation effect, where the indirect effect of ICT workload on emotional exhaustion via psychological detachment is stronger for individuals working under higher extended work availability.Therefore, we expect the following: Hypothesis 3: Extended work availability moderates the indirect effect of ICT workload on emotional exhaustion via psychological detachment.Specifically, the indirect relationship between ICT workload and emotional exhaustion via psychological detachment is stronger at high levels of extended work availability than at low levels of extended work availability.
Workplace telepressure is the combination of (a) thinking about the need to reply to work-related ICT messages and (b) feeling the urge to directly respond to such messages, both during work and non-work time (Barber & Santuzzi, 2015).In contrast to extended availability as an organisational demand that is likely to be shared between employees who work in the same context, we regard workplace telepressure as a personal factor given that the urge to respond and the preoccupation with ICT messages are likely to differ between employees in the same work context.Grawitch et al. (2018) showed that differences in workplace telepressure between employees can be accounted for by personality variables (i.e.neuroticism, workaholism, low self-control) and employees experiences of work demands.We assume that employees experiencing high levels of workplace telepressure strongly identify themselves with their work role and are mostly internally driven to stay connected (Grawitch et al., 2018).Based on boundary theory (Ashforth et al., 2000), we assume that experiencing high levels of workplace telepressure affects individual boundary management because employees tend to create more flexible and permeable boundaries around the role they highly identify with.Employees who experience a high amount of workload caused by ICTs in combination with feeling a strong internal drive to stay connected to the workplace regardless of time and space, tend to mentally gravitate to their work role to satisfy their need.Thus, the negative activation caused by ICT workload is prolonged and boosted by workplace telepressure, strengthening the negative effect on psychological detachment.Indeed, previous research has shown that employees experiencing workplace telepressure have difficulties in mentally switching off from work during non-work time (Barber et al., 2019;Pfaffinger et al., 2020), thus leading to emotional exhaustion (Hu et al., 2019).Applied to the stressor-detachment model (Sonnentag & Fritz, 2015), we expect that the indirect effect of ICT workload on emotional exhaustion via psychological detachment is stronger under conditions of high workplace telepressure.Accordingly, we hypothesise: Hypothesis 4: Workplace telepressure moderates the indirect effect of ICT workload on emotional exhaustion via psychological detachment.Specifically, the indirect relationship between ICT workload and emotional exhaustion via psychological detachment is stronger at high levels of workplace telepressure than at low levels of workplace telepressure.

Sample and procedure
The participants were approached by the authors and their students via internet platforms, social network sites and in face-to-face conversations.They were required to (a) work more than 20 hours weekly, (b) own ICTs (e.g. a smartphone), (c) use ICTs for work-related purposes and (d) not have on-call duty.We asked participants to answer three online questionnaires, with a one-week interval between the questionnaires.We chose this time lag for the following reasons.First, we did not expect employees to experience ICT workload on every workday, as within a workday some processes may occur unrelated to ICTs.In addition, no theoretical findings suggested that the exposure to ICT workload would affect the ability to detach only after several months.Thus, we decided to use a rather short time lag of one week hoping to detect the effect of ICT workload on psychological detachment and to avoid a possible disappearance (Dormann & Griffin, 2015).Second, we aimed to limit participant attrition through shorter time lags.Participation was voluntary, and the participants were assured that their responses would be treated confidentially.Participants received written feedback after the third measurement occasion.
Our data collection took place in 2017 prior the COVID-19 pandemic.Of the 501 employees contacted, 397 employees answered the online questionnaire at the first measurement occasion.Overall, 240 employees participated in the study on all measurement occasions.We excluded eight participants who did not fulfil the requirements and four participants whose questionnaires included mostly missing data for the main constructs.Of the remaining sample of 228 employees, 53.95% were women; the average age was 40.47 years (SD = 12.65); 45.18% held an academic degree (bachelor's level or higher); 35.97% had a vocational school qualification; 14.04% had a polytechnic degree, and the remainder (4.82%) had no professional qualification.Most employees had a permanent job (78.95%) and worked a regular day shift (92.11%).The participants worked on average 38.99 hours per week (SD = 12.89) in various industries (largest share: consulting, health, education and research), and 53.51% had a supervisory position.
For the main study variables, the longitudinal sample did not differ from the dropouts (non-respondents at either T2 or at T3).However, respondents were older than nonrespondents (39.79 vs. 35.47years, t(372) = 3.190, p = .002).

Measures
All constructs were assessed on the three measurement occasions (T1-T3).We sent participants a link to the online survey questionnaire on each Friday and asked them to answer the items with regard to their past week.
ICT workload was measured with three items developed and validated by Day et al. (2012).Using a five-point Likert scale (1 = never; 5 = always), participants indicated the frequency with which they experienced workload due to ICT (e.g."Technology creates more work for me").The Cronbach's alphas ranged between .90 at T1-T2 and .92 at T3.
Emotional exhaustion was assessed using five items from the German version of the Maslach Burnout Inventory -General Survey (Schaufeli et al., 1996).The respondents rated each item (e.g."Over the last week, I have felt emotionally drained from my work") using a seven-point Likert scale ranging from 1 (never) to 7 (always).The Cronbach's alphas ranged between .91 at T1, .94 at T2, and .93 at T3.
Psychological detachment was measured with four items from the Recovery Experience Questionnaire (Sonnentag & Fritz, 2007) using a five-point Likert scale from 1 (strongly disagree) to 5 (strongly agree).A sample item is "After work, I didn't think about work at all."The Cronbach's alpha was .93 at all measurement points.
Workplace telepressure was assessed with the six-item scale developed and validated by Barber and Santuzzi (2015).The respondents rated each item (e.g."It was difficult for me to resist responding to a message right away)" using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree).The Cronbach's alpha varied between .89 at T1, .93 at T2, and .94 at T3.
Extended work availability was measured with four items developed and validated by Day et al. (2012) on a five-point Likert scale from 1 (never) to 5 (always).A sample item is "I'm expected to check e-mail and / or voicemail when I'm out of the office."The Cronbach's alphas ranged between .81 at T1-T2 and .85 at T3.
Confirmatory factor analyses supported the empirical distinctiveness of our scales (factor loadings ranged from λ = .62to .91,p < .001).We compared our proposed fivefactor model against a one-, three-and four-factor model.The results are not reported here but can be obtained from the corresponding author.

Data analysis
We tested our hypotheses using structural equation modelling with Mplus 8.3 (Muthén & Muthén, 1998-2017).In all our analyses, the measurement errors were allowed to correlate over time in line with recommendations for tests of longitudinal models (Little, 2013).To evaluate our models, we used chi-square tests and considered the following fit indexes as indicators of an acceptable fit: Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) values of >.90, Root Mean Square Error of Approximation (RMSEA) values of <.06 and Standardised Root Mean Square Residual (SRMR) values of <.08 (Hu & Bentler, 1999).The significance of the indirect effects was tested using bias-corrected bootstrapping (k = 1000) (Hayes, 2015).To test our moderated mediations, we conducted our analyses using the approach and Mplus code developed by Stride et al. (2015).To probe the conditional direct and indirect effects, we calculated simple slopes values at low (−1 SD from the mean) and high (+1 SD from the mean) values of the moderators.Furthermore, within our analysis we controlled for stability effects, meaning that corresponding variables predict themselves over time.
Measurement invariance is a precondition to test longitudinal effects (see Little, 2013).Test of measurement invariance showed no differences between the fit of a baseline model (model with identical measurement models on all three measurement occasions), a weak invariance model (equal loading across measurement occasions), and a strong invariance model (equal intercepts of corresponding factors across measurement occasions).Thus, we assumed strong measurement invariance in our sample.Results can be obtained from the corresponding author upon request.

Results
Table 1 shows the means, standard deviations, reliabilities and intercorrelations of all study variables.The results revealed significant correlations between the study variables ICT workload, psychological detachment, emotional exhaustion, extended work availability and workplace telepressure in the expected directions.Comparison between our hypotheses' tests with and without control variables (age, gender) yielded identical results.Therefore, we report the results without control variables.

Mediation analysis
Testing hypothesis 2, we specified a model in which we investigated whether the relationship between ICT workload (T1) and emotional exhaustion (T3) was mediated by psychological detachment (T2), including stability effects for emotional exhaustion and psychological detachment.The mediation model showed a satisfactory fit (χ 2 = 777.77,df = 568, p < .001,χ 2 /df = 1.37,CFI = .97,TLI = .97,RMSEA = .04,SRMR = .05).Inspection of the model showed that the stability effect of psychological detachment (T1) on psychological detachment (T2) was very high (b = .948,p < .001),indicating that the individuals' relative scores for psychological detachment changed very little between the first and second measurements.Therefore, the short time lag of only one week between measurement occasions seemed not to be ideal for investigating changes in psychological detachment in this sample.As a result, it did not seem feasible to include the stability effect of psychological detachment, because the very high stability explained a large proportion of the variance of psychological detachment (R 2 = 78%).As can be expected by the high stability effect, a model including the stability effect of psychological detachment (T1) on psychological detachment (T2) failed to confirm the indirect effect of ICT workload (T1) on emotional exhaustion (T3) via psychological detachment (T2) (0.006, 95% CI [−0.022, 0.045]).Because ICT workload (T1) and psychological detachment (T2) showed a high zero-order correlation (r = −.49,p < .001),we decided to explore our hypothesised indirect effect, excluding the stability effect of psychological detachment.Importantly, the stability effect of emotional exhaustion was kept in the model.

Moderated mediation analyses
In line with the previous analysis, we hereby report the moderated mediations excluding the stability effect of psychological detachment.Running the moderated mediations including the stability effect of psychological detachment did, again, not yield significant results.
In Hypotheses 3 and 4, we expected that extended work availability (T1) and workplace telepressure (T1) would moderate the indirect effect of ICT workload (T1) on emotional exhaustion (T3) via psychological detachment (T2).The results showed that extended work availability (T1) was not a predictor of psychological detachment (T2) (see Table 2).In addition, the interaction between extended work availability and ICT workload was not significant.The index of moderated mediation was not significant (0.002, 95% CI [−0.034, 0.033]), indicating that the overall moderated mediation model was not supported.Thus, Hypothesis 3 was rejected.

Discussion
The purpose of this study was to attain a greater understanding of the effect of ICT workload on emotional exhaustion along with mechanisms that might explain this relationship.By applying the stressor-detachment model (Sonnentag & Fritz, 2015) to the context of ICT in the workplace, we showed that ICT workload was positively related to emotional exhaustion and that psychological detachment mediated this relationship.Furthermore, we advanced previous research by showing that workplace telepressure enhanced the negative relationship between ICT workload and psychological detachment and moderated our proposed mediation model.Our results have theoretical implications.First, we add to the scant literature on the effects of ICT workload on health (Barber & Santuzzi, 2015;Day et al., 2012), by demonstrating longitudinally that ICT workload is positively related to emotional exhaustion.Our finding strengthens the hypothesis that exposure to workload resulting from ICTs increases the risk of emotional exhaustion.Future research might deepen the insights into the detrimental impact of ICT workload by analysing possible differences between occupational groups or job designs.Certain occupational groups might be fully adapted to working digitally (e.g.programmers) and thus primarily appraise ICT workload as a technology-driven challenge and be less likely to suffer from emotional exhaustion (Webster et al., 2011).Furthermore, our data were collected prior to the COVID-19 pandemic.Meanwhile, employees working mainly remotely might perceive higher levels of ICT workload, enhancing strain (Ingusci et al., 2021).Future research could analyse the impact of ICT workload on recovery processes and strain when employees are forced to work mainly remotely.Second, our results largely confirm the proposed relationships of the stressor-detachment model (Sonnentag & Fritz, 2015) in the context of ICT use.Specifically, we could show that high levels of ICT workload were negatively related to psychological detachment, and that psychological detachment mediated between ICT workload and emotional exhaustion.Our findings support the idea that high ICT workload impairs one's ability to detach oneself from work and, thus, increases the likelihood of emotional exhaustion.Clauss et al. (2021) reported consistent findings of the negative relationship between regular workload and emotional exhaustion being mediated by psychological detachment.Future research should test whether these effects are rather short term (e.g. by conducting a daily diary study) or long term (e.g. using time lags of multiple months), because crosslagged effects may vary depending on the time lags between measurement points (Dormann & Griffin, 2015).For example, in a daily diary study by Derks et al. (2014), psychological detachment did not mediate the relationship between daily work-related smartphone use and exhaustion.Fluctuations in ICT workload may change employees' willingness to expend effort and, consequently, influence recovery processes such as psychological detachment to a different extent (Meijman & Mulder, 1998).
Third, going beyond previously investigated relationships, in this study, we proposed extended work availability and workplace telepressure as ICT-related organisational and individual factors amplifying the effect of ICT workload on psychological detachment.This reasoning was based on boundary theory that states that personal and situational factors influence individual boundary management, e.g. by reinforcing employees to create more flexible and permeable boundaries, making boundary-crossing more likely to meet the personal or situational expectations (Ashforth et al., 2000).In contrast to our hypothesis, we did not find support for the moderated mediation model with extended work availability.The negative relationship between ICT workload and psychological detachment existed irrespective of employees' extended work availability.Experiencing high levels of extended work availability did not strengthen the negative relationship, nor did low levels of extended work availability weaken the negative relationship.Thus, ICT workload seems to be harmful for psychological detachment, regardless of the organisational condition to constantly be available.Whether the effects of extended work availability unfold, might depend on employees' commitment to their organisation (Turner et al., 2006).However, our research design might also be a reason why we did not find the proposed effect.Besides our short time lags, we operationalised extended work availability differently than previous studies.Thörel et al. (2021) measured extended work availability by asking for the average weekly time spent answering work-related calls and messages.Similar to Dettmers (2017), we defined extended work availability as the requirement to be available by ICTs after working hours, which is not the same as actually being called or as answering work-related messages.We assumed that anticipation of being contacted for work-related purposes already affects the recovery process and may cause strain (Dettmers et al., 2016).However, it might be important to detect both psychological and physical role transitions.
In line with our prediction, the moderated mediation model with workplace telepressure was supported.We found that employees who reported high levels of workplace telepressure had more difficulties to mentally detach from work when they experienced high levels of ICT workload, leading to a higher risk of emotional exhaustion.Our results corroborate previous findings on detrimental effects of workplace telepressure on psychological detachment and exhaustion (Pfaffinger et al., 2020) and moderated mediations with workplace telepressure (van Laethem et al., 2018).By showing that workplace telepressure amplifies the negative ICT workload-detachment relationship, our study provides support for the extended stressor-detachment model (Sonnentag & Fritz, 2015) with regard to ICT-specific stressors in work context.We combine the stressor-detachment model with boundary theory (Ashforth et al., 2000) by showing an impact of individual boundary management.While we did not find support that the externally generated pressure to constantly be available influences role transitions, our results support the idea that individual factors play a crucial role within the stressor-strain relationship.The internal need to constantly check and reply to work-related messages enhances the detrimental effect of ICT workload on employees' ability to mentally detach during non-work time.We recommend that future work further elaborates the impact of ICT-related stressors on well-being outcomes and on potential moderator variables, using different time lags.For example, the effects of ICT workload on psychological detachment may depend on the extent to which an employee feels control over ICT use (Day et al., 2010).

Limitations
The study has the following limitations.First, by applying a longitudinal design, we found empirical support for the negative effect of ICT workload on emotional exhaustion and the assumption that psychological detachment mediates this effect.However, our results should be interpreted with caution because we did not control for stability effect of psychological detachment throughout all analyses.Without controlling for effects of a construct on itself at a later measurement point, we cannot prevent the estimates of our cross-lagged effects to be falsely inflated (Cole & Maxwell, 2003).Thus, it is important that future studies determine whether the reported relationship can be replicated when a stability effect of psychological detachment is considered.
When the data were analysed with the statistical precondition of stability effects for all measurement variables, the mediator model tested was not supported.There may be several reasons for this result.First, controlling for stability effects may dramatically reduce the effect sizes in longitudinal designs (Adachi & Willoughby, 2015).In our study, the stability effect of psychological detachment (T2) on psychological detachment (T1) was very high (b = .95).Controlling for it removed more than half the variance in psychological detachment (T2) that was shared with ICT workload (T1) (i.e..51).Thus, controlling for stability in psychological detachment dramatically reduced the magnitude of the effect size of ICT workload (T1) as a predictor of psychological detachment (T2), consequently affecting the proposed mediator model.Second, as discussed above, our chosen time lag of one week between each measurement point turned out to be too short for detecting a significant mediation effect.Thus, future research with different time lags is needed and, in addition, studies with longer time lags might also be able to draw conclusions on how the exposure of ICT workload influences employees' health in the long run.
Second, we need to address common method bias because our study variables were assessed using self-report methods which might have led to inflation between the variables (Podsakoff et al., 2003).However, Kao et al. (2020) argue that variables such as extended work availability, workplace telepressure und emotional exhaustion are most validly operationalised by self-reports, because they are either based on an employee's perception or on the psychological experiences of the employee.In addition, we minimised the bias by using a time-lagged design (Podsakoff et al., 2003) and choosing validated scales which proved to be reliable in our study (i.e. α between .81 and .95).Future studies might extend research by using multisource measures, for example by collecting possible objective indicators of ICT workload such as the number of work tasks sent by e-mail.
A final limitation of this study might be the fact that we collected data using a snowballing technique, resulting in a sample consisting of student-and non-student-recruited participants.The external validity and generalizability of our study results to the general population might therefore be limited.However, a previous study showed that studentand non-student-recruited samples yielded comparable findings, leading to identical practical conclusions from the findings (Wheeler et al., 2014).Furthermore, external validity and generalisation are ensured because the data of our sample are heterogeneous, as it includes individuals from various industries and occupations as well as a variety of individual and contextual characteristics due to the different networks of the recruiters (Demerouti & Rispens, 2014).

Practical implications
Our study suggests that the workload that employees attribute to ICTs can impair employees' recovery processes and be detrimental for their well-being.As ICTs are largely indispensable for working life, organisations should support their employees with the handling of additional work and possible risks due to ICT-related stressors.
First, training employees to handle the perceived increased workload due to ICTs could buffer its negative impact on psychological detachment.When employees prioritise checking and answering multiple communication channels in the planning of their workday (e.g. at the beginning of the workday, the phone is checked first because missed phone calls require faster reactions than emails), they may feel capable handling their ICT workload (Casper & Sonnentag, 2020).
Second, training employees to handle ICTs could buffer the detrimental effects of ICT workload.When employees are trained to actively apply the additional features of ICTs in their workday (e.g.free slots and the preferred communication channel are communicated in the e-mail signature), they might experience technological benefits rather than an increased workload due to ICT use, enabling them to better mentally switch off during non-work time.A previous study showed that employees who have qualifications and have been trained in proper ICT use experience less strain (Korunka & Vitouch, 1999).When employees perceive ICTs as useful and reliable, they experience lower levels of work overload (Ayyagari et al., 2011) and less exhaustion (Chen et al., 2009).
Third, internal communication about responsible and well-balanced use of ICTs could demonstrate a company's awareness about ICT-related stressors and potential solutions.For instance, promoting awareness about a possible increased workload due to ICTs within organisations might encourage teams to change workflows (e.g.team members do not immediately write emails, but rather raise issues within regular meetings), preventing email overload.Communicating changes due to ICTs within the workplace reduces the negative consequences of techno-stressors on well-being (Ragu-Nathan et al., 2008).

Conclusion
The present study contributes to the scarce research on ICT workload by showing that employees experiencing high levels of ICT workload are at a higher risk of suffering from emotional exhaustion.An impaired ability to mentally switch off from work during non-work time seems to mediate this relationship.Furthermore, we found that workplace telepressure moderates the conditional indirect effect of ICT workload on emotional exhaustion via psychological detachment.We conclude that strong individual factors are likely to influence the impact of ICT-related stressors.Thus, further research on the extended stressor-detachment model with ICT-related stressors is needed because psychological detachment and ICT-related moderators seem to play a crucial role in employees' well-being in the modern work context.Overall, we encourage researchers to explore possible detrimental effects of ICT workload on employees' well-being using longitudinal designs, possibly confirming our results.For organisations, we recommend promoting awareness about possible detrimental effects of increased workload due to ICTs and fostering psychological detachment.

Table 1 .
Means, SDs, reliabilities, and correlations of all variables.

Table 2 .
Results of the hypothesised structural equation models.