Measuring Work Demands and Resources of Digitalisation: The ICT Resources and Stressors Scale

The Need for a New Scale

In our opinion, incorporating a comprehensive questionnaire to assess stressors and resources resulting from digitalisation is crucial for organisational success. Such a scale could serve as an essential tool for identifying potential risks to employee well-being stemming from the rapid changes associated with digital transformation. By systematically measuring ICT-related stressors and their effects on outcomes such as workload (e.g. Zinke et al., 2023), work intensification (Mauno et al., 2019), or technological uncertainty (e.g. Pfaffinger et al., 2020), organisations can pinpoint specific areas needing attention and intervention. This could help mitigate the risks of burnout and turnover associated with digital transformation. High levels of stress resulting from rapid changes and uncertainty can lead to employee disengagement, decreased morale, and ultimately, turnover. By proactively identifying stressors and providing targeted support, organisations can create a supportive work environment that promotes employee well-being and reduces the likelihood of burnout and turnover, preserving institutional knowledge and fostering organisational citizenship among employees. Moreover, a scale that also measures ICT-related resources would further allow organisations to understand which resources are crucial for their employees to use ICTs successfully. This enables companies to ensure that their employees have the necessary tools and support systems to effectively master digital challenges.

In addition to the practical applications within organisations, a questionnaire assessing resources and stressors in the context of digitalisation is also invaluable for research purposes. By collecting data on ICT-related stressors and resources, researchers gain insights into the digital workplace’s complexities. This data informs strategies to mitigate negative impacts and promote positive outcomes. Thus, such a questionnaire may contribute not only to understanding future challenges but also to developing interventions for individuals, organisations, and society.

Our study aims to contribute to the expanding body of literature by identifying a comprehensive range of existing ICT-related resources and stressors and developing a self-report questionnaire for their assessment in organisational settings, thereby offering practical insights for fostering a healthy and productive work environment.

Theoretical Considerations

We conducted our study based on established theories of stress research. Zapf and Semmer (2004) define stress as a perceived imbalance between demands (i.e. stressors) and a person’s ability (i.e. resources) to respond to them. This imbalance is perceived as unpleasant and reduces well-being. Chronic work-related stress can lead to illness. The present study uses the job demands-resources model (Bakker & Demerouti, 2007; Demerouti et al., 2001) as a framework to explain how stress occurs and applies it to the context of ICTs. Central to the model is the assumption that resources and demands are triggers for two different and independent processes. On the one hand, work demands deplete personal physical and psychological resources and thus impair health. On the other hand, work resources fulfil psychological needs and thus primarily influence attitudes towards work. It is further assumed that there are interactions between resources and work demands, in that the positive effects of resources have a mitigating effect on the negative influences of work stressors.

ICT demands and resources

Day et al. (2010) developed a framework for ICT demands and resources based on the job demands-resources model (Bakker & Demerouti, 2007; Demerouti et al., 2001), the transactional stress model (Lazarus, 1966; Lazarus & Folkman, 1984) and the conservation of resources theory (Hobfoll, 1989). ICT demands were defined as ‘any ICT factor or process at work involving some type of storing, transmitting, or processing technology (e.g., computer programs) or device (e.g., computer, cell phone) that have the potential to be perceived as stressful by workers” (Day et al., 2010, p. 324). ICT demands include response expectations, availability, poor communication, lack of control, ICT hassles, employee monitoring, learning expectations, and workload. ICT resources include personal assistance and ICT resources/upgrades and were defined as “any ICT factor or process at work […] that assist employees with the completion of their work, reduce the burden of job demands, or that promote personal growth and development.” (Day et al., 2010, p. 324). According to Day et al. (2012), ICT demands and resources are associated with job demands such as overload, ambiguity, and job control as well as with health outcomes like experience of stress, strain, exhaustion, and cynicism.

Technostress

Stress resulting from technological change has long been a research topic in the field of information systems. Brod (1984) introduced the term technostress, which describes a stress response resulting from the maladaptive use of ICTs. Based on the transactional stress model (Lazarus, 1966; Lazarus & Folkman, 1984), Tarafdar et al. (2007) and Ragu-Nathan et al. (2008) developed the technostress inventory. They identified technostress creators (i.e. stressors) that can lead to stress experiences as well as technostress inhibitors (i.e. resources) that reduce these effects (see Tarafdar et al., 2007). The technostress creators include techno-overload (a sensation that ICTs force their users to work more and faster), techno-invasion (all-time accessibility through ICTs), techno-complexity (feeling of inadequacy resulting from ICT use), techno-insecurity (job-insecurity because of ICTs), and techno-uncertainty (arising from continuing changes in ICTs). The technostress inhibitors include literacy facilitation (organisational support for learning new ICTs), technical support provision, and involvement facilitation (inclusion of employees in technological change and encouraging interaction). Since its conceptualisation, technostress has been extensively investigated and expanded, and there is empirical evidence of its predictive value. Technostress has been associated with job satisfaction, organisational commitment, turnover intentions, role overload, role conflict, reduced productivity, and increased role stress as well as with health outcomes such as exhaustion, burnout, and strain (see for example Ayyagari et al., 2011; D’Arcy et al., 2014; Galluch et al., 2015; Kaltenegger et al., 2023; Maier, 2014; Maier et al., 2015; Nimrod, 2018; Riedl, 2012; Riedl et al., 2012; Salanova et al., 2013; Sarabadani et al., 2018; Tarafdar et al., 2011).

Antecedents of technostress

Expanding upon the technostress framework, Ayyagari et al. (2011) identified characteristics of ICTs that function as antecedents of technostress creators, using the person-environment fit model (Edwards & Cooper, 1988) as a theoretical basis. A key characteristic is the usability of ICTs. When used voluntarily, usability is associated with the acceptance and use of ICTs (Weil & Rosen, 1997). In the work context, where the use of ICTs is not voluntary per se, employees have to use the available technologies despite a possible low perceived usability. This can lead to an increase in perceived workload and health problems such as musculoskeletal pain (Åborg & Billing, 2003). Ayyagari et al. (2011) consider usefulness an important aspect of ICT usability (next to complexity and reliability). Technologies that are considered useful reduce feelings of workload, leading to employees accomplishing work tasks faster and being more productive (Ayyagari et al., 2011).

Telepressure

Workplace telepressure is defined as “the combination of a strong urge to be responsive to people at work through message-based ICTs with a preoccupation with quick response times.” (Barber & Santuzzi, 2015, p. 172). Workplace telepressure can negate the benefits of asynchronous communication (e.g. flexible response times) if employees start to view it as a synchronous form of communication that requires an immediate response. This can lead to prioritising responding to messages and neglecting recovery times. Telepressure involves internalising existing norms of expectation in the workplace and is associated with workaholism, absenteeism, poorer sleep quality, work overload, emotional exhaustion, less detachment from work, lower satisfaction with one’s work-life balance (Barber & Santuzzi, 2015; Barber et al., 2019; Grawitch et al., 2018; Santuzzi & Barber, 2018) and its effects on a range of biological parameters is currently being researched (Semaan et al., 2023).

Defining the Item Pool

As mentioned above, several scales were created over a decade ago and did not meet the specified requirements in Podsakoff et al. (2016) to have clear conceptualisations and definitions, and consequently needed to be adapted for contemporary use. Furthermore, following an interview study, Fischer et al. (2019) recommended revising and updating the technostress inventory. In addition, Nastjuk et al. (2023) describe in their meta-analysis how technostress creators have varying relationships with different outcome variables. Moreover, they are commonly aggregated in a second-order construct. In accordance with the job-demands-resources framework, we reconceptualise each technostress creator as an individual stressor. Therefore, to ensure construct and content validity, we revised the definitions of all identified constructs using the guidelines in Podsakoff et al. (2016) and MacKenzie et al. (2011) specifying the conceptual domain and theme for each. This led to the reconceptualisation of ICT control as a resource instead of as a stressor that was defined solely through the absence of specific characteristics, and in accordance to previous research (e.g. Karasek, 1979; Semmer et al., 1995). Furthermore, we created new items in some instances, to guarantee that all facets of the constructs are captured, and adapted item wordings to be consistent within the questionnaire (e.g. regarding the use of the term information and communication technology). During this step, all items were translated into German and then retranslated by a native English speaker with a background in occupational and organisational psychology. Discrepancies were discussed and corrected where necessary. Eight psychologists then participated in the preliminary study to ensure content validity as well as comprehensibility of the scales. Most of the feedback concerned the intelligibility of the items (e.g. unclear words such as “end user”) or their content (e.g. the original items of involvement facilitation did not capture the perspective of the employees). Where modifications were possible without changing the meaning, the items were adapted accordingly. An overview of the revised items is shown in Table 3. On the reader’s request, we can provide a full detail of the steps undertaken during this phase.

Methods

Sample

From November 2020 to February 2021, a total of 375 German-speaking employees in Switzerland (269 female, 103 male, 1 other, 2 no information) between 18 and 73 years (M = 34.1, SD = 11.5) completed the online survey. Participants in retirement age were included, if they indicated that they were still working. Among the participants were 38 employees of a foundation operating in the Swiss health sector. In addition, participants were recruited via the personal environment of the author (n = 105) and social media (n = 149). These participants had the opportunity to take part in a lottery for a CHF 50.00 gift certificate at the end of the survey. Furthermore, 83 participants were working students from the University of Bern and received a course credit in exchange for their participation. Two individuals were excluded because they did not meet the requirement of employment of at least 20% or 8 work hours per week. Moreover, 17 participants were excluded for not completing the online survey carefully (see section data analysis). Thus, the sample for validation contained the data of 356 participants. See Figure 1 for an overview of the sample’s characteristics regarding most reported work areas, education level, years of employment, and importance of ICTs for their work. The participants’ education level was high, with 45% reporting a degree from a higher education institution (university, institute of technology, or university of applied sciences). In total, over 200 different job titles were reported (e.g. project managers, commercial employees, psychologists, chefs, CEO). The median employment level was 80% (min. 20%, max. 160%). 76% of all participants had been employed at their current job for five years or less. 30% reported having supervisory responsibilities and 4% stated to be self-employed. The median importance of ICTs for one’s job was 9.5 on a scale of 0 to 10 (M = 8.3, SD = 2.4).

Figure 1

Control and split variables

We assessed the participants’ age and attitudes towards digital change as control variables. The latter was measured with the subscale technology acceptance by Neyer et al. (2012) using a five-point Likert scale (example item: ‘I am very curious about new technical developments.’). The reliability of the scale was α = .82, ω = .82.

Importance of ICTs for one’s work was included as a split variable for the known-group comparison and was measured with the question ‘How important are ICTs for the execution of your work?’. The participants reported their assessment using a slider from 0 = ’ICTs are unnecessary for my job’ to 10 ‘My job cannot be carried out without ICTs’.

Outcome variables

Job satisfaction was assessed with the single-item scale general job satisfaction (‘How satisfied are you when you look at your work situation in general?’) and the job satisfaction scale (example item: ‘If certain things don’t change soon at my job, I’ll look for a new job.’) by Semmer et al. (1990). The four items were combined into one scale. The reliability was α = .79, ω = .80.

Intentions to quit were assessed with two items from Baillod (1992), adapted from Bluedorn (1982) (example item: ‘How likely is it that you will still be working at your current company in six months?’). The reliability was α = .80 and the corrected correlation of the two items was ρ = .81.

Affective commitment was assessed with four items by Allen and Meyer (1990) (example item: ‘I enjoy talking to others about my organisation.’). The reliability was α = .87, ω = .87.

Exhaustion was assessed with eight items of the Oldenburg burnout inventory by Demerouti et al. (2001) (example item: ‘After work, I now often need longer recovery times than before to get fit again.’). The reliability was α = .83, ω = .84.

Well-being was measured using the five items of the WHO-5 (Topp et al., 2015; WHO Regional Office for Europe, 1998) (example item: ‘In the last two weeks, I have been happy and in a good mood’). The reliability was α = .88, ω = .88.

General health was measured with a single-item scale (adapted from Igic et al., 2014) (‘How would you describe your health in general?’).

Assessing incremental validity

Qualitative overload and task completeness were measured with three, resp. one item from Udris and Rimann (1999) (example item ‘You have to do things for which you are actually not trained and prepared enough.’, ‘In my work, you can produce or carry out a thing or a job from A to Z.’). The reliability of qualitative overload was α = .79, ω = .79.

Problems with the organisation of work tasks (POWT), participation, and job control were assessed with four, one and six items using the corresponding subscales from the instrument for stress-oriented task analysis ISTA (Irmer et al., 2019; Semmer et al., 1995) (example items: ‘Which of the two workplaces (A or B) is more similar to your workplace? A has to spend a lot of time getting information, material or tools to continue working. B always has the necessary information, material or tools available.’, ‘In decisions affecting my situation as an employee, I have no influence at all’, ‘Can you decide for yourself the way in which you do your work?’). The reliabilities were α = .67, ω = .68 for POWT, and α = .91, ω = .92 for job control.

Social stressors in relation to work colleagues were assessed with five items by Frese and Zapf (1987) (example item: ‘One often has arguments with some work colleagues.’). The reliability was α = .80, ω = .80.

Results

Predicting health and attitudes towards work (Step VI)

To test criterion validity, we examined whether ICT resources and stressors predicted health and work-related outcomes using structural equation modelling. The fit of the measurement model was acceptable (χ² = 1371.63, df = 806, p < .001; CFI = .914; TLI = .904; RMSEA = .048, 90%–CI [.044; .053]; SRMR = .059). Figure 2 shows the significant paths of the structural equation model, and bivariate correlations of the latent variables are shown in Table 9. Telepressure was a significant predictor of exhaustion, well-being, and general health. In contrast, it did not predict any of the work-related variables, demonstrating its deteriorating effects were primarily on health outcomes. The ICT resources were primarily associated with work-related outcomes: involvement facilitation predicted job satisfaction, intentions to quit, and affective commitment as well as exhaustion and general health. ICT control was predictive of all work-related outcomes as well as exhaustion and well-being. Personal assistance only predicted job satisfaction and affective commitment, while ICT resources & upgrades predicted job satisfaction and general health.

Figure 2

Table 9

Bivariate correlations between predictors and outcome variables.

VARIABLES		1	2	3	4	5	6	7	8	9	10
1	InvoFac
2	Control	.18**
3	ICTResU	.56***	.25***
4	PersAs	.48***	.27***	.47***
5	Telepre	.05	.16*	.08	.02
6	JobSat	.38***	.23**	.36***	.35***	–.05
7	IntQuit	–.30***	–.30***	–.22**	–.28***	–.03	–.77***
8	AffComm	.36***	.35***	.28***	.33***	.10ƚ	.81***	–.88***
9	Exhaustion	–.32***	–.17*	–.26***	–.22**	.26***	–.49***	.31***	–.27***
10	Wellbeing	.19**	.15*	.15*	.17*	–.15*	.50***	–.36***	.40***	–.64***
11	GenHealth	–.01	.01	.14*	.09	–.12*	.17*	–.11ƚ	.11ƚ	–.40***	.47***

[i] Note. ƚ p < .10, * p < .05, ** p < .01, *** p < .001.

Interaction effects of ICT resources (Step VII)

In accordance with the job demands-resources model (Demerouti et al., 2001), we expected the ICT resources to moderate the relationship between ICT stressors and the outcome variables. We tested the interaction effects of ICT resources using the double-mean-centering strategy (Crowson, 2020; Lin et al., 2010). There was a significant interaction between ICT control and telepressure in association with job satisfaction (β = .17, p = .024). The model fit was very good: χ² = 370.713, df = 365, p = .407; CFI = .999; TLI = .999; RMSEA = .008, 90%–CI [.000; .024]; SRMR = .039. Simple slope analysis showed that the interaction was significant when ICT control was low (–1 SD, p = .018), but not significant at average (M, p = .178) or high (+1 SD, p = .109) levels of ICT control (see Figure 3), suggesting that low ICT control reinforces the negative relationship between telepressure and job satisfaction. No significant interactions were found for the remaining ICT resources.

Figure 3

Integration of results and final questionnaire

Following the item analysis (step III), 23 items measuring six constructs remained. The analysis of the measurement models showed that the overall model has a good fit and is significantly better than the alternative models. Literacy facilitation, however, did not meet the evaluation criteria recommended by MacKenzie et al. (2011) and was therefore excluded after step IV. The remaining scales showed scalar measurement invariance concerning gender and the importance of ICTs for one’s work. Furthermore, we established preliminary evidence for significant differences in the empirical mean values between groups with high and low importance of ICTs. However, whereas the decrease in CFI was above the threshold for accepting the stricter model, the test of small differences in fit (MacCallum et al., 2006) did not support this. Regarding criterion validity, the scales involvement facilitation, ICT control, ICT resources & upgrades, and telepressure were able to predict health (i.e. exhaustion, well-being, and general health) and/or work outcomes (i.e. job satisfaction, affective commitment, and intentions to quit). The scales explained additional variance in health and work-related outcomes and thus demonstrated incremental variance. Our analyses confirmed that telepressure mainly increased explained variance in health outcomes while the ICT resources primarily increased explained variance in attitudes towards work.

Overall, the scales involvement facilitation, ICT control, ICT resources & upgrades, as well as telepressure, showed robust results in the different steps of the validation process, particularly regarding criterion validity. Personal assistance, however, could not predict health or work-related outcomes after controlling for age and personal attitude towards digital change, and when general constructs were taken into account. Thus, we excluded it from the final questionnaire after step IX. Table 13 provides an overview of the results. The final ICT Resources and Stressors Scale contains four constructs measured by 16 items and is included in Appendix C.

Practical and theoretical implications

The results suggest that the scales are a reliable way to measure key resources and stressors in the context of digitalisation. The scales are internally consistent and can be used separately or together to identify possible areas of action and to expand occupational health management. The validation among employees of the foundation, which used the obtained results to identify potential for internal digital transformation processes, further demonstrates the practical value of the scales.

The present questionnaire can be used in various contexts. For example, the scales telepressure and ICT control could be administered to reduce stress from ICT-based messaging, which is a key form of communication for cooperation and task completion in organisations. With the increase in hybrid or working-from-home work models, it has even become more important (Greer & Payne, 2014). However, employees often feel overwhelmed by the constant influx of such messages (e.g. e-mails), which leads to lower productivity and an increased experience of workload (Sonnentag et al., 2018). Administering the questionnaire, the organisation is able to assess the extent of telepressure experienced by employees and identify strategies to mitigate its negative effects, such as introducing messaging management policies, providing training on effective communication practices, or exploring alternative communication channels. Our findings suggest that the negative effects of telepressure are higher when ICT control is low. Thus, measuring ICT control could be used to address work design decisions when seeking strategies to mitigate telepressure or implementing new processes.

In another specific scenario, the questionnaire could be applied to improve the efficiency and productivity of the organisation before the transition to new strategic software. Measuring involvement facilitation and ICT resources and upgrades is a suitable tool for organisations that aim to assess how their employees perceive support in terms of technological change in line with work design recommendations (Parker & Grote, 2022). The scale ICT resources and upgrades can help to assess the needs for more appropriate ICTs and ensure that the new resources are sufficient, and effectively address previous technological shortcomings. This guarantees that employees have the necessary tools to perform their tasks efficiently. In addition, employee involvement in change management is crucial. With the involvement facilitation scale, organisations can ensure a smooth introduction and minimise resistance from employees who rely on using the new software to complete their work tasks.

The present study provides new insights into the measurement of work-related resources and stressors in the context of digitalisation that are not captured by existing instruments. We expanded upon the work of Tarafdar et al. (2007), Ragu-Nathan et al. (2008), Ayyagari et al. (2011), Day et al. (2012), and Barber and Santuzzi (2015), reviewing the original definitions, adapting, or creating new items to capture the constructs in their entirety, and thus optimising the measurement of involvement facilitation, ICT control, ICT resources & upgrades, and telepressure. In doing so we addressed the criticism of Fischer et al. (2019) concerning the conceptualisation of technostress. We claim that our scales are a valuable supplement to other measurement instruments concerning technological stressors or resources, such as Fischer et al.’s (2021) revised and updated digital stressors scale. Our questionnaire not only enables the measurement of stressors but also of resources at the same time. The development process recommended by MacKenzie et al. (2011) and Podsakoff et al. (2016) ensures the quality of the scales. The study extends the job demands-resources model (Bakker & Demerouti, 2007; Demerouti et al., 2001) to ICT-specific stressors and resources.

Limitations and future research

The study has some limitations that need to be considered. First, the ICT resources and stressors examined in this study should not be considered exhaustive. In the literature review, we aimed to find constructs that can be captured by existing scales and are relevant in the work context. Thus, stressors and resources resulting from private ICT use were not considered. We aimed to search for scales that are widely applicable in the general population of employees. Therefore, occupation-specific resources and stressors were not included. In a specific case, it may therefore make sense to adapt the questionnaire to the specific circumstances.

Furthermore, as Fischer et al. (2019) and Marsh et al. (2022) note, work-related digital stressors and resources change over time, and it is essential that measurement instruments are regularly updated. Second, all variables were measured cross-sectionally and via self-report, which raises the issue of common method variance (Podsakoff et al., 2003). However, by using questions to assess the participants’ attention, simplifying the online survey, and randomising the order of scales and items, we increased the data quality and minimised biases due to inattentive completion. Third, sampling effects may have influenced the relationships between the variables (Kelava & Moosbrugger, 2012). Women, individual work areas as well as professions in which ICTs are a prerequisite were overrepresented, and the education level of the participants was relatively high. This reduces external validity. Fourth, the practical value of the questionnaire has been demonstrated by applying it to the employees of a healthcare organisation. However, this step should be repeated in a future study with different organisations. In addition, the results obtained should be cross-validated on a new sample to further validate the questionnaire in the field (Hinkin, 1998; MacKenzie et al., 2011).

Completing the process of scale development, scale norms should then be created for benchmarking by administering the questionnaire to a representative sample of the workforce. Additionally, the new questionnaire should be translated into other languages. This requires that the scales are revalidated for use in different language regions and that measurement invariance regarding questionnaire language is established.

Conclusion

Following the standards of MacKenzie et al. (2011), we first identified key constructs in the existing literature on resources and stressors from digitalisation. We then refined the definitions and developed an instrument to comprehensively capture these constructs. The result of our validation study is a 16-item questionnaire covering four constructs with good psychometric properties. Our findings align with the job demands-resources model (Bakker & Demerouti, 2007) and extend it to ICT-specific stressors and resources. Our questionnaire enables organisations to systematically assess and address ICT-related stressors and resources and to improve occupational health management. Thus, our instrument not only helps to understand the complexity of the digital workplace but also to formulate appropriate strategies to mitigate negative effects and promote positive outcomes when introducing new ICTs.