Introduction

The challenges created by Covid-19 pandemic have impacted the whole economy largely. Based on the prevailing unpredictable situation, flexible working models have been accelerated by the organizations in all over the world. “Work from home” (WFH) as a concept of remote working where employees work from their houses, has been adopted by all organizations. Almost all sectors have implemented the work from home policy as a measure to the maximum extent feasible for all activities. This became an important step against Covid-19 to reduce workplace contact. Many organizations switched to teleworking and virtual meetings through video conferencing to adapt and cope up with the current scenario. Working from home may continue even when the economy reopens. Companies in IT sector and financial service have invested in remote work tools as there is no trace of resuming the old way of working. The mode of working is now completely with rapid use of mobile phones, computer technology, e-mail, and collaboration tools. Long commutes have been replaced with zoom and other conference mode of working. The coronavirus pandemic changed the way of business and work system facilitated by smart technologies. According to an article, NASSCOM (The National Association of Software and Services Companies) is working to create new models of working and engagement systems. Though technology with work from home model has facilitated the organizations with positive outcomes such as increase in overall efficiency and production, saving the commute time at the same time it has posed a threat to the wellbeing and motivation of employees resulting in technostress with the more use of it. There are certain challenges with significant changes in the professional and personal life of individuals now. One of the challenges is psychological impact of technology on individuals which is known as “technostress”. When employees go through a certain level of stress and anxiety, they lose confidence and interest in taking responsibilities, and their performance go down. Stress related to technology at workplaces has been investigated since 1980s, and many studies have addressed this problem for decades. The same problem has gained the researcher’s much needed attention during Covid-19 pandemic due to ongoing work from home setting in IT industry all over the world. While organizations are constantly putting efforts to make virtual working mode effective in the current crisis, many people are finding it more stressful with this work-extension technologies. There is a negative impact on the psychological wellbeing of the employees due to this excessive and compulsive use of technologies. “Technostress”—stress caused by use of technology—leads to fatigue, anxiety, lack of sleep, depression and reduced performance. During this outbreak of Covid-19, IT industry in India has adopted “work from Home” as per government’s mandate. There are numerous stressors due to information and communication technology at workplaces, and these stressors have been highlighted in many of the studies previously. The gap still exists in the area of technostress, and more is to be explored.

In view of the above, the objectives of this research are to explore and identify the techno-stressors impacting IT professionals in the work from home mode and deriving the weights of each stressor in order to prioritize them. Ranking of techno-stressors has been done using Analytical hierarchy process that is widely accepted method for multi-criteria decision analysis. This helps the decision support systems in structuring the potential stressors, assess their severity and take remedial action. This result may also provide support to the decision makers in order to assess ongoing or future decisions. For this purpose, four techno-stressors (techno-insecurity, techno-overload, techno-invasion and techno-complexity) originally classified by Tarafdar et al. (2007) and later widely used by the researchers for examining, prioritizing or comparing technology induced stress (Spagnoli et al. 2020) were selected. In this study, these four techno-stressors dimensions have been included because of their relevance to the present Covid-19 pandemic work settings where these techno-stressors have been highlighted as the more prevalent ones (Molino et al. 2020). The reason for selecting these techno-stressors is guided by the industry experts’ opinion as well in addition to the literature survey. During the process, telephonic discussions were held with selected executives of reputed IT organizations, psychologists and university level academicians in order to understand the distressing factors affecting most individuals in the present scenario of work from home.

Next, exploratory factor analysis followed by confirmatory factor analysis of the techno-stressors has been performed and analysed. At the end, the analytic hierarchy process (AHP), based on mathematics and psychology, is a structured technique for analysing complex decisions. AHP deals with defining decision making processes taking in to consideration decision makers’ perceptions, feelings, inputs, and judgements. It provides the decision makers an objective and logical structure to categorize a decision problem, for weighing its components (Karpak 2017; Saaty and Vargas 2012). The hierarchical model describes the criteria objects aligned with the main goal. Thus, it simplifies the whole decision making process by structuring the complex issue with multiple facets in a systematic way (Abdullah and Najib 2016).

At the outset, the aim of the study is to design techno-stressors framework based which would facilitate the policy makers, governments and decision makers to prioritize the stressors based on the weights assigned to them and later, decide upon the proposal of extending work from home post-Covid-19 pandemic. As government has lifted lockdowns in many parts of the world, employers are in huge discussion about resuming the regular office work culture. Therefore, in this context, the significant contribution of this study is the nature of the research technique adopted which is evidence based and can be replicated in different contexts and culture. Section 2 presents the literature review. Section 3 describes the used methodology and data. Section 4 presents and analyses the obtained results, and Sect. 5 concludes the paper.

Review of literature

‘Technostress’—stress caused by technology—has been defined as the negative psychological impact of technology on people. Technostress is an emergent psychological illness induced by individual’s lack of ability to adapt to new technology in an efficient and right way (Brod 1984). The model on technostress explains how IT create stressors which ultimately affect employees at workplaces (Ayyagari et al. 2011). Individuals depend on technologies, such as mobile and communication technologies, database technologies and social media technologies. Therefore, anything which causes technostress while working is an important consideration.

Use of ITC depends on both a user’s attitude towards the using technology and what he/she perceives about the benefits or advantages of using the same (Davis 1989). More categorically, perceived usefulness can be explained as the extent to which an individual perceives that utilizing computer technology would lead to more efficiency and productivity. However, some individuals may possess a negative impression of the utility of ICT, still they are required to use it at their workplace. This perceptual subjectivity provokes the attitude to work harder to complete more and more assignments beyond one’s capacity (Åborg and Billing 2003). The IT requirements also results in almost constant “connectivity” with frequent use of e-mail and the phone which demands the employees to be connected and available always. This leads to working beyond office hour where individuals are not able to control working hours, which creates feelings of being stressed out. This is one of the main reasons of dissatisfaction about the work which ultimately causes techno-invasion and triggers negative feelings towards technology (Weil and Rosen 1997). The techno-stressors related to ICT at the workplaces have been scientifically classified as: techno-overload (pressure to work much faster with extended hours), techno-invasion (constant connectivity causing blur boundary between professional and family life), techno-complexity (feeling of inadequate knowledge and skills related to ICT and pressure to learn and update), techno-insecurity (job insecurity due to new ICT upgradation or new ICT skilled person) and techno-uncertainty (difficulty to match with frequently changing ICT) (Tarafdar et al. 2007). Ayyagari et al. (2011) developed the P–E fit model of stress to describe computer-related stress. In sum, all the above-mentioned studies indicated individual’s way of interaction with technology and the subjective evaluation of the ICT eventually leads to psychological reaction to ICT. The ICT also leaves individuals in organizations with obsolete skills (Ayyagari et al. 2011). The rapid changes in technologies also create a problem of higher level of techno-insecurity. He also stated that the constant changes of technologies have negative impact on the perceptions of individual on the use of ICT as they have a pressure to be updated to face the competition.

IT projects assigned to the users are causing health related problems, as such the complexity of technology pose a particular challenge to users in acquiring new skills to handle such projects (Morris and Venkatesh 2010). Some specific intrinsic characteristics of IT jobs including frequent changes in the technology, unexpected demand of new skills, and pressing deadlines always create a fear of obsolescence of skills of the users. While organizations foster new IT advancements, the workforce in IT sector is on a learning curve showing reduced productivity (Sethi et al. 1999). Technological changes resulting in increased job demands have pressurized individuals to acquire new job skills to handle new set of work environment (Sami and Pangannaiah 2006). In order to complete the tasks, employees are adopting new technologies which creates a scenario of competition with a sense of techno-insecurity. Individuals, who are in the pressure of facing competition with fear of losing jobs, are experiencing more anxiety and frustration. There are many observed impacts of technostress such as lower job commitment, reduction in professional effectiveness, low output, more conflict and isolation and more employee turnover (Igbaria and Siegel 1992). There is also an impact of technostress creators on employee innovation based on the nature of each creator (Chandra et al. 2019). This requires an awareness on the adverse effect of technostress and effective measures and also implementation of desired practices and strategies for managing technostress.

After a review of literature, it was found that a lot of work has been done in identifying the causes of technostress and its association with productivity and performance. But more attention needs to be paid on the indicators of technostress on employee working from home in detail. The current scenario of lockdown imposed by the widespread coronavirus pandemic has taken this technostress to the next level. Though work from home has reported increased productivity and efficiency during pandemic, there are evidences of IT employees working longer than usual hours, mixing work and family time and facing anxiety and depression. In the light of these speculations, it is vitally important for the employers or decision makers to adopt a systematic approach to analyse the repercussion of technostress on IT professionals and reconsider the proposal of extending work from home post-pandemic. As Covid-19 pandemic has emerged as a rare phenomenon, its consequences on mental stress have not been adequately reported. Consequently, in the year 2020–2021 use of data science and mathematical approaches to develop decision support system during pandemic for decision makers has grown extensively (Sardar et al. 2020). Further, Pileggi (2020) has revealed that a systematic approach in which all outcomes (positive, neutral and negative) are taken for evaluation produces accurate results and leads to better decision making. Unlike Chatterjee and Shukla (2020), the present study applies AHP approach (specifically designed for multi-criteria decision making) to identify and rank the most important factors creating technostress during work from home so that they should be considered while deciding upon the implementation and extension of work from home during and post-lockdown, respectively.

Methodology

Data collection

For collecting data, researchers exercised their judgement to select the 17 IT companies of India which have moved to work from home policy due to outbreak of Covid-19 pandemic. Based on extensive review of literature and inputs taken from the IT industry experts, two sets of questionnaire were prepared in two different phases in order to meet the research objectives.

In first phase, a questionnaire of 25 items with a five-point Likert scale was prepared to go for exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) to identify the stressors. A total of 398 employees were approached randomly out of which only 334 questionnaires were usable after eliminating incomplete and missing data. Based on the EFA and CFA, four stressors were identified.

In the second phase, a second set of questionnaire was developed based on the fundamental scale of Saaty (1980) after identifying stressors and sub-stressors from first phase to develop pairwise comparison matrices. This set of questionnaire was administered to the respondents of first phase, and out of 334 respondents, 225 responses were recorded.

Research method

Tests for reliability and validity of self-developed questionnaire items were performed on responses of first phase of questionnaire. Techno-stressors were identified with EFA followed by CFA with all 25 items of first phase. Using these techno-stressors, sub-stressors were identified under each stressor based once again with EFA, CFA and expert views. Using these techno-stressors and sub-stressors as criteria and sub-criteria for analytic hierarchy process (AHP), weights are derived for each factor and sub-factor and then ranking is assigned. In AHP, weights of the criteria as well as sub-criteria have been derived using eigenvector method (EM) as principal eigenvector captures transitivity to obtain the correct ranking on a ratio scale of the alternatives (Saaty and Hu 1998).

When the element dominance properties are violated, linear programmes (LP) are formulated based on pairwise comparison matrix derived from AHP. A two-stage LP approach for generating a priority vector was used. On the first stage, a linear programme was formulated that provides a consistency bound for a pairwise comparison matrix. In the second stage, we use the consistency bound in a linear programme whose solution is a priority vector of alternatives. This method is known as LP-AHP approach (Chandran et al. 2005; Patel et al. 2016).

Results

Exploratory factor analysis (EFA)

Initially, to verify the factorability of the data for exploratory factor analysis, Kaiser–Meyer–Olkin (KMO) test was performed. The KMO value was 0.950 which is above the threshold value indicated the adequacy of the sample to perform exploratory factor analysis (Hair 2009). Finally, EFA was executed for all 25 items of measurement scale using Principal Component Analysis with varimax rotation to identify underlying factor structure and test the construct validity. All 25 items were considered in four stressors based on their loadings. Total four factors/stressors were extracted with factor loadings greater than 0.4 and eigenvalue greater than 1 (Nunnally 1994). These items are explaining 67.756 per cent of total variance. The factor loadings of each item are shown in Table 1.

Table 1 Factor loading, Cronbach's alpha (α) and KMO measures of sampling adequacy
Full size table

Confirmatory factor analysis (CFA)

CFA for techno-stressors

Confirmatory factor analysis confirmed how well the measured techno-stressors, i.e. techno-insecurity, techno-overload, techno-Complexity and techno-invasion (F1, F2, F3 and F4, respectively), represent the number of latent variables (f1 to f25). Structural equation modelling technique has been used to conduct CFA. Figure 1 shows the measurement model reflecting standardized estimates for all 25 variables. All are satisfying the minimum requirement except CFI (0.893) and average variance explained of techno-invasion (0.483) which was missing marginally by 0.007.

Fig. 1
figure 1

The measurement model developed in R

Full size image

The model fit indices are reported in Table 2.

Table 2 Model fit indices for criteria
Full size table

CFA for sub-stressor techno-overload

All 10 items under techno-overload were subjected to CFA, and model fit indices are shown in Table 3. All the parameters are satisfying the minimum requirements. Sub-stressors for techno-insecurity, techno-complexity and techno-invasion have 7, 4 and 4 items, respectively, for which we have gone for sub-stressors based on expert views. Sub-stressors with variables of these techno-stressors are shown in Table 4.

Table 3 Model fit indices for sub-criteria
Full size table
Table 4 Sub-criteria based on expert views
Full size table

Analytical hierarchy process (AHP)

Goal setting and developing hierarchical structure of criteria

Total 10 sub-stressors categorized under 4 main techno-stressors (techno-insecurity, techno-overload, techno-complexity and techno-invasion) derived in Sects. 4.1 and 4.2, have been included for conducting AHP. Figure 2 provides the list of criteria and sub-criteria of techno-stressors.

Fig. 2
figure 2

The hierarchical structure of techno-stressors (Criteria for AHP)

Full size image

Developing pairwise comparison matrices

Two criteria having a common parent are compared. The numerical representation of the comparison in the matrix form is known as pairwise comparison square matrix. In a matrix, for a set of \(n\) criteria, \(n\left( {n - 1} \right)/2\) comparisons are needed. All elements of the diagonal are 1 because on the diagonal, elements are compared with themselves. The remaining half of the judgements are reciprocals. The scale was proposed by Saaty (1980). The numbers 1, 3, 5, 7 and 9 represent the intensity of comparison with even numbers 2, 4, 6 and 8 being used for intermediate judgments. The elements of a pairwise comparison matrix are designated as \(a_{ij}\) where \(a_{ij}\) the geometric mean of the responses of respondents to minimize the effect of outliers.

Let \(w_{i}\) be the weight of ith criteria and \(w_{j}\) the weight of jth criteria, then we have

$$a_{ij} = {\raise0.7ex\hbox{${w_{i} }$} \!\mathord{\left/ {\vphantom {{w_{i} } {w_{j} }}}\right.\kern-\nulldelimiterspace} \!\lower0.7ex\hbox{${w_{j} }$}}\quad i,j = 1,2, \ldots ,n$$
(1)
$$A = \left( {\begin{array}{*{20}c} {w_{1} /w_{1} } & {w_{1} /w_{2} } & \ldots & {w_{1} /w_{n} } \\ {w_{2} /w_{1} } & {w_{2} /w_{2} } & \ldots & {w_{2} /w_{n} } \\ \vdots & \vdots & \vdots & \vdots \\ {w_{n} /w_{1} } & {w_{n} /w_{2} } & \ldots & {w_{n} /w_{n} } \\ \end{array} } \right) = \left( {a_{ij} } \right)$$

For the matrix A with eigenvalue n, the vector \(w = \left( {w_{1} {, }w_{2} ,{ } \ldots w_{n} } \right)\) is the principal eigenvector. The matrix \(A\) is consistent if \(Aw = nw\)

$${\text{If}}\;A\;{\text{is}}\;{\text{not}}\;{\text{consistent,}}\;{\text{we}}\;{\text{write}}\;Aw = \lambda_{\max } w,$$
(2)
$${\text{Where}},\quad \lambda_{\max } = \mathop \sum \limits_{i = 1}^{n} \mathop \sum \limits_{j = 1}^{n} a_{ij} w_{j}$$
(3)

Eigenvector method (EM)

An eigenvector vector of a linear transformation is a nonzero vector that changes at most by a scalar vector during application of that linear transformation. The corresponding eigenvalue is the factor by which eigenvector is scaled. Principal eigenvector of the comparison matrix \(A\) can be used as the desired priority vector. EM is based on solving the following equation.

$$Aw = \lambda w,\; e^{T} w = 1$$

The principal eigenvector \(\lambda_{\max }\) of A is determined after solving the characteristic equation,

$$\det \left( {A - \lambda_{\max } I} \right) = 0$$

where I denotes the identity matrix.

Then, using the value of \(\lambda_{\max }\), the eigenvector \(w = \left( {w_{1} , w_{2} , \ldots w_{n} } \right)\) is found out from the set of simultaneous linear equations:

$$\left( {A - \lambda_{\max } I} \right)w = 0$$

Consistency check

Consistency is checked to determine to what extent an assessment made by the decision makers is reliable. Consistency ratio (CR) is to measure how consistent the judgements are relative to large samples of random judgements. CR is calculated by dividing the consistency index of pairwise comparison matrix by the set of judgements by the index for the corresponding random matrix. Saaty suggested (1980) that if the ratio exceeds 0.1, the set of judgements may be too inconsistent to be reliable. The consistency index (CI), random consistency index (RI), and consistency ratio (CR) have been computed to check the inconsistency of comparison matrix using the formula proposed by Saaty (1980).

$${\text{CI}} = \frac{{\lambda_{\max } }}{n - 1},\;{\text{CR}} = \frac{{{\text{CI}}}}{{{\text{RI}}}}$$

where λmax is the principal eigenvector.

Table 5 represents the random consistency index values (RI) for different matrix sizes (n). Maximum acceptable limit for CR is 0.1. If it is greater than 0.1, then the decision assessment is considered to be inconsistent.

Table 5 Random consistency index (RI) values
Full size table

Table 6 presents the details of pairwise comparison matrix. Weights using eigenvector method (EM) are shown in the last column and consistency ratio in the last row.

Table 6 Pairwise comparison matrix with geometric mean
Full size table

In Table 6, it can be observed that \(a_{42} = 1.253\), which implies that the weight of fourth criteria (techno-invasion) be greater than or equal to weight to second criteria (techno-overload), which have been violated here. To obtain weights, without the violation of element dominance, LP-AHP approach discussed in previous studies by Chandran et al. (2005) and Patel et al. (2016) has been used.

Deriving weights for techno-stressors and their sub-factors

The local weights of other matrices have been calculated with eigenvector method and they did not violate element dominance property. The global weights and the local weights of techno-stressors are mentioned in Table 7. All comparison matrices are satisfying the consistency ratio less than 0.1.

Table 7 Weights, local weights, global weights and ranking of the main techno-stressors and associated sub-factors
Full size table

The global weight 0.134 of “pressure to upgrade skills” is calculated as 0.223 multiplied with the weight 0.611 of “techno-insecurity”, where “techno-insecurity” is the higher level of “pressure to upgrade skills”. Similarly, the global weight 0.038 of “work beyond office hour” is calculated as 0.287 multiplied with the weight 0.138 of “techno-overload”, where “techno-overload” is the higher level of “work beyond office hour”. In this way, the weights of all items have been derived.

Assigning rank to techno-stressors and their sub-factors

The numbers mentioned inside the bracket in the first and last column of Table 7 represents the ranking of each techno-stressor criteria. The ranks have been assigned to each techno-stressors based on their weights. “Techno-insecurity” (F1) is the most important dimension, while “techno-overload” (F2), “techno-complexity” (F3) and “techno-invasion” (F4) are of same importance. The top three most important indicators with respect to their global weights are “pressure to upgrade skills”, “threat from new recruits” and “fear of job loss due to new ICT” and all these fall under the criteria “techno-insecurity”. It implies that the higher the weight, more will be contribution to technostress. Thus, the most contributing factor to technostress in IT sector during work from home is “fear of Job loss due to new ICT” under the criteria “techno-insecurity” while the factor perceived by the employees as least contributing is found to be “work beyond office hour” under the criteria “techno-overload”.

Discussion

During coronavirus pandemic, “work from home (WFH)” policy proved to be a saviour for employers as well as employees. WFH was embraced all over the world especially in IT industry as a “win–win” strategy. Industrialists claimed to have increased employee productivity during remote working. But at the same time, WFH is taking a toll on employees’ mental health in the form of increased technostress, anxiety, work-overload and inability to separate work from their personal lives, violence, etc. (Piquero et al., 2020). Psychologists have also claimed that extensive use of ICT has adverse effect on employee’s mental health.

But, these types of virtual workplace arrangements reduce the employer’s overall cost and ensure high connectivity with contractors and other stakeholders as well (Spreitzer et al., 2017). Consequently, months after the lockdown was relaxed and removed in many countries, a heated debate on extending WFH post-pandemic has sparked. Gartner (2020) has also confirmed in his study that current Covid-19 situation is likely to trigger the WFH trend after the pandemic too.

Grounded on the above-mentioned fact, the purpose of the present study is twofolded. First, it provides the weighted framework of present techno-stressors prevalent due to WFH policy among IT employees. This will give a practical insight to the policy makers and employers to prioritize these technostress creators and take appropriate measures in order to reduce them. Second, due to the uncertain and vague nature of Covid-19 pandemic AHP methodology which is appropriate for multi-criteria decision making (Abdullah and Najib 2016) has been applied so that the weighted hierarchy model of techno-stressors would help them in designing better post-pandemic work policy. The present result that prioritizing the indicators/factors would help the policy makers and governments to design post-lockdown work strategy. Recent work on technostress by Spagnoli et al. (2020) has highlighted multi-tasking, pressure to acquire new skills, often ICT upgradation, technical glitches and consequent insecurities and threat related to job to be the most prominent technostress creators during pandemic.

Our finding shows that “techno-insecurity” having weight 0.601 has been perceived as the biggest stressor among employees in IT sector during work from home. In congruence with the findings of Spagnoli et al. (2020), fear of job loss due to new recruits, frequent ICT system upgrade and pressure to upgrade one’s own skills have been important indicators of techno-insecurity. Recent works have extensively reported how WFH has triggered fear of job loss among workforce which is not only because of the pandemic situation but also due to the rapid technological advancements that employees are not able to match up. Our findings are consistent with these evidences. Notably, dimensions like had equal rank in terms of prioritization. That does not mean they are irrelevant or less important. Clearly, in such complex pandemic situation, technostress creators under the dimension “techno-insecurity” need immediate attention while others are long-term priorities.

Conclusions

Working from home due to Covid-19 has transformed the whole economy with its industries and jobs. Managers and the policy makers are facing decision fatigue in this pandemic characterized by uncertain and complex nature. In the light of this, this study brings insights to form managerial implications for understanding the emerging phenomenon of the techno-stressors in connection with working from home and take tactical decisions later. Our finding shows that “fear of job loss due to new ICT” under the criteria “techno-insecurity” has been perceived as the biggest stressor among the employees of IT sector. Techno-insecurity is due to the reason that employees in IT sector feel immense pressure to upgrade their ICT skills to keep their jobs safe. These employees are even worried that new recruits may perform their jobs posing a threat to them. Moreover, the advancement in new ICT is posing the threat employees. Thus our evidence on the reasons associated with techno-insecurity has potentially important implications for the management in IT sector. While advances in technology, skill gap and new recruits are increasing the stress and anxieties among the employees, their effects could be offset by stronger policies in IT organizations. Further, post-pandemic strategy (e.g. WFH) must ensure stability and job security among the employees as a robust Covid-19 containment policy tends to decrease the unemployment rate (Tran et al. 2020). These policies can include transparency in hiring process, enhancing trust in management, increasing the sense of procedural fairness, raising employees’ efforts and achieving greater workforce flexibility. The existing work system in a working from home mode can be restructured and redesigned with appropriate collaboration of technology.

At the outset, the findings of this paper intend to provide an insight to the managers and policy makers in various ways. The managers can use this approach as an investigation tool to explore the degree to which each stressor affects the remote employees and to take necessary actions to deal with technostress in the present scenario. However, It is imperative to note that multi-criteria decision making methods indicates towards the possibility of best answer therefore, it should only be used as decision support tool not as the method to derive final solution. As, the scope of this study was limited to only four out of the five techno-stressors originally reported by previous researcher, in future, a more detailed and comparative study on technostress can be conducted in other sectors and technostress coping alternatives can also be discussed in greater detail. The findings of this study also guide the future researchers to replicate this study in different context.