Is There a Sampling Bias in Research on Work-Related Technostress? A Systematic Review of Occupational Exposure to Technostress and the Role of Socioeconomic Position

Abstract

Technostress is a widespread model used to study negative effects of using information communication technologies at work. The aim of this review is to assess the role of socioeconomic position (SEP) in research on work-related technostress. We conducted systematic searches in multidisciplinary databases (PubMed, PubMed Central, Web of Science, Scopus, PsycInfo, PsycArticles) in June 2020 and independently screened 321 articles against eligibility criteria (working population, technostress exposure, health or work outcome, quantitative design). Of the 21 studies included in the narrative synthesis, three studies did not collect data on SEP, while 18 studies operationalised SEP as education (eight), job position (five), SEP itself (two) or both education as well as job position (three). Findings regarding differences by SEP are inconclusive, with evidence of high SEP reporting more frequent exposure to overall technostress. In a subsample of 11 studies reporting data on educational attainment, we compared the percentage of university graduates to World Bank national statistics and found that workers with high SEP are overrepresented in nine of 11 studies. The resulting socioeconomic sampling bias limits the scope of the technostress model to high SEP occupations. The lack of findings regarding differences by SEP in technostress can partly be attributed to limitations in study designs. Studies should aim to reduce the heterogeneity of technostress and SEP measures to improve external validity and generalisability across socioeconomic groups. Future research on technostress would benefit from developing context-sensitive SEP measures and quality appraisal tools that identify socioeconomic sampling biases by comparing data to national statistics.

How Are Techno-Stressors Associated with Mental Health and Work Outcomes? A Systematic Review of Occupational Exposure to Information and Communication Technologies within the Technostress Model

“…The included studies contained considerable heterogeneity in the selection and analysis of ICT exposure and outcome measures. We found that when operationalising ICT exposures with the technostress model, technooverload was the most commonly examined techno-stressor in 17 studies (81% of studies), followed by techno-invasion in 16 studies (76%), techno-insecurity in 11 studies (52%), techno-complexity in 8 studies (38%) and techno-uncertainty in 7 studies (33%), as previously described elsewhere [15]. The exposure measures, sampled occupations, and reported outcome measures varied substantially.…”

“…This systematic review follows the PRISMA statement guidelines [14] and we registered a review protocol in the PROSPERO database (CRD42020199960). More detailed information on the search strategy can be found in our previous publication based on our research process [15]. This systematic review is the second paper to result from one review and data collection process.…”

“…The majority of studies considered at least one potential confounder, while seven studies did not describe any confounders. Furthermore, an additional analysis regarding potential sociodemographic differences in included study populations was carried out and reported separately with a focus on limitations to the generalisability of study findings and the role of socioeconomic position [15].…”

“…As shown in Appendix A, a majority of six studies were conducted in the USA and fourteen were published in journals for Information Systems, i.e., ICT-oriented business research [15]. The 16 convenience samples were collected in diverse organisations with a range from 152 to 14,757 employees.…”

“…Similarly, some potentially positive implications of technostress for work engagement may also be influenced by socioeco-nomic factors, although the findings have been inconclusive. Potential sociodemographic differences regarding specific techno-stressors have been suggested elsewhere [15].…”