Flexible digital work carries a hidden strain cost, study of Chinese employees finds

Digital technologies have loosened many of the temporal and spatial limits that once defined where work could be done and when it should end, and the evidence on what this means for employees points in more than one direction. A new study from China suggests that the same flexibility that lets employees escape the office can be associated with two opposing psychological experiences at once: a genuine sense of control over their work, and a measurable rise in technology-related strain. The research, published in Frontiers in Psychology, analyzed survey responses from 462 full-time employees across Jiangxi Province and found that digital work flexibility was positively associated with both digital autonomy and technostress, two experiences that pulled employee well-being in opposite directions.

The study, conducted by Weina Liao of the School of Economics and Management at Nanchang University, was designed around a deceptively simple question. If flexible digital arrangements give people more discretion, why does the evidence on their benefits remain so mixed? The answer, according to the statistical model, is that a single overall coefficient can hide two meaningful pathways running in opposite directions. When both were estimated simultaneously, the positive and negative indirect relationships were nearly equal in magnitude and almost canceled each other out, leaving the apparent overall association between flexibility and well-being small and statistically indistinguishable from zero.

To untangle these pathways, the study drew on two theoretical frameworks. Job Demands-Resources theory distinguishes between job resources that support goal achievement and functioning, and job demands that require sustained effort and produce strain when excessive. The autonomy paradox, documented in earlier research on mobile and digitally mediated work, describes how greater discretion can coexist with stronger expectations of constant accessibility. Together, these perspectives allowed the researcher to treat digital work flexibility not as an inherently beneficial resource, but as an arrangement that may be associated with both resource-related and strain-related experiences at the same time.

A crucial conceptual distinction underpins the analysis. Digital work flexibility was defined narrowly as employees’ perceived discretion over when and where work is performed, measured with two items adapted from a well-established operationalization. Digital autonomy, by contrast, captured experienced control over work decisions and work methods, measured with six items from the Work Design Questionnaire covering decision-making and work-methods autonomy. Scheduling autonomy was deliberately excluded to avoid overlapping with the temporal component of flexibility. This separation matters because having freedom over your schedule does not necessarily mean having a say in how the work itself gets done.

The survey was conducted between June and December 2025 through the online platform SOJUMP, with the Jiangxi Association of Small and Medium Enterprises helping reach eligible participants. Of 818 employees invited, 539 returned questionnaires, and after screening for eligibility failures, missing data, unusually short completion times, straight-lining, failed attention checks, and duplicate submissions, 462 usable responses remained. Only one employee per organization was retained, producing a sample drawn from 462 organizations. Respondents were 52.2 percent women, most commonly aged 31 to 40, and 43.1 percent reported hybrid work arrangements, with the remainder mainly on-site or mainly remote.

The structural analysis used partial least squares structural equation modeling with 10,000 bootstrap resamples. Digital work flexibility was positively associated with digital autonomy (standardized coefficient 0.400) and also positively associated with technostress (0.316), the latter captured by a 16-item scale covering anxiety, fatigue, skepticism, and inefficacy. Digital autonomy was in turn positively associated with employee well-being (0.346), a higher-order construct combining life, workplace, and psychological well-being, while technostress was negatively associated with it (−0.425). The indirect relationship through digital autonomy was positive (0.139), the indirect relationship through technostress was negative (−0.135), and their net sum was not significant.

The most striking findings concerned after-hours availability expectations, defined as employees’ perceptions that their supervisors expect them to remain reachable and responsive beyond normal working hours. This perception acted as a boundary condition that shifted the balance between the two pathways. The interaction between flexibility and availability expectations was negative for digital autonomy (−0.123) and positive for technostress (0.131). In practical terms, the positive link between flexibility and autonomy weakened as perceived availability pressure rose, falling from 0.523 at low expectations to 0.277 at high expectations, while the positive link between flexibility and technostress strengthened, rising from 0.185 to 0.448 across the same range.

The conditional indirect effects told a consistent story. At low perceived availability expectations, the positive autonomy pathway reached 0.181 while the negative technostress pathway was only −0.079, producing a positive net indirect association of 0.102. At high expectations the pattern reversed: the autonomy pathway shrank to 0.096 while the technostress pathway deepened to −0.190, yielding a net indirect association of −0.094. The conditional total association between flexibility and well-being was not statistically distinguishable from zero at low or mean levels of availability expectations but was negative at high levels (−0.141). The author emphasizes that this describes a conditional pattern of associations rather than a simple positive-to-negative reversal of the overall relationship, and that the cross-sectional design means these are statistical associations, not established causal mechanisms.

Several limitations temper the conclusions. All constructs were self-reported at a single time point, so reverse relationships remain plausible; employees with lower well-being might, for example, perceive more technostress or less autonomy. A marker-variable procedure using an unrelated attitude measure, attitudes toward the color blue, reduced but did not eliminate concerns about common method variance, and re-estimating the model with demographic controls left the focal coefficients essentially unchanged. The availability-expectations measure captured perceptions of supervisor expectations, not actual after-hours connectivity behavior, message volume, or organizational policies, none of which were directly observed.

For organizations navigating the expansion of remote and hybrid work, the practical message is that formal flexibility policies alone may not capture employees’ day-to-day experience. The same arrangement can coincide with perceived empowerment or with perceived pressure to remain reachable, depending on the expectations surrounding it. The author suggests that clearer communication about urgency, expected response windows, and the legitimacy of disconnecting after hours may be promising managerial practices, while cautioning that such interventions were not tested in this study and require direct evaluation. Future research using longitudinal, experience-sampling, or supervisor-employee dyad designs is needed to establish whether the competing autonomy and strain pathways unfold over time as the conditional model proposes.

Subject of Research: Competing autonomy and technostress pathways linking digital work flexibility to employee well-being

Article Title: When flexibility feels like control: competing autonomy and Technostrain pathways linking digital work flexibility to employee well-being

Article References: Liao, W. (2026). When flexibility feels like control: competing autonomy and Technostrain pathways linking digital work flexibility to employee well-being. Frontiers in Psychology, 17, Article 1977600. https://doi.org/10.3389/fpsyg.2026.1977600

Image Credits: AI Generated

DOI: 10.3389/fpsyg.2026.1977600

Keywords: digital work flexibility, digital autonomy, technostress, employee well-being, after-hours availability expectations, autonomy paradox, job demands-resources theory, PLS-SEM, remote work, hybrid work, organizational psychology, China