Nassen et al. (11) proposed a definition of digital disconnection through a systematic review: “Digital disconnection is a deliberate (i.e., chosen by the individual) form of non-use of devices, platforms, features, interactions, and/or messages that occurs with higher or lower frequencies, and for shorter or longer periods of time, after the initial adoption of these technologies, and with the aim of restoring or improving one’s perceived overuse, social interactions, psychological well-being, productivity, privacy and/or perceived usefulness”. This definition not only encompasses the specific behavioral strategies individuals adopt but also deeply reveals the underlying motivations, providing a comprehensive perspective that lays a solid foundation for subsequent research.

To date, most research on digital disconnection has examined only one behavior at a time, primarily by evaluating the effects of interventions, such as using digital detox apps (12), blocking online distractions (13), or batching smartphone notifications (14). While these studies contribute to understanding the implications of partial disconnection, employees can engage with multiple ICTs disconnections in various ways, making it imperative to adopt a more comprehensive approach to measuring disconnection behaviors. Therefore, Verlinden et al. (10) developed the EDDS. This scale incorporates employees into the study of digital disconnection, measuring their intentional disconnection behaviors rather than the degree of disconnection. By combining the content of disconnected ICTs (work-related and private) with the scenarios of disconnection behaviors (during work and outside work), they identified four contexts of employee digital disconnection: digital disconnection from private ICTs outside work (DD POW), digital disconnection from work-related ICTs outside work (DD WOW), digital disconnection from private ICTs during work (DD PDW), and digital disconnection from work-related ICTs during work (DD WDW). Verlinden et al. (10) argued that the first context (DD POW) is not unique to employees but applies to any media user, thus focusing on the latter three contexts and developing the EDDS for these three scenarios. However, the context of DD POW should also be included in research. Although Verlinden et al. (10) did not address it on the grounds that this context “is not specific to employees,” this precisely indicates that employees commonly face such disconnections. Therefore, to understand employees’ digital disconnection behaviors more comprehensively, it is crucial to develop measurement tools that encompass this dimension.

The EDDS developed by Verlinden et al. (10) captures employees’ engagement in digital disconnection, advancing the field. However, this scale has several limitations. The participants were mainly from the public sector, which restricts the generalizability of the findings. Additionally, the participants were from Europe, where Western countries exhibit individualistic tendencies, whereas the collectivist culture in China emphasizes that individual interests are subordinate to collective interests. People will experience stronger social expectations (15), so it may be difficult for employees to achieve digital disconnection. Existing studies have shown that national culture can significantly influence employees’ behavioral choices (16). Therefore, exploring the manifestations of employees’ digital disconnection behavior in different cultural contexts is crucial for forming context-specific management approaches.

Scholarly research on digital disconnection has been conducted within the Chinese context (17), indicating that digital disconnection is a significant phenomenon of interest in China. Therefore, this study was conducted among Chinese employees without restricting participants’ industries to increase sample diversity.

Digital disconnection has certain distinctions and connections with technology non-use, media resistance, digital detox, and digital minimalism.

This study employed a questionnaire survey method and collected data online from employees in Guangdong Province, Jiangsu Province, Shandong Province, Zhejiang Province, and other provinces. The sample was diverse, covering not only the public sector but also industries such as manufacturing, real estate, construction, software, and commercial services, ensuring industry diversity and representativeness. To improve response rates and data quality, participants were compensated for completing the questionnaire. The survey was distributed twice, with 500 questionnaires each time. After excluding invalid responses (e.g., short completion time, patterned answers), sample 1 included 400 valid questionnaires (80% valid response rate), and sample 2 included 395 valid questionnaires (79% valid response rate). Sample 1 was used for item analysis and exploratory factor analysis, sample 2 was used for confirmatory factor analysis, and the entire dataset was used for reliability analysis. The demographic characteristics are shown in Table 1.

In this study, SPSS 27.0 was used for item analysis, exploratory factor analysis, reliability analysis, and criterion-related validity analysis, and Mplus 8.3 was used for confirmatory factor analysis. Controlling for the effects of an unmeasured latent methods factor (ULMC) indicated that there were no severe common method biases in this study (∆RMSEA = 0.002, ∆SRMR = 0.008, ∆CFI = 0.011, and ∆TLI = 0.005).

Conduct item analysis on the data of sample 1. Item analysis was performed via item-total correlation and the extreme group method. First, the correlation coefficient between each item score and the total score was used as an indicator for item analysis. The results revealed that the correlation coefficients between all 40 items and the total score were significant (p < 0.01). Second, the total questionnaire scores were ranked, and the highest and lowest 27% were divided into high and low groups. An independent-samples t-test was used to compare the differences between each item in the high and low groups. The results indicated that the t-values ranged from −21.733 to −2.408, and all the values reached significance levels (p < 0.05), as shown in Table 2.

Exploratory factor analysis (EFA) was conducted on sample 1 data (n = 400). On the basis of the results of the item analysis, EFA was performed on the 40 items retained in sample 1. The relevant indicators (KMO = 0.901, χ² = 10850.779, df = 4,780, and p < 0.001) indicate that the data are suitable for EFA.

Principal component analysis and the varimax method were used for analysis. The results revealed that Factor 5 was composed of items 9 and 10 from four dimensions. Because items 9 and 10 in the four contexts have similar wording (both involving “notify” or “I think it’s important”), they naturally aggregated into one factor in the EFA. According to the division principle of the four contexts in the original EDDS (work/nonwork time and work-related/private content) (10), this new factor does not belong to any of the four contexts. Therefore, items 9 and 10 from all four dimensions were removed, resulting in 8 deleted items. EFA was conducted again, following deletion criteria requiring factor loadings above 0.4, at least 3 items per factor, and cross-loading differences not less than 0.20. Items DD WDW8 and DD POW5 were sequentially removed, leaving 30 items. Four common factors with eigenvalues of 9.637, 5.021, 2.739, and 2.300 were extracted, and the cumulative variance contribution rate was 65.658%. The remaining items had loadings between 0.633 and 0.857. After EFA, the Chinese-EDDS has 30 items across four dimensions: factors 1 to 4 correspond to DD PDW, DD WOW, DD POW, and DD WDW. The specific data are presented in Table 3.

Using data from sample 2 (n = 395), a confirmatory factor analysis (CFA) was conducted. The CFA fitting result of the 4-factor model is better than that of the 8-factor model (see Table 4). This indicates that the 4-factor model is more suitable for China. This might be caused by cultural differences and the degree of sample diversity. The research of EDDS mainly used samples from the European public sector (10). The samples we used were not limited by industry and were more diverse.

Reliability analysis revealed that the Cronbach’s α for the total Chinese-EDDS was 0.925, with the Cronbach’s α for each dimension ranging between 0.875 and 0.944. The Cronbach’s α of the DD POW factor, the DD WOW factor, the DD PDW factor, and the DD WDW factor were 0.888, 0.938, 0.944, and 0.875, respectively, as shown in Table 5.

Using sample 2 (n = 395) for convergent validity and discriminant validity analysis, as shown in Table 6, the composite reliability (CR) of the four-factor model was greater than 0.7, and the average variance extracted (AVE) was greater than 0.5, indicating that the Chinese-EDDS has good convergent validity. The square root of the AVE for each factor was greater than the correlation coefficient values between each pair of factors, demonstrating good discriminant validity of the scale.

Job autonomy was used as the criterion tool. As shown in Table 7, the total score of the Chinese-EDDS was significantly positively correlated with the criterion tool, and the scores of each dimension of the Chinese-EDDS were also significantly positively correlated with the criterion tool. All correlation coefficients ranged between 0.105 and 0.318, which aligns with the expected hypothesis, indicating that the Chinese-EDDS has good criterion-related validity.

Sample 2 (n = 395) was used to conduct cross-group measurement equivalence testing for gender (male = 0, female = 1), and the model fit indices for each model are shown in Table 8. The results indicate that the configural invariance model, weak invariance model, and strong invariance model of the Chinese-EDDS all have good fit indices (∆CFI ≤ 0.01, ∆TFI ≤ 0.01, and ∆RMSEA≤0.01). This scale exhibits cross-group measurement invariance.

First, this study further clarifies the distinctions and connections between digital disconnection and related concepts. By sorting out the differences and connections between digital disconnection and technology non-use, media resistance, digital detox, and digital minimalism, its theoretical boundaries are further clarified. Unlike concepts such as “digital detox,” which presupposes that ICTs inevitably have negative effects (20), digital disconnection assumes that ICTs are neutral, with a focus on the dynamic interaction between humans and technology. This characteristic is reflected in the measurement items across the four dimensions. Whether it involves technical restrictions for private or work communications during work or outside work (such as turning off notifications or enabling airplane mode) or proactive communication behaviors (such as informing others of unavailability) reflects employees’ autonomous adjustment strategies in different contexts without entirely negating technology. This perspective breaks through the premise that ICTs inevitably have negative effects on users, revealing the dual-path mechanism (11) through which digital disconnection may simultaneously trigger both positive and negative impacts, thereby providing a clear framework for a deeper understanding of the theoretical uniqueness and complexity of this construct.

Second, in this study, a four-dimensional theoretical structure of employee digital disconnection in the Chinese context is constructed and validated. This structure not only identifies the cross-context of work/nonwork time and work-related/private content but also reveals that digital disconnection is not a vague one-dimensional concept but a context-driven, multilayered construct (10). This provides a theoretical framework and measurement foundation for future research, enabling in-depth exploration of the differentiated antecedents and consequences of digital disconnection in various contexts, thereby promoting the transition from qualitative to empirical research in this field.

Third, this study provides a comprehensive perspective on digital disconnection in the Chinese context, presenting a critical addition to existing research focused mainly on the Western context. Nassen et al. (11) clarified the concept of voluntary digital disconnection, but existing studies are predominantly based on Western contexts (30); thus, the related discussions in the Chinese context are insufficient. Given China’s collectivist culture, this study, which is grounded in the local context, analyzes four scenarios of employee digital disconnection, urging domestic academics to focus on and explore this topic further. Additionally, on the basis of Verlinden et al.’s (10) EDDS, a Chinese-EDDS was developed and validated in the Chinese context in response to calls for localized tool development (7, 17) and the use of diverse samples, providing reliable measurement foundations for empirical research in China.

Finally, digital well-being is a concept that emphasizes the importance of mental, physical, and emotional health in individuals’ interactions with technology (31). It seeks to mitigate the adverse effects of technology while fostering a more balanced and effective utilization of technological tools in social contexts (32). Recent research indicates that digital disconnection is not an isolated act but a crucial part of a series of strategies for achieving digital well-being (33). Therefore, the revised scale in this study goes beyond merely measuring the “digital disconnection” behavior itself, and has become a key measurement tool that connects the micro-individual behavior with the macro-health construct.

First, organizational managers should pay attention to employees’ digital disconnection behaviors. When employees actively engage in digital disconnection, managers need to focus on granting them greater job autonomy, allowing them to flexibly and independently allocate work time and attention according to their own rhythms and needs, thereby achieving a healthier and more sustainable work environment.

Second, organizational managers should emphasize the technological application and development of employees’ digital disconnection. By providing tools with multiple self-configurable disconnection modes (such as the work mode, rest mode, or deep focus mode), employees can autonomously choose the intensity and frequency of digital connectivity in various scenarios while limiting the scope and quantity of available functions. This promotes the evolution of the employee-technology-society-information ecosystem in a “people-oriented” sustainable development direction.

Finally, organizations should foster a cultural atmosphere that supports employees’ digital disconnection. Given China’s collectivist cultural background, employees often experience disconnection anxiety because of fears of detaching from the collective. Therefore, organizations should actively develop a culture that understands and respects individual digital boundaries. Managers can publicly affirm the value of reasonable digital disconnection and clearly delineate the boundaries between work hours and private time. For example, managers advocate that employees avoid sending work messages through tools such as WeChat Work or DingTalk outside work or explicitly inform employees that they do not need to reply to work messages during their private time. This cultural shaping not only helps alleviate employees’ anxiety and guilt associated with digital disconnection but also enhances their sense of identification with and belonging to the organization.

The study is not without limitations. First, the data come from employee self-assessments, so there is a potential social desirability bias, posing a risk of common method bias. Future research could incorporate measurement methods using objective data (such as the usage time of different software, the number and types of notifications closed or silenced, etc.) to further enhance the reliability and comprehensiveness of the study findings. Additionally, we relied on cross-sectional data, but to clarify the long-term consequences of digital disconnection, longitudinal research is needed.

Second, this study was conducted solely within the Chinese context. Influenced by the national cultural background and the institutional environment, the manifestation dimensions of employees’ digital disconnection and their psychological mechanisms may have regional specificity, and the cross-cultural generalizability of the research conclusions remains to be tested. Future studies could select countries or regions with distinct cultural characteristics for cross-cultural comparisons, further validating the applicability and robustness of the scales developed in this study. This would expand the external validity of digital disconnection theory and provide an empirical foundation for constructing a more universally explanatory theoretical framework.

Third, approximately 61% of the two samples in this study were female. This imbalance in proportion may be caused by multiple factors. The main reason might stem from the systematic recruitment bias of the online survey panel. While this distribution incidentally mirrors that of the original validation study (10), enhancing comparability, it may limit the generalizability of our findings to male employee populations. Nonetheless, we proactively addressed this concern by conducting a measurement equivalence analysis across genders. The results confirmed configural, weak, and strong invariance, indicating that the factor structure and the meaning of the latent constructs are equivalent between males and females. This provides strong evidence that the observed relationships among variables are not an artifact of the uneven sample split, thereby substantially mitigating validity concerns related to the gender imbalance.

Finally, because digital disconnection is a newly introduced construct, its theoretical system is still immature. The causes of digital disconnection are complex and diverse, and scholars are divided on the results of digital disconnection (11). Therefore, efforts to clarify the antecedent (e.g., personal characteristics, job characteristics, and organizational factors) and outcome variables (e.g., employee mental health and individual thriving) of digital disconnection and reveal its underlying mechanisms will contribute to the development of the digital disconnection field.