Connectivity in crisis: the contrasting roles of mobile and non-mobile Internet on subjective well-being during the COVID-19 pandemic

The study investigates how different modes of Internet access—mobile (smartphones/tablets) versus traditional non-mobile (PCs/laptops)—affect individuals’ subjective well-being, and how these effects change during societal disruptions such as the COVID-19 pandemic. Motivated by the rapid global expansion of Internet use and the pandemic-driven surge in online activity, the paper addresses three questions: (1) How does the impact of mobile Internet on subjective well-being compare with traditional non-mobile Internet? (2) How do these impacts change during social disruption (COVID-19)? (3) How do mobile and non-mobile Internet interact in shaping subjective well-being? Grounded in uses and gratifications theory, the authors argue that although both Internet types satisfy overlapping gratifications (content, process, social), their differing capabilities and usage contexts likely yield heterogeneous effects on well-being, particularly under conditions of uncertainty and isolation generated by the pandemic.

The paper integrates uses and gratifications theory (UGT) to explain media choice and outcomes, emphasizing that users actively seek gratifications—content/information, process/utility, and social connection—from technologies. Prior work has shown the Internet can enhance well-being by facilitating communication, information access, and varied activities. However, mobile versus non-mobile Internet differ in technical capabilities and personalisation: traditional Internet offers richer resources and multimedia performance, while mobile affords portability, immediacy, and personal attachment. Research during COVID-19 documents increased online time and a shift toward gratifications addressing stress, uncertainty, and social isolation. Mobile devices, viewed as more personal and identity-linked, can provide psychological comfort and a sense of control, potentially buffering stress during disruptions. The review also notes potential downsides of extensive multichannel online engagement—greater time displacement, reduced offline interactions, and dependence—associated with poorer mental health outcomes. Hypotheses derived include: H1, non-mobile Internet has a stronger positive effect on subjective well-being than mobile Internet; H2, during disruptions (COVID-19), the impact of mobile Internet on well-being strengthens relative to non-mobile; H3, mobile and non-mobile Internet use substitute rather than complement each other in their effects on well-being; H4, this substitution effect diminishes during societal disruptions.

Data come from the China Family Panel Studies (CFPS), a nationally representative longitudinal survey conducted biennially. The analysis uses three waves (2016, 2018, 2020), with 2020 data collected in June–July 2020 after the COVID-19 outbreak. A balanced panel was constructed including respondents with non-missing key variables, yielding 46,803 person-wave observations from 15,601 individuals. Key independent variables are binary indicators: Mobile equals 1 if the respondent used mobile devices (smartphone/tablet) to access the Internet; Non_Mobile equals 1 if the respondent used traditional devices (PC/laptop). The dependent variable, WellBeing, is derived from the CES-D depression scale; scores were inversely transformed and normalized so higher values indicate better well-being. Controls include age, female, higher education, marital status, employment status (employed, unemployed, exited labor market, other), relative income (five-point scale), rural residence, province fixed effects, and a COVID dummy equal to 1 in the 2020 wave. Panel regression with random effects is used. The baseline model regresses WellBeing on Mobile and Non_Mobile with controls. To assess pandemic effects, the model includes COVID and interactions with Mobile and Non_Mobile. To examine the relationship between Internet types, the interaction Mobile × Non_Mobile is added; due to potential collinearity with higher-order terms, subgroup analyses (pre- vs post-COVID) further assess the substitution effect.

• Both mobile and non-mobile Internet use are positively associated with subjective well-being, but traditional non-mobile Internet has a larger effect. In baseline estimates, non-mobile: b ≈ 0.010 (p < 0.001); mobile: b ≈ 0.003 (p ≈ 0.048). After transforming CES-D scores, mobile Internet use corresponds to an average reduction of ~0.210 CES-D points, and non-mobile by ~0.541 points. Confidence intervals further confirm the larger effect of non-mobile use (mobile 90% CI ≈ [0.001, 0.006]; non-mobile 90% CI ≈ [0.007, 0.014]), supporting H1. • The COVID-19 pandemic had a negative main effect on well-being, but the effect of mobile Internet on well-being increased during COVID-19 (interaction with COVID significant and positive in interpretation; narrative b ≈ -0.011 with sign conventions tied to transformed scale). The effect of non-mobile Internet does not significantly change during COVID-19. This supports H2. • Mobile and non-mobile Internet exhibit a substitution relationship in their effects on well-being: the interaction Mobile × Non_Mobile is negative and statistically significant (b ≈ -0.010, p = 0.003), indicating no synergistic benefit from concurrent use, supporting H3. • Subsample analyses show that the substitution effect weakens during the pandemic: pre-COVID interaction is marginally negative (b ≈ -0.012, p = 0.058), while post-COVID it becomes null (b ≈ -0.000, p = 0.995), supporting H4. • Descriptives: WellBeing mean 0.753 (SD 0.153) on the transformed scale; Mobile users 52.5% of observations; Non-Mobile users 21.8%. Model fit (overall R-sq ~0.07) is consistent across specifications.

The findings address the core questions by demonstrating that while both mobile and non-mobile Internet use correlate with higher subjective well-being, the traditional non-mobile mode generally confers greater benefits in typical times. During COVID-19, mobile Internet’s role becomes particularly salient, consistent with UGT’s notion that gratifications sought shift under stress and uncertainty. The portability and personalization of mobile devices likely provide psychological soothing, a sense of control, and private space that mitigate the pandemic’s negative effects on well-being. The observed substitution between modes suggests that engaging with both types does not yield added well-being benefits and can be associated with time displacement and potential psychosocial downsides. However, this substitution weakens in crisis conditions, likely because mobility restrictions and social isolation increase the value of any online engagement for emotional support and connection, thus buffering the typical trade-offs. Overall, technology’s contribution to well-being is context-dependent, with mobile connectivity offering unique protective effects during societal disruptions.

This study advances understanding of how different forms of Internet access relate to subjective well-being and how these relationships evolve during crises. Using a large, nationally representative longitudinal panel from China, the study shows that non-mobile Internet use generally has stronger positive associations with well-being than mobile Internet use. Yet in times of social disruption like COVID-19, mobile Internet’s protective role becomes more pronounced, and the usual substitution between modes weakens. The work contributes theoretically by applying a dynamic uses and gratifications lens, highlighting psychological soothing as a distinct gratification provided by mobile Internet during crises. Practically, it underscores the potential of mobile technologies to mitigate well-being losses during disruptions and informs policies aiming to leverage mobile access for mental health support. Future research should examine usage intensity and device granularity (e.g., smartphones vs tablets, desktops vs laptops), incorporate broader well-being measures beyond depressive symptoms, and validate findings across diverse cultural and national contexts.

The study measures Internet use with binary indicators rather than intensity or time spent, limiting insights into dose–response relationships. Mobile and non-mobile categories aggregate heterogeneous devices (e.g., smartphones versus tablets; desktops versus laptops), potentially obscuring device-specific effects. Well-being is proxied solely via the CES-D; results may differ for other subjective well-being dimensions. The data are from China, and cultural, economic, and infrastructural differences may limit generalizability; cross-country replications are needed.