In plain language
Most research treats problematic internet use as uniformly harmful: on average, people who report more compulsive or avoidance-driven internet use also report lower well-being. But does that average describe every individual? This study asked whether the moment-to-moment link between problematic internet use and mood is actually the same for everyone, using an idionomic (within-person-first) design.
Over 10 days, 84 young adults (average age 23.5) completed five short surveys a day, reporting on four problematic internet behaviors (two reflecting time control and compulsivity, two reflecting emotion dysregulation and avoidance) and six mood states. The researchers first modeled each person individually using time-series (ARIMAX) analysis, then combined those individual results with random-effects and Bayesian meta-analysis, and finally explored subgroups with clustering methods (partitioning around medoids and growing self-organizing maps).
On average, when people were above their own usual level of problematic internet use, their mood was slightly lower. But the average hid enormous variation: some people felt clearly worse when their internet use spiked, others showed no change, and some actually felt better in the moment. The findings argue for measurement-based, person-specific intervention: for people whose internet use lifts mood now but may carry later costs, practitioners might emphasize delay-of-gratification strategies; for those whose mood drops immediately, awareness and functional-mapping approaches may be more useful.
Key findings
- Within-person, being above one's own baseline of problematic internet use was associated with slightly lower momentary mood on average (correlations ranging from −0.12 to −0.04 across behaviors).
- Bayesian multilevel meta-analysis found a small pooled negative effect (β = −0.06, 95% CI [−0.11, −0.02]) but large between-person variability (SD = 0.18): individual effects were estimated to range from about −0.42 to +0.30.
- Cluster analysis (partitioning around medoids) identified one group (n = 50) for whom problematic internet use had a neutral-to-slightly-positive relationship with mood, and another (n = 34) for whom it was clearly linked to poorer mood.
- Growing self-organizing map analysis revealed finer-grained subgroups, including people for whom avoidance-driven or compulsive internet use was linked to more positive mood, not less.
- Between people, habitual problematic internet use was consistently associated with lower average mood (medium-sized correlations, e.g., preferring the internet over people correlated with feeling bad, r = −0.31), showing that group-level harm can coexist with highly variable individual-level effects.
- The results support tailoring interventions to each person's own internet-use/mood pattern rather than applying a one-size-fits-all approach.
How to cite
APA
Ciarrochi, J., Hernández, C., Hamed, E., Alahakoon, D., Adikari, A., Yap, K., Ranapanada, I., Hayes, S. C., Fraser, M. I., & Sahdra, B. (2026). Is problematic internet use always problematic? An experience-sampling study of compulsive and avoidance-driven internet behaviors and momentary mood. Psychology of Addictive Behaviors, 40(4), 422–431. https://doi.org/10.1037/adb0001142
BibTeX
@article{ciarrochi2026is,
title = {Is Problematic Internet Use Always Problematic? An Experience-Sampling Study of Compulsive and Avoidance-Driven Internet Behaviors and Momentary Mood},
author = {Ciarrochi, Joseph and Hern{\'a}ndez, Crist{\'o}bal and Hamed, Eman and Alahakoon, Damminda and Adikari, Achini and Yap, Keong and Ranapanada, Isuru and Hayes, Steven C. and Fraser, Madeleine I. and Sahdra, Baljinder},
journal = {Psychology of Addictive Behaviors},
year = {2026},
volume = {40},
number = {4},
pages = {422--431},
doi = {10.1037/adb0001142}
}
Related work
- All publications by Joseph Ciarrochi (searchable, with free PDFs)
- Process-Based Therapy & Idionomic Analysis
Author: Joseph Ciarrochi (ORCID 0000-0003-0471-8100). Free copy hosted with permission for scholarly use. Please cite the published version via the DOI above.