Addictive use of social media scales are not necessarily measuring problematic behavior: Evidence from the international 2017 Health Behavior in School-Aged Children (HBSC) study

Maurage, Pierre;Maurage, Pierre;Maurage, Pierre;Billieux, Joël
(2026) Journal of Psychopathology and Clinical Science — Vol. 1, n° 1, p. 1 (2026)

Files

Ciudad_JPCS_2026_Abstract.docx
  • Open Access
  • Microsoft Word XML
  • 18.93 KB
Download

Details

Authors
Abstract
Background and aims The dominant problematic social media use (PSMU) screening questionnaires rely on criteria derived from substance use disorders and assess a unitary construct. This approach generates a single total score for classification (e.g., pathological vs. recreational), also used to determine PSMU prevalence. Recent evidence questions the validity of such unidimensional model, suggesting that current scales may conflate genuinely problematic (“core” criteria) with merely intensive but non-problematic involvement (“peripheral” criteria), potentially inflating prevalence rates and pathologizing social media use. Here we re-analyzed the structural validity and dimensionality of one of the most widely used questionnaires evaluating PSMU, the Social Media Disorder Scale (SMD), comparing three different theory-based latent structures for this highly used screening questionnaire. Design: Secondary analysis of data from the cross-sectional 2017 international Health Behavior in School-Aged Children (HBSC) study. Setting: Schools from 43 regions from 40 different countries Participants: 190,335 students (48.13% males; 51.87% females) with a mean age of 13.58 years (SD = 1.62). Measurements: Participants completed the SMD to measure PSMU, the Symptom Checklist to measure psychological problems and sleep problems, and the Cantril Ladder to measure life satisfaction. Findings: Across most regions, a two-factor model that separates salience and tolerance as peripheral criteria from all other SMD items as core criteria provided a superior fit compared to a unidimensional solution. Configural invariance CFA corroborated the distinction between core and peripheral (χ²(1118) = 7,692.318, p < .001, CFI = .981, TLI = .974, RMSEA = .036, SRMR = .038), surpassing the one-factor model (χ²(1161) = 8,815.970, p < .001, CFI = .978, TLI = .971, RMSEA = .038, SRMR = .042). Alignment optimization showed that 23% of the parameters were invariant across regions under this two-factor solution. Weak positive associations emerged between the core factor and psychological, sleep problems and life satisfaction, while the peripheral factor was negligibly or non-significantly associated with these outcomes. Conclusions: This study challenges the longstanding use on unidimensional scoring of PSMU measures. There is a need to refine screening instruments in order to distinguish core features, directly related to problematic use, from peripheral ones, associated with non-problematic involvement. This emphasizes an urgent need to renew and improve the assessment of PSMU, and behavioral addictions in general.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Maurage, P., Maurage, P., Maurage, P., & Billieux, J. (2026). Addictive use of social media scales are not necessarily measuring problematic behavior: Evidence from the international 2017 Health Behavior in School-Aged Children (HBSC) study. Journal of Psychopathology and Clinical Science, 1(1), 1. https://hdl.handle.net/2078.5/274361 (Original work published 2026)