Asia Pacific Education Review Education

Problematic social media use, study habits, and academic identification in a Thai undergraduate sample: a cross-sectional SEM study

Zhang, Yuxin, Alivi, Mumtaz Aini Binti
Springer Nature · 2026-11-13 · DOI 10.1007/s12564-026-10167-1

Abstract

This study examined associations among problematic social media use (PSM), study habits (SH), and academic identification (AI), with self-regulated learning as the proximal conceptual context for study habits, in a Thai undergraduate sample ( N = 337). The six PSM items were adapted from the Bergen Social Media Addiction Scale, but the original frequency format was changed to a five-point agreement format; the administered Thai questionnaire was not retained. In the baseline cross-sectional structural equation model (SEM), the standardized PSM → SH, SH → AI, and PSM → AI paths were 0.678, 0.456, and 0.320, respectively, and the coefficient product was 0.309. A free-loading orthogonal common-method-factor sensitivity model attenuated and reversed the study-habit paths (PSM → SH: β = −0.302, p = .580; SH → AI: β = −0.112, p = .832) but not the PSM → AI path ( β = 0.326, p = .001), whereas an equal-signed-loading random-intercept item-factor model estimated method variance at the boundary and therefore left the baseline paths unchanged. Thus, the study-habit associations were sensitive to the chosen method-factor specification. The findings do not establish a learning benefit, mediation mechanism, or temporal sequence; they foreground the measurement uncertainties of an adapted addiction-oriented measure and the need for separately assessed response style and educational-use purpose.

Read at publisher →