EUVIMEDEUVIMED entdecken
Crossref

Modelling the Latent Dynamics of Digital Device Use, Parental Involvement, Psychological Capital and English Achievement: A Bayesian Finite Mixture Structural Modelling Approach with Data-Driven Class Enumeration

Palash Majumder, Nimisha Beri, Salil Biswas

International Journal of Innovative Science and Research Technology · 14.04.2026 · 541

DOI: 10.38124/ijisrt/26apr458 ↗

Kurz gefasst

Fachartikel in der Zeitschrift International Journal of Innovative Science and Research Technology, online erschienen am 14.04.2026. Verfasst von Palash Majumder und 2 weiteren Autor:innen. Die Arbeit nennt 50 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This study examined heterogeneous relationships between digital device use (DDU), parental involvement, psychological capital, and English as a Second Language (ESL) achievement among secondary students from the marginalized Matua community of West Bengal, India. Conventional structural equation modeling (SEM) fails to account for unobserved heterogeneity in DDU patterns and the need for data-driven identification of latent subgroups. This study addressed these limitations by implementing a Bayesian Finite Mixture Structural Model (BFM-SEM) using an overfitted finite Gaussian mixture with posterior sparsity for data-driven class discovery, with district membership modeled as a fixed covariate to account for data clustering across two districts (Nadia: n = 301; North 24 Parganas: n = 299; total N = 600). The posterior converged on three empirically distinct DDU profiles: Passive Consumers (40.7%), Balanced Users (35.2%), and Educational Engagers (24.1%; note: percentages sum to 100.0% after rounding adjustment); mean posterior classassignment probabilities exceeded .86 for all classes, indicating high classification certainty. Preliminary conventional SEM confirmed significant structural associations—including negative associations between unguided DDU and psychological capital (β = −.281, p < .001) and positive associations between parental involvement and ESL achievement (β = .185, p < .001)—but revealed poor model fit (RMSEA = 0.165; CFI = .831), and was decisively outperformed by the BFM-SEM (ΔELPD = 367.4, SE = 28.3). Profile-specific structural estimates revealed substantial heterogeneity: the negative association between DDU and psychological capital was strongest among Passive Consumers (β = −.412, 95% CI [−.501, −.323]) and non-significant among Educational Engagers (β = −.048, 95% CI [−.157, .061]). All Bayesian diagnostics confirmed model convergence (R̂ ≤ 1.01; ESS > 400). These findings may help guide differentiated, culturally responsive digital literacy intervention design.

Zitieren

APA

Majumder, P., Beri, N., & Biswas, S. (2026). Modelling the Latent Dynamics of Digital Device Use, Parental Involvement, Psychological Capital and English Achievement: A Bayesian Finite Mixture Structural Modelling Approach with Data-Driven Class Enumeration. International Journal of Innovative Science and Research Technology, 541. https://doi.org/10.38124/ijisrt/26apr458

Vancouver

Majumder P, Beri N, Biswas S. Modelling the Latent Dynamics of Digital Device Use, Parental Involvement, Psychological Capital and English Achievement: A Bayesian Finite Mixture Structural Modelling Approach with Data-Driven Class Enumeration. International Journal of Innovative Science and Research Technology. 2026:541. doi:10.38124/ijisrt/26apr458

BibTeX
@article{Majumder2026,
  author = {Majumder, Palash and Beri, Nimisha and Biswas, Salil},
  title = {Modelling the Latent Dynamics of Digital Device Use, Parental Involvement, Psychological Capital and English Achievement: A Bayesian Finite Mixture Structural Modelling Approach with Data-Driven Class Enumeration},
  journal = {International Journal of Innovative Science and Research Technology},
  year = {2026},
  pages = {541},
  doi = {10.38124/ijisrt/26apr458}
}
Kostenlos starten →