Screens and Self-Belief: Digital Habits Shape Student Confidence Differently in Korea and Australia

Every three years, the Programme for International Student Assessment hands researchers one of the richest datasets in education science, and the 2022 cycle arrived with a twist that few analysts could ignore: it was the first PISA round conducted largely amid the aftermath of pandemic-era remote schooling, when screens had shifted from supplementary tools to the primary medium of learning for millions of teenagers. A new study published in Current Psychology by Ahrong Beik of Chinju National University of Education and Jinsil Kim of SkillsFuture Korea seizes on that moment, asking a deceptively simple question about two very different digital societies. When tenth-grade students in South Korea and Australia pick up their devices, does what they use them for — studying, or simply passing time — relate to how confident they feel about their academic abilities? And does the digital skill they carry into that interaction act as the bridge between usage and self-belief?

The study drew on PISA 2022 responses from 14,465 tenth-grade students across the two countries, applying structural equation modeling to trace pathways that raw correlations alone cannot reveal. The researchers distinguished between two broad purposes of device use: academic purposes, such as looking up information for schoolwork or using learning software, and leisure purposes, such as entertainment and social browsing. On the other end of the proposed chain sat subject-related academic self-beliefs, the confidence students express in their own competence in specific school subjects — a construct long tied in psychological research to motivation, persistence, and eventual achievement. Between the two sat digital competency, the students’ self-assessed fluency with digital tools, which the authors hypothesized would carry part of the effect from how devices are used to how capable students feel.

Before any structural model could be estimated, the team confronted a methodological hurdle that quietly governs the credibility of nearly all cross-national psychology: measurement invariance. If a survey item about digital confidence means something subtly different to a student in Seoul than to a student in Melbourne, comparing the resulting scores across countries is like comparing temperatures with two differently calibrated thermometers. Multi-group invariance testing examines whether the same constructs are measured with the same structure, the same loadings, and ultimately the same intercepts across groups. In this study, the analysis confirmed configural and metric invariance but found that scalar invariance did not hold between the Korean and Australian samples — meaning the intercepts of the items differed systematically, likely reflecting cultural and linguistic response-style differences of the kind documented in classic work on the reference-group effect in Likert-scale comparisons.

The consequence of that failed scalar invariance was decisive for the study’s design. Rather than pooling the two national samples into a single cross-national model — a move that would have implicitly assumed the constructs were measured identically — the authors estimated separate structural models for Korea and Australia. This choice is methodologically conservative and, the researchers argue, honest: it reframes the entire enterprise from a head-to-head comparison of nations into an exploration of within-country association patterns. The authors are explicit that the resulting findings should be read as exploratory, not as evidence of national differences, and that they do not support broad cross-national policy generalizations. In an era when PISA league tables are routinely over-interpreted, that restraint is itself a finding worth headline treatment.

Within the Korean model, a clear pattern emerged. Academic-purpose device use showed a small positive association with students’ subject-related academic self-beliefs, and part of that association was model-implied through digital competency. In plain terms, Korean students who reported using their devices more for learning tended to report slightly higher confidence in their school subjects, and some of that relationship ran through their sense of digital skill. The pathway is theoretically coherent: using technology for study may give students repeated, low-stakes opportunities to practice and master digital tools, and that growing fluency may feed back into a broader sense of academic capability. The authors situate this within Korea’s intensely digital educational environment, where online learning platforms and device-based study are deeply embedded in secondary schooling, and within cultural frameworks emphasizing collectivist norms that shape how students engage with shared educational technologies.

The Australian model told a subtly different story. There, both academic-purpose and leisure-purpose device use showed small positive associations with academic self-beliefs via digital competency. That the leisure pathway appeared at all in the Australian sample is intriguing: it suggests that for Australian students, even recreational screen time may contribute, however modestly, to the digital fluency that in turn relates to academic confidence. The authors caution against romanticizing this interpretation — the effect sizes are small, and the cross-sectional design forbids any claim that gaming or scrolling builds self-belief. But the pattern is consistent with a body of research on digital inequality, notably Hargittai and Hinnant’s work showing that what young people actually do online, not merely how much they are online, differentiates their digital skills. It also resonates with the technology acceptance literature, in which perceived ease of use and perceived usefulness jointly shape how people adopt and benefit from technology.

The contrast between the two national patterns invites cultural interpretation, and the authors provide it with appropriate hedging. Korea’s educational culture, shaped by collectivist traditions and a highly standardized, exam-driven system, may channel device use more narrowly toward academic ends, so that leisure use contributes little to the confidence pathway. Australia’s more individualist educational context, with greater curricular flexibility and a different relationship between students and technology, may allow a wider range of digital activities to feed into competency. The study draws on Hofstede’s cultural dimensions and on Markus and Kitayama’s influential account of independent versus interdependent self-construals to frame these possibilities, while stopping well short of claiming that culture explains the difference — the design simply cannot support that inference.

What makes the paper methodologically notable is its transparency about limits. The authors emphasize that the pathways detected are small in empirical magnitude, that the data are cross-sectional and therefore silent on causality, and that the failure of scalar invariance rules out direct comparison of effect sizes across the two countries. They also note that the study analyzed secondary data with no identifying information and was accordingly exempt from institutional review board approval, and that the underlying datasets will be available from the authors upon reasonable request. The modeling choices — including attention to non-normal and categorical data in structural equation modeling, and standard fit criteria from the covariance-structure literature — reflect established best practice, and the decision to test Little’s missing-data assumptions before proceeding shows the kind of diligence that often goes unreported in secondary analyses of large international assessments.

For educators and policymakers, the takeaways are modest but actionable. The study lends cautious support to the idea that the purpose of device use matters more than raw screen time: academic-purpose use was consistently linked, through digital competency, to healthier academic self-beliefs in both countries. That aligns with a growing consensus in educational technology research that active, goal-directed engagement with digital tools — rather than passive consumption — is what converts access into learning benefit. It also underscores digital competency itself as a plausible lever: if schools can build students’ genuine fluency with technology, that fluency may carry secondary benefits for how capable students feel across their subjects, independent of which devices they own or how many hours they log.

The broader lesson may be about how the field studies digital life at all. As debates over screens in schools grow increasingly polarized, studies like this one model a middle path: use large, internationally comparable data, but respect the psychometric limits of cross-cultural measurement; report small effects as small; and resist the temptation to convert correlational pathways into policy slogans. Beik and Kim’s analysis of the PISA 2022 samples suggests that the relationship between our digital habits and our academic self-image is real but fragile, filtered through culture, purpose, and skill — and that understanding it will require precisely the kind of careful, country-specific, deliberately humble modeling this study exemplifies.

Subject of Research: Associations between digital device usage, digital competency, and academic self-beliefs among Korean and Australian students in PISA 2022

Article Title: Country-specific associations among digital device usage, digital competency, and subject-related academic self-beliefs in Korean and Australian PISA 2022 samples

Article References: Beik, A., & Kim, J. (2026). Country-specific associations among digital device usage, digital competency, and subject-related academic self-beliefs in Korean and Australian PISA 2022 samples. Current Psychology, 45(20), Article 1594. https://doi.org/10.1007/s12144-026-10149-x

Image Credits: AI Generated

DOI: 10.1007/s12144-026-10149-x

Keywords: digital competency, academic self-beliefs, PISA 2022, structural equation modeling, measurement invariance, South Korea, Australia, educational technology, screen time, self-efficacy, cross-cultural psychology, adolescents