How Effective is Technology in Language Education? A Meta-analysis of Experimental Studies

Abstract

This meta-analysis examines the effect of technology-enhanced instruction on Arabic language achievement relative to traditional instruction. A systematic search of Scopus and Web of Science yielded 688 records, from which 15 primary studies providing 19 independent comparisons met the inclusion criteria. Effect sizes were computed as Hedges' g under a random-effects model with the DerSimonian–Laird estimator. The pooled effect was g = 1.41 (95% CI [0.80, 2.02], p < .001), a very large positive effect, accompanied by substantial heterogeneity (I2 = 97.46%). Moderator analyses showed that technology type (QB = 313.37), education level (QB = 40.53), Arabic language status (QB = 87.18), target language skill (QB = 141.52), and intervention duration (QB = 227.70) all significantly moderated the effect (all p < .001). Mobile learning produced the largest effect (g = 2.57), followed by game-based learning (g = 1.75); undergraduate students benefited more than K-12 learners, and second-language learners showed the highest gains. Productive skills yielded the strongest effects, whereas grammar instruction produced the most consistent results. Publication bias tests indicated no significant asymmetry. Interpreted through Mayer's Cognitive Theory of Multimedia Learning and Krashen's Input Hypothesis, the findings offer evidence-based guidance for integrating technology into Arabic language education