Inclusive Mobile Learning: How Technology-Enabled Language Choice Supports Multilingual Students

Paper Title

Inclusive Mobile Learning: How Technology-Enabled Language Choice Supports Multilingual Students

Publication Info

• Topic area: Multilingual education in mobile learning for resource-constrained environments.
• Keywords: Multilingual education, mobile learning, EdTech, STEM education, low-resource settings, language choice, rural learners, non-formal education, Uganda, accessibility.

Background and Problem

• Problem / challenge: Traditional education systems and many EdTech platforms are limited to monolingual instruction, which excludes learners with limited proficiency in the dominant language. There is a lack of large-scale evidence on how multilingual EdTech can improve access and learning outcomes in non-formal, resource-constrained settings.
• Significance: Multilingual education can broaden participation and improve learning outcomes, particularly for disadvantaged learners in rural areas with limited formal education and low prior knowledge. This is critical for addressing educational inequities.
• Motivation and related work: Prior research highlights the benefits of mother-tongue instruction and the potential of mobile learning in low-resource settings. However, existing studies often focus on formal education or language learning, neglecting the delivery of disciplinary content in multiple languages. Few studies investigate the impact of flexible language choice in EdTech, particularly in sub-Saharan Africa.

Solution

• Proposed approach: A multilingual mobile learning platform offering a STEM course in English, Leb Lango (a local language), or a Hybrid option combining both languages, allowing learners to self-select their preferred language of instruction.
• Novelty:
  1. Conceptualizing language choice as a personalization mechanism to expand access and inclusion in EdTech.
  2. Providing the first large-scale quasi-experimental evidence on multilingual EdTech in rural, resource-constrained settings.
  3. Offering design insights for implementing language support in low-tech, multilingual learning environments.
• Procedure and key techniques:
  1. Conducted a quasi-experimental study with 2,931 learners in rural Uganda.
  2. Learners self-selected one of three language options for a 15-week STEM course delivered via radio and mobile phones.
  3. Measured accessibility, engagement, learning trajectories, and performance using demographic data, baseline knowledge, and assessment scores.
  4. Analyzed data using multinomial logistic regression, linear mixed-effects models, and propensity score matching.

Results

• Concrete findings:
  • 43% of learners chose the Hybrid option, 31% chose English, and 26% chose Leb Lango.
  • Learners with lower education levels, from rural areas, and with lower prior knowledge were more likely to choose Leb Lango.
  • Leb Lango learners demonstrated faster learning gains and closed performance gaps with English and Hybrid learners by the final exam.
  • Engagement with radio broadcasts was higher among Leb Lango learners compared to English learners.
• Advantage over baselines:
  • Offering a local language option increased participation among disadvantaged learners.
  • Faster learning gains for Leb Lango learners compared to English and Hybrid learners, despite lower initial performance.
• Experiments / evaluation:
  • Conducted in Lira District, Uganda, with a diverse sample of learners.
  • Evaluated engagement (radio prize codes), learning trajectories (weekly tests), and final performance (exams).
  • Used statistical models to analyze demographic, engagement, and performance data.
• Limitations and future work:
  • Lack of randomization introduces potential self-selection bias.
  • Limited baseline data and reliance on multiple-choice assessments.
  • Translation quality and equivalence across languages were not systematically assessed.
  • Future work includes randomized controlled trials, qualitative studies on learner motivations, and improving translation consistency.

Summary

This study demonstrates that multilingual mobile learning platforms can expand access and improve learning outcomes for disadvantaged learners in resource-constrained settings. By offering a STEM course in English, Leb Lango, or a Hybrid option, the platform enabled broader participation, particularly among rural learners with limited prior knowledge. Leb Lango learners showed faster learning gains and closed performance gaps with English and Hybrid learners by the final exam. These findings highlight the potential of language choice as a personalization mechanism in EdTech and underscore the importance of designing multilingual, low-tech learning environments to support diverse learner needs.