Voice-Based Conversational User Interfaces for Enhancing Accessibility in Learning Environments: A Mixed-Methods Study in Enugu State, Nigeria

Abstract

Accessible digital learning remains uneven in Nigerian public education, including in Enugu State, where inclusive-education policy exists but assistive tools, teacher preparation, and multimodal content are still limited. This mixed-methods study examined whether a bilingual (English and Igbo) voice-based conversational user interface (CUI) could improve accessibility, engagement, usability, and short-term learning outcomes relative to conventional graphical learning platforms. A pretest–posttest control-group design was implemented over six weeks with 248 senior secondary and first-year tertiary students from 10 public institutions in Enugu urban and peri-urban local government areas. Seventy-six participants had documented visual, motor, or specific learning difficulties. The experimental group (n = 124) used OluLearn, a speech-in/speech-out tutoring and content-access assistant aligned with Universal Design for Learning; the control group (n = 124) continued with existing teacher-led and GUI-based materials. Data were collected with an accessibility scale, the System Usability Scale, an engagement inventory, Technology Acceptance Model items, curriculum-aligned quizzes, task logs, and 24 semi-structured interviews. After the intervention, the CUI group reported higher perceived accessibility (M = 3.89 vs. 2.54), higher engagement (M = 4.06 vs. 3.14), higher quiz scores (M = 73.18 vs. 62.41), and higher usability (SUS = 79.12 vs. 57.86). Gains were larger among students with visual or motor impairments. Perceived usefulness and ease of use jointly explained 56% of the variance in intention to use. Qualitative themes highlighted hands-free access, Igbo voice support, and reduced dependence on print, alongside constraints of noise, connectivity, and speech-recognition accuracy for local accents. The findings support voice-first CUIs as a feasible accessibility layer for Enugu learning environments if they are bilingual, multimodal, and paired with teacher training and reliable power and network infrastructure.