Comparative Analysis of E-Learning Effectiveness in STEM vs. Humanities

Thesis Abstract

This study addresses the persistent question of whether e-learning environments yield equivalent learning outcomes across STEM and humanities disciplines, given differences in epistemic practices, assessment methods, and interaction patterns. The problem stems from inconsistent evidence in higher education literature regarding discipline-specific effectiveness of digital learning platforms, which has implications for curriculum design, instructional strategy, and resource allocation. The aim is to compare e-learning effectiveness between STEM and humanities programs, operationalized through learning gains, engagement, and perceived learning quality, with the objective of identifying discipline-specific facilitators and barriers to success. Specific objectives are (1) to quantify differences in cognitive gains between STEM and humanities students enrolled in fully online or blended courses; (2) to examine differences in learner engagement, motivation, and self-regulated learning strategies; (3) to assess teacher support, feedback quality, and technical infrastructure as mediators of learning outcomes; (4) to explore student and instructor perceptions of e-learning affordances and constraints; and (5) to develop a conceptual model linking disciplinary context to e-learning effectiveness. The study adopts a comparative cross-sectional design combining quantitative and qualitative methods. The population comprises undergraduate and graduate students enrolled in accredited online or blended courses across three public universities. A stratified random sample of 600 students (300 STEM, 300 humanities) and 60 instructors will be drawn, ensuring representation across program levels and delivery modes. Data collection instruments include a standardized achievement test to measure learning gains (pre-test/post-test design with validated alignment to course learning outcomes), the Instructional Materials Engagement Scale, the Online Learning Readiness Inventory, and the Motivation and Engagement Scale. Additional instruments are a course-embedded survey to capture perceived learning quality and instructor feedback quality, and platform analytics to capture participation, time-on-task, and assessment submission patterns. For qualitative insights, semi-structured interviews will be conducted with 24 students (12 STEM, 12 humanities) and 10 instructors, complemented by 6 focus groups (three per discipline). Validity and reliability procedures include pilot testing, Cronbach’s alpha assessment (target ? ? .80 for scales), factor analyses, and triangulation across instruments and data sources. Analytical approaches comprise descriptive statistics to profile the sample, multiple-group structural equation modeling (SEM) to test a disciplinary moderation model of learning gains as a function of engagement, readiness, and instructional quality, and multilevel regression to account for nested data at the course and university levels. Mediation analyses will examine the roles of instructor feedback quality and technical support in linking engagement to outcomes; moderation analyses will test whether discipline (STEM vs. humanities) alters the strength of these relationships. Thematic analysis will be applied to interview and focus group transcripts to extract themes related to pedagogical affordances, assessment alignment, and cultural norms within each discipline. A conceptual model will be proposed that integrates the quantitative and qualitative findings, drawing on constructivist and sociocultural learning theories to interpret disciplinary differences. Expected findings anticipate that STEM students will show higher cognitive gains in courses with frequent, timely feedback and active problem-solving tasks, whereas humanities students may exhibit stronger gains when courses emphasize discussion-based activities and interpretive tasks, albeit with potential variability due to assessment formats and measurement sensitivity. Engagement metrics are expected to mediate the relationship between instructional quality and learning outcomes across both groups, with stronger effects in courses that provide structured feedback loops and clear navigation. The study aims to contribute to knowledge by clarifying discipline-specific determinants of e-learning effectiveness and by offering an empirically grounded framework for designing adaptive, discipline-informed e-learning interventions. The practical implications include evidence-based guidelines for curriculum designers and instructors to tailor e-learning strategies to STEM and humanities contexts, recommendations for university policymakers on resource allocation for platform features and instructional training, and a validated model for anticipating e-learning success across disciplines. The study concludes with actionable recommendations to optimize e-learning delivery, including enhancing formative assessment practices in STEM, fostering collaborative discourse in humanities, strengthening instructor feedback protocols, and investing in robust technical support to reduce cognitive load and improve learner satisfaction.