Instructors of large online courses often rely on asynchronous online discussions (AOD) to engage students, build community, and enhance collaborative learning. Despite these documented benefits, AOD frequently encounter challenges such as superficial interactions, uneven participation, and motivational barriers. In contrast, peer assessment---an engagement strategy involving structured peer feedback---has been shown to promote deep cognitive engagement. Despite their complementary potential, few studies have directly compared peer assessment and AOD as distinct, coexisting engagement strategies within the same learning environment. This study addresses this critical gap by examining the relative impacts of AOD and peer assessment on student learning performance and perceptions in a large online graduate course in computer science (N = 1,451) over five years (Spring 2020--Fall 2024). Students were classified based on participation: Collaborators (active in both activities), Reviewers (active primarily in peer assessment), Discussants (active primarily in online discussions), and Limited Contributors (limited participation). ANOVA results showed Collaborators and Reviewers significantly outperformed Discussants and Limited Contributors, indicating that peer assessment is strongly linked to improved performance. Additionally, Discussants outperformed Limited Contributors, reaffirming the engagement value of AOD. Regression analysis further revealed that peer assessment was a stronger predictor than AOD for both perceived peer support and course effectiveness, with a larger predictive gap observed for peer support. These findings highlight the complementary potential of combining peer assessment with AOD in enhancing engagement and performance in large-scale online learning environments. Educators are encouraged to strategically combine these strategies to leverage their distinct strengths, while future research should explore interventions to better engage minimally active students.

Artificial intelligence (AI) is increasingly integrated into project-based learning (PjBL), with research primarily addressing two dimensions: learning about AI through PjBL and learning with AI as a supportive tool. However, studies on AI's role in PjBL remain limited, often focusing on small-scale, short-term contexts. Additionally, there is little focus on a third dimension---learning through AI Creation---where students deepen engagement by developing AI-driven projects. This study investigates AI's dual role in online PjBL at scale, analyzing how students use AI tools and engage in AI-driven projects within a graduate computer science course over six semesters (2023-2024). Using a mixed-methods approach, we analyzed survey responses from 467 students and conducted a thematic analysis of 436 unique project keyword sets. Findings reveal that students primarily use AI for technical tasks, research, and content generation, supporting PjBL's planning and execution phases. The analysis of students' project keywords highlights AI in Education as the most prevalent project theme, with subthemes like Large Language Models, Intelligent Tutoring Systems, and Personalized Learning, indicating strong student interest in leveraging AI to address educational challenges related to personalization and scalability. The diversity of AI subthemes, including AI Ethics and AI for Accessibility, suggests that students are exploring AI creation through multiple perspectives, considering both technical and societal implications. This study contributes to the growing body of research on AI in education by offering large-scale, longitudinal insights into AI's dual role in online PjBL: as a tool for enhancing learning and as a medium for deeper engagement through AI creation. The findings have important implications for curriculum design, such as scaffolded AI creation and ethical training, preparing students for a future in which they are not mere consumers of AI technologies but also responsible AI innovators.

Research consistently shows that peer assessment affects student achievement and attitudes across various subjects and contexts. However, most studies have focused on anonymous peer assessments in small, one-time, and non-iterative settings. This large-scale longitudinal study explores the effects of open, collaborative, and AI-augmented peer assessment in a large online graduate course with 1,636 students across 12 semesters from 2018 to 2022. The research investigated how different groups of students participated in the peer assessment in an online graduate computer science class. We also explored how their participation influenced their learning performance and their perceptions. Key findings reveal that students' age and gender significantly affect engagement levels and the perception of peer assessment effectiveness. They also provide new insights into the positive relationship between providing feedback and enhanced learning performance. This paper presents the implementation of peer assessment and detailed findings of the study. The implications of the study for future research and practices are also discussed.

Despite the ubiquitous use of videos in online learning and enormous literature on designing online learning, there has been relatively little research on what pedagogical strategies should be used to make the most of video lessons and what constitutes an effective video for student learning. We experimented with a model of incorporating four pedagogical strategies, four instructional phases, and four production guidelines-in designing and developing video lessons for an online graduate course. In this paper, we share our experience as well as students' perceptions of their effectiveness. We also discuss what needs to be done for future research.