Abstract:

Undergraduate STEM courses increasingly rely on discussing primary literature to demonstrate relevance, build communication skills, and improve students’ reasoning. Presentations are especially powerful for delivering complex visual and mechanistic content in biology.

Yet two practical barriers limit their use at scale:
(1) the time cost of running fair, high-quality presentations in large and multi-section courses, and
(2) the difficulty of assessing not just performance (slides, delivery) but the level and quality of students’ thinking.

As a result, many instructors avoid presentations altogether or reduce them to fact-listing with minimal feedback, leaving both their cognitive potential and valuable learning opportunities underused. We address these barriers with a Presentation Exchange model that reorganizes presentations into an asynchronous, peer-reviewed workflow and couples it to a simple structure for eliciting higher-order thinking. Students create short, narrated presentation decks on research articles or cases, submit them to an online exchange, and both receive and provide rubric-based peer reviews. This workflow compresses in-class time, standardizes procedures across sections, and generates rich assessment data without increasing instructor grading load. To shift focus from performing to reasoning, we add lightweight scaffolds: each presentation must include at least one explicitly marked high-level move (e.g., contrasting mechanisms, challenging an assumption, tracing a causal pathway, or proposing a next experiment), and reviewers rate the depth and clarity of that move using a brief rubric aligned with upper levels of Bloom’s taxonomy or the Cognitive Demand Level framework. To further scale assessment and make cognitive data usable at the course and program level, we include AI-based scoring of presentation artifacts. Across multiple offerings of medium-size (75 students) biology courses, Presentation Exchange supports measurable gains in students’ factual knowledge of presented material and their self-efficacy for applications of specific course context (Cell Biology). Because students repeatedly see their peers’ explanations of the same figures, methods, and mechanisms, they benefit from multiple “passes” through difficult content and learn from contrasting approaches to the same problem, which strengthens conceptual understanding and promotes group cohesion. The low-stakes, low-stress structure: short narrated decks, distributed deadlines, and constructive peer review, all constructed at students' own pace and time, creates an accessible on-ramp to analyzing, presenting, and discussing primary literature, especially for students who might otherwise remain silent or peripheral in traditional, high-pressure presentation formats. Students report increased enthusiasm for science, greater confidence in reading and critiquing primary articles, and a stronger sense of belonging in the scientific community. Together, these results suggest that structured presentation exchange, supported by AI-based assessment, can convert a time-intensive teaching practice into a scalable mechanism for deep learning, peer-supported engagement, and durable gains in scientific identity.

Keywords:

Technology, education, peer, AI, research literature, presentations.