Automated short answer grading with feedback (ASAG-F) systems currently face challenges in transparency, pedagogical alignment, and cost-effectiveness that limit their real-world deployment. We introduce GraphRAG, a knowledge graph-based retrieval-augmented generation framework that addresses these limitations by grounding all large language model (LLM)-generated feedback and scores in instructor-curated atomic facts, ensuring traceability and verifiability. Using the Short Answer Feedback (SAF) dataset with 31 topics, we evaluate GraphRAG on unseen-question and unseen-answer splits. Our systematic evaluation demonstrates that GraphRAG achieves grading accuracy comparable to vector-based RAG and generally superior to a fine-tuned LLM baseline model while providing more transparent source attribution. Additional findings include: (1) Instructing the LLM to discretize continuous scores to match pedagogical rubrics, such as the 0.25 increments common in SAF, improves grading accuracy; (2) LLM-generated feedback exhibits length-dependent quality variations when unconstrained; prompt-based length control substantially enhances feedback quality and its stability, achieving optimal balance of instructional richness and conciseness; (3) Performance scaling analysis reveals that basic models like GPT-4o-mini offer cost-effective performance, while premium models like Claude-Opus-4 show diminishing returns. These results demonstrate that GraphRAG offers a robust, explainable, pedagogy-aligned, and cost-effective solution for large-scale educational applications, enabling transparent automated grading with effective pedagogical feedback and practical deployment costs.

Recently, academic institutions have started to explore the potential of commercially available multiplayer computer game engines for the development of virtual environments for instructional purposes. This trend is motivated by the fact that many students are highly accustomed to and knowledgeable about playing electronic and computer games. Computer games share a number of potential characteristics with effective instructional tools and therefore have a great potential for positively affecting learning. Achieving a sense of immersion by the students and interactive collaboration among them are two of the main goals being pursued in exploring new types of educational laboratories as alternatives or complements to traditional hands-on experiments. A scripted scenario for a game-based gear train laboratory exercise was implemented in `ME 358 Mechanisms and Machine Dynamics', a junior-level course for mechanical engineering majors at Stevens Institute of Technology (SIT). This paper discusses the learning assessment conducted in that implementation. The results are presented for two assessment studies comparing versions of game-based gear train experiments to workbook-based gear train experiments with a `hands-on' laboratory experience. The results suggest that the students learned the concepts addressed by the laboratory content better in the game-based condition and that they had a realistic understanding and appreciation of the practical advantages of game-based gear train experiments over more traditional learning experiences. In addition, a questionnaire was administered to the students and then analyzed in order to obtain further insights into the students' opinions about and attitudes toward the game-based laboratory approach.

Laboratories play a crucial role in the education of future scientists and engineers, yet there is disagreement among science and engineering educators about whether and which types of technology-enabled labs should be used. This debate could be advanced by large-scale randomized studies addressing the critical issue of whether remotely operated or simulation-based labs are as effective as the traditional hands-on lab format. The present article describes the results of a large-scale (N = 306) study comparing learning outcomes and student preferences for several different lab formats in an undergraduate engineering course. The lab formats that were evaluated included traditional hands-on labs, remotely operated labs, and simulations. Learning outcomes were assessed by a test of the specific concepts taught in each lab. These knowledge scores were as high or higher (depending on topic) after performing remote and simulated laboratories versus performing hands-on laboratories. In their responses to survey items, many students saw advantages to technology-enabled lab formats in terms of such attributes as convenience and reliability, but still expressed preference for hands-on labs. Also, differences in lab formats led to changes in group functions across the plan-experiment-analyze process: For example, students did less face-to-face work when engaged in remote or simulated laboratories, as opposed to hands-on laboratories.