NEED: Writing exercises have been used to detect student misconceptions regarding course content across a variety of STEM disciplines. Further, writing exercise often provide deeper insight into an individual’s cognitive processes that do simple computation-based problems. Recently, short conceptual-based writing exercises have shown promise not only to aid in detecting misconceptions but to help at-risk students achieve improved understanding of fundamental concepts in a circuit analysis course. While writing exercises may be a valuable tool to detect and correct common misconceptions of students in gateway STEM courses, evaluating writing exercises and providing personalized feedback is time consuming, thus limiting their use. This project seeks to leverage advances in natural language processing (NLP) to automate the evaluation and feedback tasks within web-based writing applications. Successful implementation of such web-based writing applications would enable the creation of a scalable writing-centric approach in gateway STEM courses, balancing the importance of conceptual understanding with the development of standard problem-solving skill.GUIDING QUESTIONS: 1. What is the impact of the writing-centric approach in terms of enhancing students’ conceptualunderstanding, metacognitive skill, and ultimate success in a foundational course on electric circuit analysis? 2. Is there evidence that the impact of the writing exercises depends upon the mode with which the exercises are administered, and the feedback provided?3. To what extent is the web-based writing application template to be developed in the project transferable to conceptual writing exercises across a variety of gateway STEM courses?ANTICIPATED OUTCOMES:It is anticipated that enhanced conceptual understanding and metacognitive skill will be demonstrated by at-risk students engaged in the writing-centric approach and that the effect size of the approach will increase when feedback is instantaneous and tailored to the individual. In this study, classification between the at-risk and not at-risk groups is made using an evaluation that considers not only the correctness of a student’s response to an initial conceptual-based writing exercise, but also demonstration of their consistent reasoning in justifying the response. A second key anticipated outcome is that the web-based template developed in this study will be able to be readily adapted across STEM fields, requiring minimal coding skill on the part of the instructor.BROADER IMPACTS:While the project activities are focused on a sophomore-level course in electric circuit analysis at a particular institution, instructors from a diverse set of STEM fields will be trained to implement a conceptual-based writing exercise of their choosing within the developed software template. Not only will this open a path for the adoption of a writing-centric approach in a variety of gateway STEM courses for which students often struggle, but also provide the research team a formative assessment tool for adjusting the final version of the template.