High performance computing (HPC), typically used only in research until recently, is being integrated more into undergraduate curricula across disciplines. HPC resources are expensive but lend themselves easily to sharing among institutions. However, in the context of undergraduate education, inter-institutional sharing of such resources is still relatively rare. The authors hold the belief that generating administrator support early in system-wide initiatives for sharing resources is crucial for long term success. Taking the case of a specific HPC cluster at the University of Wisconsin-Eau Claire, part of the University of Wisconsin System (UW-System) of higher educational institutions, the authors asked 15 administrators from the system about the barriers as well as the path forward for better cluster sharing in undergraduate education. The responses to the interview questions were analyzed qualitatively for themes and sentiments. Five themes arose from this analysis, related to awareness of and access to HPC, cost-sharing structures, industry needs, institutional factors, and variations in benefits depending on the field. The sentiments were overwhelmingly positive, but with variations depending on the questions asked and approaches taken by the interviewees. Overall, the responses indicated a positive attitude and willingness to incorporate HPC in undergraduate education as well as the importance of awareness of such resources. They also indicate that cost-sharing structures and institutional factors are both the biggest barriers and the keys to better sharing practices and industry partnerships. At the end, the authors make recommendations for anyone looking to establish a framework to share HPC resources across institutions.

This Innovative Practice Full Paper is part of a collaboration between a US and a Chinese university, in which two different online engineering courses, situated at the US university, were taught to mixed classes of students from both. Course content remained the same as standard offerings, and assessments had both individual and group components, specifically instructor-selected mixed-university groups. The first course was run in the spring 2021 semester, followed by a different course in the fall 2021 semester.International education offerings have been hampered by disruptions due to Covid-19, in addition to the usual barriers. Alternate models like collaborative online international learning (COIL) provide an avenue to continue international partnerships and expand their modalities. The current project is based on the COIL model but differs in two ways: the courses were not specially developed for COIL but adapted to it without changing course objectives, and the courses were taught solely by the instructors at the US university though administrative support was provided by both universities.Data for this project was collected using student surveys, personal interviews, instructor reflections, and course assessments. Student survey data is qualitatively analyzed using thematic analysis, while personal interviews and instructor reflections are summarized by the authors. Student learning outcomes in both sections were compared to those in the sections offered only to on-campus students and found to be similar or slightly better. Student perceptions of the classes were positive despite substantial difficulties.Instructors and students in these classes faced significant challenges like time zone differences and scheduling issues, technological barriers, and cultural differences leading to miscommunication. However, the collaborative aspects of the courses made them attractive for both groups of students, and opportunities for easy communication during synchronous class meetings made things run smoothly.In addition to discussing the challenges and attractions, the authors make a series of recommendations on conducting a regular online course in STEM disciplines with students from two different education systems.

The work presented in this Full Paper is categorized as Innovative Practice, as per FIE guidelines. This project is based on data collected from two coding-intensive courses in an engineering program at a mid-sized, open-enrollment primarily undergraduate institution (PUI). Notably, the engineering curriculum does not include an introductory programming course. In both courses, groups of students worked on engineering problems using code, after which they presented their initial drafts to the entire class. Their peers, watching the presentations, were asked to provide written feedback in the form of 4 objective questions and 3 short, essay-type responses. A thematic analysis of the essay-type responses indicates that the prompts determined what the responses were likely to focus on. When asked what the presenters had done well, the responses overwhelmingly related to the quality of the presentation itself without details on content. When asked for what could be improved, the responses most often included checkbox-like statements like whether the answer was correct, if there were any syntax errors, etc. When asked for one takeaway from the presentation they just watched in a live classroom, responses in both classes revolved around coding-related terms and problem-solving steps, relevant to each course. When classes went online for Fall 2020 and presentations were held over video conferencing, the responses to this prompt also included significantly more expressions of emotions like commiserating with peers or being validated by them. Themes emerging from the other two prompts did not vary based on class modality. This thematic analysis can, therefore, inform instructors teaching programming-heavy engineering courses on how best to design reflective activities that are integrated into course assessments.

Students in non-CS STEM majors often have specific difficulties with programming content in their coursework. Instructor feedback is usually limited to the results obtained and major algorithmic errors, due to the nature of coding assignments. The authors' aim is to investigate whether presenting code and receiving feedback from their peers in class makes an appreciable difference in non-CS STEM-major students' learning of programming. The poster will present data collected over two semesters from five sections of three programming-heavy chemistry, engineering and math courses taught at the University of Wisconsin-Stout. Student groups were asked to present their homework attempts to the class, their peers were asked to fill out surveys containing Yes/No questions and comments on the attempts, and this feedback was passed on to the presenting students, who were then given time to improve their work before submitting their final versions. Comparison of student performance statistics in exams, between semesters during which the study was implemented and prior semesters, is used as a tool to assess student learning. Results of the study from two semesters have shown that group composition and the type of questions asked in the survey are the two major factors at play. The poster will also discuss the enhanced design of the study during the third and final semester in spring 2019 that will take these results into account. DOI: http://doi.org/10.1145/3287324.3293820