Algorithms courses are a foundational part of an undergraduate computer science degree that require abstract thinking and creativity and are known to be challenging for many students. Recently researchers have been developing auto-graded tools to scaffold students through the problem-solving process. We examine student's perceptions of such a tool in a required upper-division Algorithms course at a R1 University. The goal of the tool is to improve student experience in three ways: (1) help students break down the problem-solving process into clear steps; (2) increase students' self-efficacy by raising their confidence and understanding of the material; (3) have low ''cost'', by being easy to use, enjoyable, and a good use of students' time. The tool itself is designed to provide these benefits to students at every level of mastery through instantaneous feedback over increasingly challenging problems. It is designed as an addition to and not complete replacement of the written homework in the course. Based on a survey of almost 1000 students across four semesters, each with a different instructor, we examine whether student feedback is favorable over all four offerings, and for groups of students with different course outcomes. Using qualitative and quantitative methods, we found that across each of the four semesters and across letter grades A, B, C, and D students favored the tool as compared to written homework.

Graph data structures and algorithms play an essential role in computer science, and one of the ultimate goals of learning graphs is to solve more complicated algorithm design problems with them. A common way to solve a novel, complex problem is to reduce the problem to a standard graph problem, which often requires modeling a graph, and one essential way to model a graph is a technique called graph layering. Graph layering is often considered difficult by students and rarely studied by computer science education researchers despite its significance in algorithm design. To understand students' struggles with graph layering and improve teaching of algorithm designs, we conducted this qualitative study using think-aloud interviews with current students from an algorithm course. Participants were asked to solve algorithm design problems meant to be solved with graph layering. We used thematic analysis to extract difficulties observed in these interviews. We share our preliminary findings in this poster, and propose next steps for this study and future research.

Background: Proof Blocks is a software tool that enables students to construct proofs by assembling prewritten lines and gives them automated feedback. Prior work on learning gains from Proof Blocks has focused on comparing learning gains from Proof Blocks against other learning activities such as writing proofs or reading.

Office hour utilization in computer science courses can spike near deadlines, producing long wait times, frustrated students, and over-worked staff. To address this problem, a large CS1 course implemented a split deadlines policy. Students were randomly divided into two groups with staggered release and due dates. Each group had the same amount of time to complete assignments, but the number of students with each due date was reduced by half. Our study evaluates the effectiveness of this policy. We measure office hour utilization and staff efficiency near deadlines, examine the policy's impact on student performance, and investigate student perception of the policy's fairness and effectiveness. Overall we found that the split deadline policy increased office hour efficiency, resulted in no significant difference in performance between groups, and was considered fair and effective by most students. Our experience report includes reflections and student feedback indicating how to implement and further improve similar policies.

Algorithm design is a vital skill developed in most undergraduate Computer Science (CS) programs, but few research studies focus on pedagogy related to Algorithms coursework. To understand the work that has been done in the area, we present a systematic survey and characterization of existing studies in the CS Education literature related to the teaching of algorithm design at the undergraduate level. Across all papers in the ACM Digital Library, we only find 97 applicable papers. We classify these papers by topic, evaluation metric, evaluation methods, and intervention target. We present the results of these classifications alongside insights about existing knowledge, rigor, and contribution rates. We hope that this work not only provides a detailed representation of the current corpus of CS Education work related to algorithm design but also demonstrates that the body of knowledge is sparse and supports further research in the area. For future work, we intend to investigate and synthesize the conclusions reached by these papers.

Background : Proof Blocks is a software tool that enables students to construct proofs by assembling prewritten lines and gives them automated feedback. Prior research has shown that students learn as much from an activity where they use Proof Blocks as where they write proofs. However, in both cases students first read a book chapter. Prior research was not able to differentiate between the learning gains achieved from reading versus proof practice. Purpose : This study aims to measure learning gains from reading a book chapter versus completing Proof Blocks. Methods : We conducted a randomized controlled trial with three experimental groups: one that only read a book chapter, one that only completed Proof Blocks, and one that did both. Findings : The group that completed only Proof Blocks had the smallest learning gains. The group that read the book chapter and completed the Proof Blocks activity performed marginally better than students who only read the book chapter, but it is not clear if the source of this improvement was the Proof Blocks or just exposure to more examples.

Collaborative learning is an evidence-based instructional practice that has been widely used in higher education, but it is not a silver bullet and requires careful design and implementation to yield the desired benefits. Prior work has also shown that collaborative learning is especially effective for female students, who are historically underrepresented in Science, Technology, Engineering, Mathematics (STEM), and Computing fields, suggesting a promising prospect for applying collaborative learning in these disciplines. In spring and fall of 2022 at a large public university, an upper-level required computer science course using collaborative learning was offered in a hybrid format, with great flexibility in meeting time (students could meet and collaborate at the “instructor-scheduled” meeting time, or they could pick their preferred “student-scheduled” time) and team consistency (students were encouraged, but not required, to work with a fixed team throughout the semester). To understand the effects of meeting time preference and team consistency, we measured students' learning outcomes (exam performance) and experience (sense of belonging and satisfaction about team dynamics in collaborative learning), and conducted linear regression analyses. We also investigated whether the effects were different for male and female students. We found that students' meeting time preference had no significant effect on their exam performance, sense of belonging, or satisfaction, and the non-significant effects were homogeneous for both gender groups. We also found that having a consistent team had significantly positive effects on exam performance and sense of belonging, but no significant effect for satisfaction. Moreover, the effect of team consistency on exam performance was significantly stronger for female students than male students. Our findings justified the option to give students meeting time flexibility since it did not hurt their learning experience and outcome, and encouraged exploring effective approaches to forming consistent teams that make students intrinsically want to work with them. The gender difference in effects of team consistency on exam performance aligned with previous literature and served as evidence to use collaborative learning in computing and STEM classrooms.

A new computational linear algebra course was developed and offered at a large public university in the Midwest. This new course traded off some of the lecture time in the pre-existing traditional linear algebra course for applied computational materials taught in a flipped-classroom lab setting. We compare exam performance in a subsequent numerical methods course from students having taken either the new computational or traditional course, while controlling for student performance in prerequisite computer science and mathematics courses. We find that for students with less mathematics background (i.e., those who needed to take Calculus 2 at the university), taking the new computational linear algebra course has significant positive impact on their average exam performance in the subsequent course. The performance of students with more initial mathematics background (i.e., those who already had credit for Calculus 2) is not significantly affected by the computational vs. traditional course backgrounds.

This Research Full Paper presents an observation protocol to explore group processing in cooperative learning. Use of structured roles, such as Process Oriented Guided Inquiry Learning (POGIL) and pair programming, can help facilitate cooperative learning and help courses scale to large classroom sizes, decrease attrition and failure rates, and improve student performance. Our observation protocol was created to capture how groups work together in online, POGIL-inspired activities. A team of graduate student researchers developed the observation protocol for a variety of courses by observing three different computer science courses during the Spring 2021 semester. A total of 77 groups across all three courses were recorded, and percent agreement using a subset of the recordings suggested good interrater reliability (91.29%). We also extend a previous equality metric to quantify the rates of student participation, and found that it offered good differentiation between groups where one student contributed the most and groups where students contributed equally. We present example applications of our observation protocol related to general participation trends, the kinds of contributions students make, student-student bonding, and help-seeking patterns. Finally, we discuss future directions for use of our coding scheme as well as implications for implementing structured role-based cooperative learning online in the future.