Ever since Large Language Models (LLMs) and related applications have become broadly available, several studies investigated their potential for assisting educators and supporting students in higher education. LLMs such as Codex, GPT-3.5, and GPT 4 have shown promising results in the context of large programming courses, where students can benefit from feedback and hints if provided timely and at scale. This paper explores the quality of GPT-4 Turbo's generated output for prompts containing both the programming task specification and a student's submission as input. Two assignments from an introductory programming course were selected, and GPT-4 was asked to generate feedback for 55 randomly chosen, authentic student programming submissions. The output was qualitatively analyzed regarding correctness, personalization, fault localization, and other features identified in the material. Compared to prior work and analyses of GPT-3.5, GPT-4 Turbo shows notable improvements. For example, the output is more structured and consistent. GPT-4 Turbo can also accurately identify invalid casing in student programs' output. In some cases, the feedback also includes the output of the student program. At the same time, inconsistent feedback was noted such as stating that the submission is correct but an error needs to be fixed. The present work increases our understanding of LLMs' potential, limitations, and how to integrate them into e-assessment systems, pedagogical scenarios, and instructing students who are using applications based on GPT-4.

Generative AI (GenAI) has seen great advancements in the past two years and the conversation around adoption is increasing. Widely available GenAI tools are disrupting classroom practices as they can write and explain code with minimal student prompting. While most acknowledge that there is no way to stop students from using such tools, a consensus has yet to form on how students should use them if they choose to do so. At the same time, researchers have begun to introduce new pedagogical tools that integrate GenAI into computing curricula. These new tools offer students personalized help or attempt to teach prompting skills without undercutting code comprehension. This working group aims to detail the current landscape of education-focused GenAI tools and teaching approaches, present gaps where new tools or approaches could appear, identify good practice-examples, and provide a guide for instructors to utilize GenAI as they continue to adapt to this new era.

Learning programming is perceived as hard by many students. To support students, many e-assessment and intelligent tutoring systems have been developed. These systems can automatically evaluate student submissions and provide feedback. Despite comprehensive research on feedback modalities, little is known about students' confidence in the correctness of their submissions when requesting feedback. Also, educators hope that students get more confident and that automatically provided feedback helps students to improve and better self-assess their work. In this paper, first-semester students are asked about their confidence in passing a requested syntax or function test before the test results are revealed to them. Students can request feedback from each of the two provided test types twice in arbitrary order. The self-rated confidence, test outcomes, time needed to enter the confidence, and correlations are analyzed in detail. The results show that the majority of students has a high confidence in their submitted work. However, students frequently over-estimate the correctness and only few under-estimate it. There is a correlation between students' confidence in their submissions and their actual performance, but this cannot be used to make reliable predictions. The test pass rate for highly confident students is higher for syntax tests than for function tests and students need more time for entering their confidence for syntax than for function tests. Over the semester, the self-rated confidence decreases. When tests are reattempted, both correctness and self-assessment abilities show improvement.

Programming courses in higher education are often attended by several hundred students. In such large-scale courses, direct instruction is often the last resort, resulting in mostly passive students and limited social interaction. The instructor may present worked examples or perform live coding and the students try to reproduce these on their devices. However, there is usually no live coding where students work on small programming assignments themselves directly in class, because the instructor does not have a timely/rapid overview of the most important common issues that are prevalent in class to support the students. This paper presents experiences with a teaching format to activate students that addresses the aforementioned issues. After students have worked on a small assignment and uploaded their solution attempt within a specified time period, an extended e-assessment system analyzes all submissions and instantly provides an overview of the number of correct submissions as well as all common errors and their frequency to assist the instructor in the immediate tailored discussion. This approach makes it possible to engage students, make student performance visible to all participants, and discuss the most common errors. Students liked "their" live coding, the discussion of "their" errors, and want to do it more often, although not many students uploaded their solution attempts. The teaching scenario, benefits, pitfalls, possible improvements, and further application scenarios are discussed.

Peer review can be used as a collaborative learning activity in which people with similar competencies evaluate other students' submissions and/or provide feedback. It provides many potential benefits such as timely feedback, high motivation, reduced workload for teachers, collaboration among the students, improving the code, and seeing other solution strategies. However, there are also challenges and contradictory results such as low motivation, participation, quality, and no improvements in the reviews. This article attempts to shed more light on these issues through an empirical investigation in a university-based introductory programming course with approx. 900 students. In the evaluation, this paper empirically investigates the effects of reviewing other solutions on the view of one's own solution and how students can be motivated to regularly work on voluntary homework assignments. Furthermore, there is an analysis of the peer reviews regarding their quality (length and correctness), and the students' participation and perceptions. The results indicate that giving feedback can change the view on one's own submission regarding the complete correctness, the majority of feedback is rather short, peer review assignments are a major driver for working on the assignments, and the majority of students like seeing other solutions. The majority of students seems to be able to identify correct submissions as correct, however, (partly) incorrect submissions are also often classified as completely correct. Possible measures to address these weaknesses are discussed.

Career choice is a life-changing decision for young adults. Online career guidance systems facilitate information access and provide recommendations for career paths. However, they bear the risk of imposed decisions, as they often neglect the non-cognitive processes of career decision-making. Mobile apps are promising for supporting career choices, since smartphones are a constant companion of adolescents. Following a user-centered design process, we developed “FindMyself”, a mobile app helping users discover their personal strengths and interests. FindMyself includes multiple sessions, prompting users with short challenges and reflection tasks building a personal profile. Our approach leaves time for self-reflection and discussion with peers. In a field study (N = 46), we found that FindMyself was perceived as easy, flexible, and fun to use, and was able to support self-reflection. For future iterations, some users requested specific career recommendations. We discuss these results and derive implications for further research on career decision support tools.

Traditional recommender systems are well established in scenarios in which "enough"items, users and ratings are available for the algorithms to operate on. However, automatic recommendations are also desirable in smaller online communities which only contain several hundred items and users. Collaborative filters, as one of the most successful technologies for recommender systems, do not perform well here. This paper argues that recommender systems can make use of contextual information and domain specific semantics in order to be able to generate recommendations also for these smaller usage scenarios. The new hybrid recommendation approach presented in the paper enhances traditional neighborhood-based collaborative filtering techniques through the use of new kinds of data and a combination of different recommendation methods (rule, demographic, and average based). While the algorithmic techniques presented in this paper are suitable (especially) for smaller online communities, they can also be applied to improve the quality of recommendations in larger communities. The approach was implemented and evaluated in a small regional bound parent education community. A multi-staged evaluation was conducted in order to determine the quality of recommendations: A cross-validation (recall), an expert questionnaire (recommendation quality) and a field study (user satisfaction). The results show that recommenders even for smaller communities are possible and can produce high quality recommendations.