arXiv subjects

Natalie Kiesler

Publications and source records attributed to Natalie Kiesler.

21 records · Page 2Linked to original sources

arXiv2023

Large Language Models in Introductory Programming Education: ChatGPT's Performance and Implications for Assessments

This paper investigates the performance of the Large Language Models (LLMs) ChatGPT-3.5 and GPT-4 in solving introductory programming tasks. Based on the performance, implications for didactic scenarios and assessment formats utilizing LLMs are derived. For the analysis, 72 Python tasks for novice programmers were selected from the free site CodingBat. Full task descriptions were used as input to the LLMs, while the generated replies were evaluated using CodingBat's unit tests. In addition, the general availability of textual explanations and program code was analyzed. The results show high scores of 94.4 to 95.8% correct responses and reliable availability of textual explanations and program code, which opens new ways to incorporate LLMs into programming education and assessment.

cs.SE↗

arXiv2022

An Exploratory Analysis of Feedback Types Used in Online Coding Exercises

Online coding environments can help support computing students gain programming practice at their own pace. Especially informative feedback can be beneficial during such self-guided, independent study phases. This research aims at the identification of feedback types applied by CodingBat, Scratch and Blockly. Tutoring feedback as coined by Susanne Narciss along with the specification of subtypes by Keuning, Jeuring and Heeren constitute the theoretical basis. Accordingly, the five categories of elaborated feedback (knowledge about task requirements, knowledge about concepts, knowledge about mistakes, knowledge about how to proceed, and knowledge about meta-cognition) and their subtypes were utilized for the analysis of available feedback options. The study revealed difficulties in identifying clear-cut boundaries between feedback types, as the offered feedback usually integrates more than one type or subtype. Moreover, currently defined feedback types do not rigorously distinguish individualized and generic feedback. The lack of granularity is also evident in the absence of subtypes relating to the knowledge type of the task. The analysis thus has implications for the future design and investigation of applied tutoring feedback. It encourages future research on feedback types and their implementation in the context of programming exercises to define feedback types that match the demands of novice programmers.

cs.HC↗

arXiv2020

Zur Modellierung und Klassifizierung von Kompetenzen in der grundlegenden Programmierausbildung anhand der Anderson Krathwohl Taxonomie

This research paper focusses on the competences expected from computer science novices in the domain of basic programming and how they can be classified. By means of a qualitative content analysis of current learning objectives at German universities and the perspective of university teachers, basic programming competencies are identified. Since the competency model proposed by the German Society of Computer Science (GI) reveals several deficits, competencies are classified along the Anderson Krathwohl Taxonomy (AKT) of learning, teaching and assessing. As a result, dimensions and subtypes of the AKT are revised towards a model specific to computer science aiming at the classification of programming competencies according to their cognitive complexity and knowledge dimension. The adaptation of the educational model can thereby help standardize curricula, and develop assessments and corresponding items in the future.

cs.CY↗