자료유형 | 학위논문 |
---|---|
서명/저자사항 | Assessment of Students' Computer Science Focal Knowledge, Skills, and Abilities in Game-Based Learning Environments. |
개인저자 | Akram, Bita. |
단체저자명 | North Carolina State University. |
발행사항 | [S.l.]: North Carolina State University., 2019. |
발행사항 | Ann Arbor: ProQuest Dissertations & Theses, 2019. |
형태사항 | 132 p. |
기본자료 저록 | Dissertations Abstracts International 81-02B. Dissertation Abstract International |
ISBN | 9781085645560 |
학위논문주기 | Thesis (Ph.D.)--North Carolina State University, 2019. |
일반주기 |
Source: Dissertations Abstracts International, Volume: 81-02, Section: B.
Advisor: Carrier, Sarah |
이용제한사항 | This item must not be sold to any third party vendors. |
요약 | Computer science has become a critical competency for all students. With national K-12 education initiatives such as CSforAll, block-based programming environments have emerged as widely used tools for teaching computational thinking (CT) and computer science (CS) concepts to novice programmers. A key challenge posed by block-based programming environments is assessing students' CT and programming competencies within these environments. While this assessment role would traditionally be fulfilled by a teacher, the dramatic growth of computer science as both a necessary skill and area of interest for students presents the need for automated and semi-automated assessment tools that support teachers in this role. Developing assessment methods that can evaluate students' CS focal knowledge, skills, and abilities help teachers evaluate students' learning and provide appropriate scaffolding.It is also important to foster students' motivation and engagement toward learning CT and programming. Game-based learning environments utilize the motivational elements of games to promote students' learning and engagement. Studies have shown that for a broad range of student populations and subject matters, students who engage in game-based learning experience higher motivation compared to those who learn with conventional methods. Intelligent game-based learning environments integrate the adaptive learning support of intelligent tutoring systems and the motivational elements of games to provide learners with adaptive scaffolding without interfering with their state of flow. As a result, intelligent gamebased learning environments have emerged as a promising tool that can provide students with engaging learning opportunities while providing educators with ample opportunity to unobtrusively assess students' knowledge and skills.This dissertation introduces a novel temporal analytics framework for stealth assessment based on students' problem-solving strategies. The strategy-based temporal analytic framework incorporates long short-term memory networks to analyze students' problem-solving behaviors across consecutive tasks to inform an evidence model used in stealth assessment. When evaluating the temporal analytic framework on a dataset collected from middle grade students' interactions with the ENGAGE game-based learning environment, results show that it outperforms competitive baseline models with respect to predictive accuracy when used for predicting students' post-test scores.Furthermore, this dissertation presents the development and evaluation of a stealth assessment framework that utilizes a blended hypotheses-driven assessment design approach to assess middle grade students' CS focal knowledge, skills and abilities (FKSAs) within game-based learning environments. We follow an Evidence-centered design-based assessment to evaluate students' CS FKSAs based on evidence extracted from their programming trajectories in a block-based programming environment. The results reveal distinctive patterns in students' approaches to problem solving for CT challenges, which provides first steps toward identifying and assessing productive CS practices.Finally, this dissertation presents a semi-automated assessment framework that utilizes a corpus of graded programming artifacts to infer students' ability to develop a bubble sort algorithm based on their submitted programming artifacts. The dataset containing graded submissions is prone to noise due to graders' subjective idea about the quality of students' submitted artifacts. Utilizing a systematic approach to labeling the training dataset and applying Gaussian process regression help reduce this noise. The semi-automated assessment framework utilizes a supervised learning approach that infers the algorithmic quality of a submitted programming artifact based on its hierarchical and ordinal encoded n-grams. Our results demonstrate the effectiveness of the proposed approach in inferring artifacts' algorithmic quality. Furthermore, according to our results, Gaussian process models outperform other models that are unable to accommodate the level of noise in our dataset. Overall, our stealth assessment framework has shown to be an effective approach to unobtrusively assess students' CS and CT competencies within a game-based learning environment. |
일반주제명 | Computer science. |
언어 | 영어 |
바로가기 | ![]() |