An official website of the United States government
Creating equitable performance outcomes among students is a focus of many instructors and researchers. One focus of this effort is examining disparities in physics student performance across genders, which is a well-established problem. Another less common focus is disparities across racial and ethnic groups, which may have received less attention due to low representation rates making it difficult to identify gaps in their performance. In this investigation we examined associations between Learning Assistant (LA) supported courses and improved equity in student performance. We built Hierarchical Linear Models of student performance to investigate how performance differed by gender and by race/ethnicity and how LAs may have moderated those differences. Data for the analysis came from pre-post concept inventories in introductory mechanics courses collected through the Learning About STEM Student Outcomes (LASSO) platform. Our models show that gaps in performance across genders and races/ethnicities were similar in size and increased from pre to post instruction. LA-support is meaningfully and reliably associated with improvement in overall student performance but not with shifts in within-course performance gaps.
more »
« less
On learning platform metrics as markers for student success in a course
Abstract Adaptive learning platforms are increasingly being used as part of varying instructional modalities. Particularly relevant to this paper, adaptive learning is a critical component of personalized, preclass learning in a flipped classroom. Previously inaccessible, data generated by adaptive learning platforms regarding student engagement with the course content provides an invaluable opportunity to gain a deeper understanding of the learning process and improve upon it. We aim to investigate the relationships between adaptive learning platform interactions and overall student success in the course and identify the variables most influential to student success. We present a comprehensive analysis of our adaptive learning platform data collected in a Numerical Methods course, including aggregate statistics, frequency analysis, and Principal Component Analysis, to determine which variables exhibited the most variability and, therefore, the most information in the data. Subsequently, we used the Partitioning Around Medoids clustering approach to investigate naturally occurring clusters of students and how these clusters relate to overall performance in the course. Our results show that overall performance in the course, as measured by the final course grade, is strongly associated with (1) the behavioral interactions of students with the adaptive platform and (2) their performance on the adaptive learning assessments. We also found distinct student clusters (as defined by success in the course) that exhibited distinctly different behaviors. These findings provide qualitative and quantitative information to identify students needing support and to craft an evidence‐based support strategy for these students.
more »
« less
- Award ID(s):
- 2013271
- PAR ID:
- 10418446
- Publisher / Repository:
- Wiley Blackwell (John Wiley & Sons)
- Date Published:
- Journal Name:
- Computer Applications in Engineering Education
- Volume:
- 31
- Issue:
- 5
- ISSN:
- 1061-3773
- Format(s):
- Medium: X Size: p. 1412-1432
- Size(s):
- p. 1412-1432
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
The ability to predict student performance in introductory programming courses is important to help struggling students and enhance their persistence. However, for this prediction to be impactful, it is crucial that it remains transparent and accessible for both instructors and students, ensuring effective utilization of the predicted results. Machine learning models with explainable features provide an effective means for students and instructors to comprehend students' diverse programming behaviors and problem-solving strategies, elucidating the factors contributing to both successful and suboptimal performance. This study develops an explainable model that predicts student performance based on programming assignment submission information in different stages of the course to enable early explainable predictions. We extract data-driven features from student programming submissions and utilize a stacked ensemble model for predicting final exam grades. The experimental results suggest that our model successfully predicts student performance based on their programming submissions earlier in the semester. Employing SHAP, a game-theory-based framework, we explain the model's predictions, aiding stakeholders in understanding the influence of diverse programming behaviors on students' success. Additionally, we analyze crucial features, employing a mix of descriptive statistics and mixture models to identify distinct student profiles based on their problem-solving patterns, enhancing overall explainability. Furthermore, we dive deeper and analyze the profiles using different programming patterns of the students to elucidate the characteristics of different students where SHAP explanations are not comprehensible. Our explainable early prediction model elucidates common problem-solving patterns in students relative to their expertise, facilitating effective intervention and adaptive support.more » « less
-
The ability to predict student performance in introductory programming courses is important to help struggling students and enhance their persistence. However, for this prediction to be impactful, it is crucial that it remains transparent and accessible for both instructors and students, ensuring effective utilization of the predicted results. Machine learning models with explainable features provide an effective means for students and instructors to comprehend students' diverse programming behaviors and problem-solving strategies, elucidating the factors contributing to both successful and suboptimal performance. This study develops an explainable model that predicts student performance based on programming assignment submission information in different stages of the course to enable early explainable predictions. We extract data-driven features from student programming submissions and utilize a stacked ensemble model for predicting final exam grades. The experimental results suggest that our model successfully predicts student performance based on their programming submissions earlier in the semester. Employing SHAP, a game-theory-based framework, we explain the model's predictions, aiding stakeholders in understanding the influence of diverse programming behaviors on students' success. Additionally, we analyze crucial features, employing a mix of descriptive statistics and mixture models to identify distinct student profiles based on their problem-solving patterns, enhancing overall explainability. Furthermore, we dive deeper and analyze the profiles using different programming patterns of the students to elucidate the characteristics of different students where SHAP explanations are not comprehensible. Our explainable early prediction model elucidates common problem-solving patterns in students relative to their expertise, facilitating effective intervention and adaptive support.more » « less
-
The ability to predict student performance in introductory programming courses is important to help struggling students and enhance their persistence. However, for this prediction to be impactful, it is crucial that it remains transparent and accessible for both instructors and students, ensuring effective utilization of the predicted results. Machine learning models with explainable features provide an effective means for students and instructors to comprehend students' diverse programming behaviors and problem-solving strategies, elucidating the factors contributing to both successful and suboptimal performance. This study develops an explainable model that predicts student performance based on programming assignment submission information in different stages of the course to enable early explainable predictions. We extract data-driven features from student programming submissions and utilize a stacked ensemble model for predicting final exam grades. The experimental results suggest that our model successfully predicts student performance based on their programming submissions earlier in the semester. Employing SHAP, a game-theory-based framework, we explain the model's predictions, aiding stakeholders in understanding the influence of diverse programming behaviors on students' success. Additionally, we analyze crucial features, employing a mix of descriptive statistics and mixture models to identify distinct student profiles based on their problem-solving patterns, enhancing overall explainability. Furthermore, we dive deeper and analyze the profiles using different programming patterns of the students to elucidate the characteristics of different students where SHAP explanations are not comprehensible. Our explainable early prediction model elucidates common problem-solving patterns in students relative to their expertise, facilitating effective intervention and adaptive support.more » « less
-
Dr. Alice Suroviec (Ed.)Approaches to student-centered active learning have evolved. The progression in course-design has led to the development of new learning paradigms such as collaborative, problem based, and project-based learning. Course-based undergraduate research experiences (CUREs) are a learning pedagogy that infuses research experiences within the curriculum. This method of instruction increases opportunities for students to participate in more authentic education experiences and is especially beneficial in the science education pathway. CUREs encourage students to be autonomous and emphasize teamwork. Our research proposes methodologies that can maximize student performance, particularly benefiting underrepresented and underprepared female students. Pre- and post- assessments of a CURE classroom were administered to gauge student engagement and success in a General Chemistry course. Specifically, our research focuses on female engagement in CURE projects and overall success and retention rates to test if the teaching methods will support increased gender equity in STEM.more » « less
