Jennifer Bivens
ABSTRACT
In the last decade, data analytics has been successfully applied in the field of education to predict student performance. There exists an obvious opportunity for this educational data to make a positive impact on computer science instruction. Machine learning models can use historical data containing behavioral and education-related attributes, such as previous course work, grades and time spent in class discussions, to make predictions about academic performance for prospective students. Even with proven predictive success, many questions related to the application of performance prediction remain unanswered, particularly in the context of larger debates about risk identification, grouping, and bias. This BoF will provide a platform for exploring the following questions: (a) How should computer science instructors use prediction data? Could results be used to group students by predicted academic performance levels? Could predictions help in the identification of students with low performance predictions for additional mentoring? (b) Should predictions be shared with students/instructors? (c) If so, how could instructor bias resulting from these predictions be minimized to ensure fair evaluation of students' actual performance? (d) Do computer science instructors attending this BoF currently implement any predictive tools or risk grouping? Would they consider doing either? (e) How much importance would instructors place on the results of performance predictions? To what degree would the accuracy of a model affect adoption?
Index Terms
- Pros and Cons of Using Data Analytics for Predicting Academic Performance in Computer Science Courses: (Abstract Only)
Recommendations
Building a Community of First Year Students Improves Student Retention and Performance in Computing Courses: (Abstract Only)
SIGCSE '18: Proceedings of the 49th ACM Technical Symposium on Computer Science EducationFirst-year, college-level computing courses are gateway courses with low passing rates, resulting in student attrition and transfers out of computer science degrees. Learning Communities (LC) are a group of students who enroll in two or more courses, ...
Bringing security into traditional computer science courses: challenges and support (abstract only)
SIGCSE '13: Proceeding of the 44th ACM technical symposium on Computer science educationIn today's world, it is critical that our students get exposure to security education as part of their Computer Science (CS) curriculum. The lack of security consciousness in digital workforce is costing businesses, government, and citizens severely ...
Using student performance predictions in a computer science curriculum
ITICSE '06: Proceedings of the 11th annual SIGCSE conference on Innovation and technology in computer science educationProfessors often develop anecdotal guidelines about how each student's past performance in their academic major relates to their performance in later courses. While these guidelines can be useful, a more formal statistical analysis of these ...
Comments