Abstract:Background: Programming skills are advantageous to navigate today's society, so it is important to teach them to students. However, failure rates for programming courses are high, and especially students who fall behind early in introductory programming courses tend to stay behind. Objective: To catch these students as early as possible, we aim to develop an early warning system, so we can offer the students support, for example, in the form of syntax drill-and-practice exercises. Method: To develop the early warning system, we assess different cognitive skills of students of an introductory programming course. On several points in time over the course, students complete tests that measure their ability to develop a mental model of programming, language skills, attention, and fluid intelligence. Then, we evaluated to what extent these skills predict whether students acquire programming skills. Additionally, we assess how syntax drill-and-practice exercises improve how students acquire programming skill. Findings: Most of the cognitive skills can predict whether students acquire programming skills to a certain degree. Especially the ability to develop an early mental model of programming and language skills appear to be relevant. Fluid intelligence also shows predictive power, but appears to be comparable with the ability to develop a mental model. Furthermore, we found a significant positive effect of the syntax drill-and-practice exercises on the success of a course. Implications: Our first suggestion of an early warning system consists of few, easy-to-apply tests that can be integrated in programming courses or applied even before a course starts. Thus, with the start of a programming course, students who are at high risk of failing can be identified and offered support, for example, in the form of syntax drill-and-practice exercises to help students to develop programming skills.

Pass/Fail Prediction in Programming Courses

Predicting Student Performance in Introductory Programming Courses

Toward Finding and Supporting Struggling Students in a Programming Course with an Early Warning System

A Simple, Language-Independent Approach to Identifying Potentially At-Risk Introductory Programming Students.

Deep learning for early performance prediction of introductory programming students: a comparative and explanatory study

Scale Up Predictive Models for Early Detection of At-Risk Students: a Feasibility Study

An Artificial Neural Network Based Early Prediction of Failure-Prone Students in Blended Learning Course

Is Initial Performance in a Course Informative? Machine Learning Algorithms as Aids for the Early Detection of At-Risk Students

Predicting Students Success in Blended Learning—Evaluating Different Interactions Inside Learning Management Systems

Contrasting prediction methods for early warning systems at undergraduate level

Course Success Prediction and Early Identification of At-Risk Students Using Explainable Artificial Intelligence

Programming Course Student Performance Prediction Based on Feature Construction

Predicting students' performance of an offline course from their online behaviors

Artificial neural network model to predict student performance using nonpersonal information

A Machine Learning Approach for Tracking and Predicting Student Performance in Degree Programs

Behavior Pattern and Compiled Information Based Performance Prediction in MOOCs

Predicting Early Dropout: Calibration and Algorithmic Fairness Considerations

Predicting Course Performance by Check-in Data in Classrooms

Multi-Class Phased Prediction of Academic Performance and Dropout in Higher Education

Real-time learning analytics for C programming language courses

An early warning system to identify and intervene online dropout learners