Improve student attainment and slash your marking time with GitHub Classroom and Travis CI

The design of meaningful authentic assessments is extremely important in education. Students are much more likely to actively engage with assignments that reflect industry practices and assess the knowledge, skills, and technical competencies that employers’ value. This is particularly true in computer programming where students are often required to analyse complex problems, design an appropriate solution, and implement it in code. While assessing a student’s understanding of the underlying theory and core principles or programming can be done using online tests in a VLE, such an approach provides the educator with a limited opportunity to assess the student’s mastery of the higher levels of Bloom’s taxonomy (analyse, evaluate, and create).

To comprehensively assess a student’s ability to program we must devise assessments that require the students to design, implement and evaluate software solutions. This however presents several challenges for the educator. How do we devise an appropriate marking rubric? How can we provide students with continuous feedback and guidance? Marking individual student submissions consisting of thousands of lines of complex code is extremely time consuming. How can we efficiently assess large student cohorts?

In this presentation Dr. Shane Wilson will discuss how he used Github Classroom and Travis CI to build an automated marking and feedback system that boosted his students’ performance and slashed the time he spent marking. The system automatically provided programming students with an indicative mark and feedback within minutes of them submitting a draft assignment submission. This continuous feedback loop offered students guidance and hints on how to improve their draft coding submission, in turn providing them with the opportunity to reflect, evaluate and refine their approach. This approach increased retention from 76 to 90 percent while decreasing the overall time spent marking by 80 percent.