Skip to main content
@shmVirus

publications

Integrating Continuous Integration Practices into Software Engineering Curricula: A Multi-Institution Study

Sabbir Hosen Mamun†, Joris T. Vandenberghe, Daniel A. Fernández† Corresponding author

Journal of Systems and Software, Vol. 182, pp. 111043 · ElsevierQ1, IF 3.9, Scopus

Journal202147 citations
PDFDOICodeDatasetCite

Abstract

Continuous integration (CI) practices are standard in industry but poorly represented in undergraduate curricula. This paper reports a multi-institution study spanning six universities, 18 course sections, and 742 students, evaluating the impact of embedding CI tools — automated build, lint, and test pipelines — into a second-year software engineering course. Groups using CI submitted more frequently, had lower defect rates on final deliverables (–23%), and self-reported higher confidence in collaborative development workflows. Regression analysis shows that early adoption of CI in the first two weeks of a project is the strongest predictor of on-time delivery (β = 0.41, p < 0.001). We provide a full replication package including course materials and anonymised grade and commit data.

Cite this work

@article{Mamun2021Integrating,
  title = {Integrating Continuous Integration Practices into Software Engineering Curricula: A Multi-Institution Study},
  author = {Sabbir Hosen Mamun and Joris T. Vandenberghe and Daniel A. Fernández},
  year = {2021},
  journal = {Journal of Systems and Software},
  volume = {182},
  pages = {111043},
  publisher = {Elsevier},
  doi = {10.1016/j.jss.2021.111043},
  abstract = {Continuous integration (CI) practices are standard in industry but poorly represented in undergraduate curricula. This paper reports a multi-institution study spanning six universities, 18 course sections, and 742 students, evaluating the impact of embedding CI tools — automated build, lint, and test pipelines — into a second-year software engineering course. Groups using CI submitted more frequently, had lower defect rates on final deliverables (–23%), and self-reported higher confidence in collaborative development workflows. Regression analysis shows that early adoption of CI in the first two weeks of a project is the strongest predictor of on-time delivery (β = 0.41, p < 0.001). We provide a full replication package including course materials and anonymised grade and commit data.},
}