Training students in evidence-based software engineering and systematic reviews: a systematic review and empirical study

Context Although influential in academia, evidence-based software engineering (EBSE) has had little impact on industry practice. We found that other disciplines have identified lack of training as a significant barrier to Evidence-Based Practice. Objective To build and assess an EBSE training proposal suitable for students with more than 3 years of computer science/software engineering university-level training. Method We performed a systematic literature review (SLR) of EBSE teaching initiatives and used the SLR results to help us to develop and evaluate an EBSE training proposal. The course was based on the theory of learning outcomes and incorporated a large practical content related to performing an SLR. We ran the course with 10 students and based course evaluation on student performance and opinions of both students and teachers. We assessed knowledge of EBSE principles from the mid-term and final tests, as well as evaluating the SLRs produced by the student teams. We solicited student opinions about the course and its value via a student survey, a team survey, and a focus group. The teachers’ viewpoint was collected in a debriefing meeting. Results Our SLR identified 14 relevant primary studies. The primary studies emphasized the importance of practical examples (usually based on the SLR process) and used a variety of evaluation methods, but lacked any formal education methodology. We identified 54 learning outcomes covering aspects of EBSE and the SLR method. All 10 students passed the course. Our course evaluation showed that a large percentage of the learning outcomes established for training were accomplished. Conclusions The course proved suitable for students to understand the EBSE paradigm and to be able to apply it to a limited-scope practical assignment. Our learning outcomes, course structure, and course evaluation process should help to improve the effectiveness and comparability of future studies of EBSE training. However, future courses should increase EBSE training related to the use of SLR results.