Using rapid review to support software engineering practice: a citation analysis and a replication study.

Abstract

Context: A few years ago rapid reviews (RR) were introduced in software engineering (SE) to address problems that standard systematic reviews take too long and too much effort to be of value to practitioners. Prior to our study, few practice-driven RRs had been reported and none of them considered collaborating with RRs for practitioners lacking SE research experience. Objective: To investigate practitioners’ perspectives on the use of RRs in supporting SE practices, we aimed to validate and build upon the findings of the seminal RR in SE study, specifically considering practitioners without explicit SE research experience. Method: First, we studied previously conducted RRs in SE using citation analysis. Second, we carried out an external replication of the first study that proposed the use of RRs in SE. Specifically, we conducted an RR for an agile software development team looking to improve its knowledge management practices. Results: Most of the participants’ perceptions about RR results were positive and strongly consistent with previous research. In particular, RR results were considered more reliable than other sources of information and adequate to address the problems detected. Some months later they confirmed using some of the recommendations. Conclusions: The results show that practitioners without explicit SE research experience appreciate the value of evidence and can make use of the results of RRs. However, SE research may need to be translated from broad recommendations to specific process change options. Our research also reveals that SE RRs reporting needs to be substantially improved.