Systematic reviews

Abstract

During the past few years, I have become interested in evidence-based software engineering [5]. The evidence-based paradigm began in medicine and has totally revolutionised medical research [7]. It is now being adopted by a number of humancentered disciplines such as nursing, social policy, education etc. Although I have some reservations about whether software engineering can adopt fully the evidencebased paradigm, I am convinced that we should adopt certain aspects of the approach immediately. In particular, I believe we should adopt systematic review in place of ad hoc literature reviews, and recognise that a systematic review is a research method in its own right. Currently, all PhD students need to conduct a literature review as a part of their research but such reviews are seldom performed with the rigour now being required in other disciplines (including Information Systems Research, [6]). A systematic review is rigorous method of synthesising existing research related to a topic of interest. Quantitative meta-analysis is one form of systematic review, but it is not mandatory to provide a quantitative summary. Systematic reviews aim to provide a synthesis that is complete (with respect to defined criteria) and unbiased. Systematic reviews must be rigorous i.e. be conducted according to a well-defined procedure and open i.e. the research process must be reported in a manner is understandable to other researchers.