Combining evidence and meta-analysis in software engineering

Abstract

Recently there has been a welcome move to realign software engineering as an evidence-based practice. Many research groups are actively conducting empirical research e.g. to compare different fault prediction models or the value of various architectural patterns. However, this brings some challenges. First, for a particular question, how can we locate all the relevant evidence (primary studies) and make sense of them in an unbiased way. Second, what if some of these primary studies are inconsistent? In which case how do we determine the 'true' answer? To address these challenges, software engineers are looking to other disciplines where the systematic review is normal practice (i.e. systematic, objective, transparent means of locating, evaluating and synthesising evidence to reach some evidence-based answer to a particular question). This chapter examines the history of empirical software engineering, overviews different meta-analysis methods and then describe the process of systematic reviews and conclude with some future directions and challenges for researchers. © Springer-Verlag 2013.