DETERMINANTS OF INSPECTION EFFECTIVENESS IN SOFTWARE DEVELOPMENT: AN EMPIRICAL ANALYSIS

Software inspections are formal evaluations of the intermediate work products (artifacts) of the development process. These artifacts are examined to ensure that a high quality work-product is delivered to the testers and ultimately to the end-users of the software product. The crucial role of inspections in determining quality of the software makes it important to assess the effectiveness of inspections. While prior research has identified several factors that influence effectiveness of software inspections, our understanding of the influence of team composition (personnel mix and team size) and the type of the inspected artifact (project plan, requirements specification, design document, code) on effectiveness of inspections is minimal. We develop hypotheses for the factors affecting inspection effectiveness and attempt to validate these hypotheses in a field setting. Our preliminary results show that, during early stages of software development, an increase in the proportion of experienced reviewers (with greater domain experience) is associated with both an increase in the total number of defects discovered in the inspection process as well as an increase in the likelihood of detecting high severity defects. However, during later stages, we find that greater programming experience is associated with both an increase in the total number of defects discovered in the inspection process as well as an increase in the likelihood of detecting high severity defects. These results have important implications for both practice and research.