Analysis of human-centric software engineering experiments: a systematic mapping study

Abstract

Software Engineering professionals need to have information about new support mechanisms to decide, not at random, what option is best adapting it needs. One way of obtaining this information is through empirical studies that make the mechanisms used to support the design and development of software, be evaluated in practice. Thus, scientific research through experiments and empirical studies are fundamental during the evaluation of any new technology to software development. In this context, researchers perform experiments to check their proposals under controlled conditions. Therefore, experiments are an important category of empirical studies and are the classical approach for identifying cause-effect relationships. The goal of this dissertation is qualitatively and quantitatively characterizes and analyze human-centric experiments in software engineering, published in three journals andthree conferences proceedings from 2003 to 2013. To reach this objective was performed a systematic mapping study that includes all full papers published at EASE, ESEM, ICSE, ESEJ, JSS, TSE. Based on manual searches in those important conferences and journals in Software Engineering, were analyzed 3671 papers. 244 primary studies were identified as relevant, reporting experiments. In these experiments, we obtained qualitative and quantified data about authors and institutions, subjects, tasks performed, environment, replication and threats to validity. From the analysis performed, this work conclude that despite guidelines exist now, there is a large gap in the report of the experiments. The main contribution of this work is to provide the reporting status of human-centric software engineering experiments and how this field has matured. This work also proposes a list of information that an experiment report expected to have.