一句话：系统梳理软件工程系统文献综述中定性研究综合方法的使用现状，揭示方法理解缺口与选择因素。
该文关注软件工程（SE）中证据驱动实践（EBSE）与系统文献综述（SLR）对定性数据综合（QRS）方法的需求。作者系统审查了2005至2015年间发表的328篇SLR，并向274篇SLR的作者发函确认是否使用QRS方法，获得116份有效回复。分析发现SE研究中实际使用了八种QRS方法，其中叙事综合（narrative synthesis）和主题综合（thematic synthesis）最常被采用。研究还表明，社区对已定义QRS方法的认识存在明显缺失和不完整，且识别出一组可能影响QRS方法选择与使用的初步因素。
Synthesizing data extracted from primary studies is an integral component of the methodologies in support of Evidence Based Software Engineering (EBSE) such as System Literature Review (SLR). Since a large and increasing number of studies in Software Engineering (SE) incorporate qualitative data, it is important to systematically review and understand different aspects of the Qualitative Research Synthesis (QRS) being used in SE. We have reviewed the use of QRS methods in 328 SLRs published between 2005 and 2015. We also inquired the authors of 274 SLRs to confirm whether or not any QRS methods were used in their respective reviews. 116 of them provided the responses, which were included in our analysis. We found eight QRS methods applied in SE research, two of which, narrative synthesis and thematic synthesis, have been predominantly adopted by SE researchers for synthesizing qualitative data. Our study determines that a significant amount of missing knowledge and incomplete understanding of the defined QRS methods in the community. Our effort also identifies an initial set factors that may influence the selection and use of appropriate QRS methods in SE.
DOI 原文 ·
@inproceedings{paperbot4438,
title = {Synthesizing qualitative research in software engineering},
author = {Xin Huang and He Zhang and Xin Zhou and Muhammad Ali Babar and Song Yang},
booktitle = {ICSE},
year = {2018},
doi = {10.1145/3180155.3180235}
}