This report proposes a guideline for systematic reviews in software engineering research. A systematic review is a method to evaluate and interpret all available research relevant to a specific research question, topic, or phenomenon. Systematic reviews aim to provide a fair evaluation of a research topic using a trustworthy, rigorous, and auditable methodology. The guideline is adapted from three existing medical guidelines to suit software engineering research. It covers three phases: planning, conducting, and reporting the review. It is at a high level and does not consider the impact of question type on review procedures or specify mechanisms for meta-analysis.
The report introduces the concept of systematic reviews as a significant research method for software engineering researchers. It discusses the importance of systematic reviews in ensuring thorough and unbiased literature reviews. Systematic reviews are more rigorous than traditional reviews and provide information about the effects of phenomena across various settings and methods. They can also help identify gaps in research and support the generation of new hypotheses.
Systematic reviews involve several discrete activities, including defining a review protocol, identifying research, selecting studies, assessing study quality, extracting data, and synthesizing data. The review process is iterative, with many stages involving refinement as the review progresses. The report outlines the stages of planning, conducting, and reporting a systematic review, emphasizing the need for transparency and replicability.
The report discusses the importance of defining a research question, including its structure (population, intervention, outcomes), and the types of studies appropriate for answering the question. It also addresses the challenges of identifying and assessing study quality, particularly in software engineering, where empirical research is limited compared to medical research. The report highlights the need for quality assessment instruments and the limitations of such assessments.
The report outlines the process of generating a search strategy, including the use of electronic databases, reference lists, and expert consultation. It discusses the issue of publication bias and the need to address it through methods such as scanning grey literature and consulting experts. The report also addresses the challenges of managing bibliographies, retrieving documents, and documenting the search process.
The report discusses the selection of studies based on inclusion and exclusion criteria, the importance of reliability in inclusion decisions, and the need for sensitivity analysis. It emphasizes the importance of assessing study quality, using quality instruments, and the limitations of such assessments. The report also discusses the design of data extraction forms, the contents of data collection forms, and the procedures for data extraction.
The report outlines the process of data synthesis, including descriptive and quantitative synthesis. It discusses the importance of presenting quantitative results in a clear and consistent manner and the use of statistical techniques for quantitative synthesis. The report also addresses the challenges of synthesizing data from different studies and the need for sensitivity analysis to assess the robustness of results.This report proposes a guideline for systematic reviews in software engineering research. A systematic review is a method to evaluate and interpret all available research relevant to a specific research question, topic, or phenomenon. Systematic reviews aim to provide a fair evaluation of a research topic using a trustworthy, rigorous, and auditable methodology. The guideline is adapted from three existing medical guidelines to suit software engineering research. It covers three phases: planning, conducting, and reporting the review. It is at a high level and does not consider the impact of question type on review procedures or specify mechanisms for meta-analysis.
The report introduces the concept of systematic reviews as a significant research method for software engineering researchers. It discusses the importance of systematic reviews in ensuring thorough and unbiased literature reviews. Systematic reviews are more rigorous than traditional reviews and provide information about the effects of phenomena across various settings and methods. They can also help identify gaps in research and support the generation of new hypotheses.
Systematic reviews involve several discrete activities, including defining a review protocol, identifying research, selecting studies, assessing study quality, extracting data, and synthesizing data. The review process is iterative, with many stages involving refinement as the review progresses. The report outlines the stages of planning, conducting, and reporting a systematic review, emphasizing the need for transparency and replicability.
The report discusses the importance of defining a research question, including its structure (population, intervention, outcomes), and the types of studies appropriate for answering the question. It also addresses the challenges of identifying and assessing study quality, particularly in software engineering, where empirical research is limited compared to medical research. The report highlights the need for quality assessment instruments and the limitations of such assessments.
The report outlines the process of generating a search strategy, including the use of electronic databases, reference lists, and expert consultation. It discusses the issue of publication bias and the need to address it through methods such as scanning grey literature and consulting experts. The report also addresses the challenges of managing bibliographies, retrieving documents, and documenting the search process.
The report discusses the selection of studies based on inclusion and exclusion criteria, the importance of reliability in inclusion decisions, and the need for sensitivity analysis. It emphasizes the importance of assessing study quality, using quality instruments, and the limitations of such assessments. The report also discusses the design of data extraction forms, the contents of data collection forms, and the procedures for data extraction.
The report outlines the process of data synthesis, including descriptive and quantitative synthesis. It discusses the importance of presenting quantitative results in a clear and consistent manner and the use of statistical techniques for quantitative synthesis. The report also addresses the challenges of synthesizing data from different studies and the need for sensitivity analysis to assess the robustness of results.