The treatment of incomplete data: Reporting, analysis, reproducibility, and replicability

Area of Science:

• Statistics
• Data Science
• Research Methodology

Background:

• Incomplete data analysis is a persistent challenge across disciplines.
• Existing methods like complete case analysis and multiple imputation require careful application.
• The work by Nguyen et al. (2018) highlights the need for better practices in handling missing data.

Purpose of the Study:

• To discuss common misconceptions regarding incomplete data analysis, specifically multiple imputation.
• To emphasize the link between proper missing data handling and research reproducibility.
• To propose actionable recommendations for improving publication standards related to data analysis.

Main Methods:

• Commentary and discussion of existing literature and practices.
• Analysis of common misunderstandings surrounding multiple imputation.
• Development of a checklist for researchers, editors, and reviewers.

Main Results:

• Identified key aspects and misconceptions in incomplete data analysis.
• Positioned missing data as a component of broader reproducibility challenges.
• Proposed a checklist to elevate publication standards.

Conclusions:

• Improving the analysis and reporting of incomplete data is essential for research reproducibility.
• The scientific community needs to adopt more rigorous standards for handling missing data.
• Recommendations are provided to facilitate better practices and enhance scientific rigor.