Dropout analysis: A method for data from Internet-based research and dropR, an R-based web app and package to analyze...

Area of Science:

• Psychology
• Computer Science
• Statistics

Background:

• Internet-based research presents unique challenges with participant non-response, including item non-response and dropout.
• The voluntary nature and large sample sizes in online studies necessitate robust methods for analyzing dropout.
• Dropout analysis is crucial for understanding data integrity and research validity in digital environments.

Purpose of the Study:

• To develop and discuss methodologies for analyzing dropout in Internet-based research.
• To introduce dropR, an R package and web service for analyzing and visualizing dropout data.
• To provide researchers with accessible tools for understanding and presenting participant attrition.

Main Methods:

• Development of the dropR R package and Shiny web application.
• Utilizing R for statistical computing and graphical representation of dropout curves.
• Implementation of statistical tests including chi-square, odds ratios, Kaplan-Meier survival analysis, Kolmogorov-Smirnov, and rho-family statistics.

Main Results:

• dropR generates accessible, publication-ready visualizations of dropout curves.
• The package calculates key dropout parameters like chi-square values and odds ratios.
• Automated inferential components identify critical dropout points and differences between experimental conditions.

Conclusions:

• dropR offers a user-friendly solution for analyzing and visualizing dropout in internet-based research, requiring no programming knowledge.
• The tool enhances data interpretation by providing statistical insights and visual displays.
• Freely available as a web application and R package, dropR supports researchers in addressing participant attrition.