babebi: An R Package for Bayesian Estimation and Validation in Small-N Two-Rater Pre-Post Designs

Area of Science:

• Statistics
• Psychometrics
• Behavioral Science

Background:

• Pre-post rating designs are common in various fields.
• Analyzing data from two-time, two-rater designs presents unique statistical challenges.
• Existing methods may not fully account for rater variability in pre-post analyses.

Purpose of the Study:

• To introduce babebi, an R package for analyzing complete two-time, two-rater pre-post rating designs.
• To provide accurate estimation of pre-post effects and adjusted change.
• To offer tools for evaluating the inferential performance of statistical methods in these designs.

Main Methods:

• Utilizes a linear model incorporating a rater indicator as a covariate.
• Employs Bayesian approximations for posterior summaries and BIC-based Bayes factors.
• Includes Monte Carlo validation routines calibrated to the observed data design.

Main Results:

• The package estimates pre-post effects and provides adjusted estimates of change.
• Offers posterior summaries and BIC-based Bayes factor approximations for model comparison.
• Monte Carlo validation routines assess inferential performance under specific study conditions.

Conclusions:

• babebi offers a comprehensive solution for analyzing complex pre-post rating data.
• The package facilitates robust estimation of change and model evaluation.
• It aids researchers in understanding and improving the reliability of their measurement designs.