A new approach to modeling positive random variables with repeated measures

Area of Science:

• Statistics
• Biostatistics
• Econometrics

Background:

• Repeated measures experiments are common, requiring models that account for intra-unit dependency.
• Existing models for positive continuous data with repeated measures have limitations.

Purpose of the Study:

• To propose a novel generalized beta prime regression model for positive continuous data with repeated measures.
• To incorporate the modeling of intra-unit dependency in such experiments.
• To provide practical tools for applying the proposed methodology.

Main Methods:

• Generalization of the beta prime regression model to accommodate repeated measures.
• Development of residuals and diagnostic tools for model assessment.
• Monte Carlo simulation study to evaluate estimator performance under various correlation structures and distributions.

Main Results:

• The proposed generalized beta prime regression model effectively handles intra-unit dependency in repeated measures data.
• Simulation results demonstrate the good finite-sample performance of the proposed estimators.
• The methodology is validated through application to a real-world dataset.

Conclusions:

• The generalized beta prime regression model offers a flexible and effective approach for analyzing positive continuous repeated measures data.
• The developed methodology and accompanying R package provide valuable tools for researchers in various fields.
• Accurate modeling of intra-unit dependency is crucial for reliable inference in repeated measures studies.