metagroup v1.0.0: First Release

metagroup v1.0.0: First Release

This is the first stable release of metagroup, an R package for creating and exploring meaningful subgroups in meta-analysis.

The package provides a suite of tools that follow a two-step process:

1. Group: Partition studies into statistically homogeneous clusters based on their effect size data.
2. Interpret: Analyze the composition of these newly formed groups to provide a deeper, data-driven interpretation of heterogeneity.

Included Functions

This initial release includes 10 primary functions to handle the full workflow, from grouping to interpretation.

Grouping Functions

A set of functions to perform iterative homogeneity-based grouping for various data types:

• mgbin(): For binary outcome data (e.g., odds ratios, risk ratios).
• mgcont(): For continuous outcome data (e.g., mean differences).
• mgcor(): For correlational data.
• mgprop(): For single proportions.
• mginc(): For two-group incidence rate data (e.g., incidence rate ratios).
• mgrate(): For single incidence rates.
• mgmean(): For single means.
• mggen(): For generic inverse-variance data (pre-calculated effect sizes and standard errors).

Interpretation and Visualization

Functions to analyze and visualize the newly created subgroups:

• meaning(): The core interpretation function. It analyzes the composition of subgroups based on study-level characteristics (e.g., country, setting) to help explain what makes them distinct.
• plot() and print(): S3 methods for the meaning() output, providing a clean summary table and a faceted bar chart to visualize the subgroup compositions.

Thank you for your interest in metagroup. Feedback and contributions are welcome!