Marginaleffects: Predictions and Comparisons for Over 100 Statistical Models

Marginaleffects is an R package for computing predictions, comparisons, slopes, and marginal means from statistical models. It supports over 100 classes of models and calculates uncertainty via the delta method, bootstrapping, or simulation. The package is authored by Vincent Arel‐Bundock and detailed in a 2024 Journal of Statistical Software article.

Use Cases

• Conduct linear and non-linear hypothesis tests based on model outputs.
• Calculate and compare marginal effects or slopes for model interpretation.
• Generate predictions and contrasts (risk ratios, odds) for statistical reporting.
• Perform equivalence tests between model estimates.

Strengths

• Supports over 100 classes of statistical and machine learning models.
• Implements multiple inference methods: delta method, bootstrapping, and simulation.
• Peer-reviewed methodology documented in a 2024 Journal of Statistical Software article.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• License is listed as closed, which may restrict use.

Provenance

Source

Vincent Arel‐Bundock

Collection Method

Software package for statistical computation.

Time Range

Publication details reference a 2024 article.

Freshness

Last update date is unknown; freshness unverified.

Geography

Spatial coverage is unknown.