New SSC package: multispline -- nonlinear multilevel spline modeling

[Subir Hait]

Dear Statalist,

I would like to share a new SSC package, -multispline-, for nonlinear multilevel modeling using natural cubic splines. The command provides a unified workflow that integrates spline construction, mixed-effects estimation, ICC computation, prediction, and visualization into a single interface.

Install:
ssc install multispline

Example:
sysuse nlsw88, clear
multispline wage age, cluster(industry) nknots(4) plot

GitHub repository: https://github.com/causalfragility-l...tiSpline-Stata

Feedback is welcome.

Best regards,
Subir Hait
PhD Candidate (MQM), Michigan State University

[Subir Hait]

Dear all,

multispline v0.2.0 is now available from SSC. The command fits nonlinear relationships using spline functions in both single-level and multilevel (mixed-effects) models.

Main features include:

• automatic selection of spline degrees of freedom,
• support for mixed-effects models (Gaussian and binary outcomes),
• marginal and conditional R-squared for multilevel models,
• derivative-based interpretation (slopes and turning points),
• model comparison (linear, polynomial, spline),
• visualization of fitted curves and marginal effects.

Install from SSC:
ssc install multispline

Feedback and suggestions are very welcome.

Best regards,
Subir Hait
Michigan State University