Skip to main content
Implementation Note: This module uses Henderson Method III (moment estimator on OLS residuals + GLS via Sherman-Morrison), not iterative REML/ML as in lme4. Random slopes and crossed/nested effects are not supported. For those, use R lme4.
Home
› Mixed Models
Data & Demo
Load your data (CSV) or use built-in demo data
IMPORT DATA
ICC Calculator
Intraclass Correlation Coefficient — assess rater agreement or consistency
SETTINGS
Linear Mixed Model (LMM)
Y ~ X + (1 | Group) — choose REML (default, matches lme4) or Henderson III (fast)
⚠ Single random intercept only (no random slopes, no crossed/nested effects). REML/ML: exact profile likelihood via golden-section search (Bates et al. 2015). Henderson III: fast closed-form moment estimator. For random slopes or complex structures, use the WebR R engine below.
MODEL SPECIFICATION

ℹ️ Supports random-intercept models only. For random slopes, crossed random effects, or nested designs, use the WebR Verify button to run lme4::lmer() directly.

Generalized Linear Mixed Model (GLMM)
Y ~ X + (1 | Group) for binary (logistic) or count (Poisson) outcomes — Laplace approximation
⚠ GLMM with Laplace — single random intercept or cross-classified (1|A) + (1|B). Binomial (logit) / Poisson (log). Matches lme4::glmer within ~1%.
MODEL SPECIFICATION
Select to fit Y ~ X + (1|A) + (1|B) — for non-nested factors (e.g. subjects × items).
Linear Growth Curve
Plot individual trajectories and group means over time
SETTINGS
Advanced R Engine
Run full lme4::lmer and nlme in the browser via WebR
WEBR STATUS

Full mixed models (ML/REML) require an R engine. The JavaScript approximations are limited to random-intercept OLS. To run lme4::lmer or nlme::lme, you can load WebR directly in the browser.

ⓘ WebR downloads ~20 MB on first load. Ensure a stable internet connection.