model_performance.merMod: Performance of Mixed Models

View source: R/model_performance.mixed.R

model_performance.merModR Documentation

Performance of Mixed Models


Compute indices of model performance for mixed models.


## S3 method for class 'merMod'
  metrics = "all",
  estimator = "REML",
  verbose = TRUE,



A mixed effects model.


Can be "all", "common" or a character vector of metrics to be computed (some of c("AIC", "AICc", "BIC", "R2", "ICC", "RMSE", "SIGMA", "LOGLOSS", "SCORE")). "common" will compute AIC, BIC, R2, ICC and RMSE.


Only for linear models. Corresponds to the different estimators for the standard deviation of the errors. If estimator = "ML" (default), the scaling is done by n (the biased ML estimator), which is then equivalent to using AIC(logLik()). Setting it to "REML" will give the same results as AIC(logLik(..., REML = TRUE)).


Toggle warnings and messages.


Arguments passed to or from other methods.


Intraclass Correlation Coefficient (ICC)

This method returns the adjusted ICC only, as this is typically of interest when judging the variance attributed to the random effects part of the model (see also icc()).

REML versus ML estimator

The default behaviour of model_performance() when computing AIC or BIC of linear mixed model from package lme4 is the same as for AIC() or BIC() (i.e. estimator = "REML"). However, for model comparison using compare_performance() sets estimator = "ML" by default, because comparing information criteria based on REML fits is usually not valid (unless all models have the same fixed effects). Thus, make sure to set the correct estimator-value when looking at fit-indices or comparing model fits.

Other performance indices

Furthermore, see 'Details' in model_performance.lm() for more details on returned indices.


A data frame (with one row) and one column per "index" (see metrics).


model <- lme4::lmer(Petal.Length ~ Sepal.Length + (1 | Species), data = iris)

performance documentation built on Nov. 2, 2023, 5:48 p.m.