mlmhelpr
Multilevel/Mixed Model Helper Functions
A collection of miscellaneous helper function for running multilevel/mixed models in 'lme4'. This package aims to provide functions to compute common tasks when estimating multilevel models such as computing the intraclass correlation and design effect, centering variables, estimating the proportion of variance explained at each level, pseudo-R squared, random intercept and slope reliabilities, tests for homogeneity of variance at level-1, and cluster robust and bootstrap standard errors. The tests and statistics reported in the package are from Raudenbush & Bryk (2002, ISBN:9780761919049), Hox et al. (2018, ISBN:9781138121362), and Snijders & Bosker (2012, ISBN:9781849202015).
mlmhelpr
A package of helper functions for multilevel models fit using the lme4
package
Authors
Installation
To install the latest development version directly from Github, type:
install.packages(“remotes”)
remotes::install_github(“lrocconi/mlmhelpr”)
Functions
boot_se
Compute bootstrap standard errors and confidence intervals for fixed effects
center
Automatically grand- or group-mean center variables and re-estimates the model
design_effect
Calculate the design effect to determine if multilevel modeling is needed
hausman
Perform a Hausman test to test for differences between random- and fixed-effects models (experimental)
icc
Calculate the intraclass correlation
ncv_tests
Computes three different Non-constant variance tests (experimental)
plausible_values
Compute the plausible value range for random effects
r2_cor
Calculate the squared correlation between the observed and predicted values
r2_pve
Compute the proportion of variance explained for each random effect in the model
reliability
Calculate reliability coefficients for random effects
robust_se
Computes robust standard errors for lmer models
taucov
Calculate correlation between random intercepts and slopes