Rdecom: R-squared based on ANOVA decomposition in multilevel model

In PeterVerboon/cyclicpkg: Analysis of (ESM/EMA) models with cyclic terms

Description

Function that computes the R-squared measure based on ANOVA decomposition measure (Edoardo Costantini, 2018). The explained variance is split into the explained variance for level-1 (within subjects), for the level-2 (between subjects) and the total explained variance.

Usage

Rdecom(dat, y, grpid, cov, xvar1, xvar2 = NULL, fit)

Arguments

dat	data set
y	dependent variable of the multilevel model
grpid	group identity (cluster variable)
cov	covariates in the MLA model
xvar1	name of first time variable from the cyclic process
xvar2	name of optional second time variable from the cyclic process
fit	lmer or lmerTest object (fit multilevel model)

Value

vector containing three values, named: "level_1_R2","level_2_R2", "total_R2"

References

Constantini, E. (2018). R-squared measures in Multilevel Modeling: The undesirable property of negative R-squared values. Downloaded from http://arno.uvt.nl/show.cgi?fid=146739, August 21, 2019.

PeterVerboon/cyclicpkg documentation built on March 4, 2020, 1:27 p.m.