modelDiagnostics.merMod: modelDiagnostics method for merMod objects

In multilevelTools: Multilevel and Mixed Effects Model Diagnostics and Effect Sizes

Description

This function creates a number of diagnostics for merMod models.

Usage

## S3 method for class 'merMod'
modelDiagnostics(
  object,
  ev.perc = 0.001,
  robust = FALSE,
  distr = "normal",
  standardized = TRUE,
  ...
)

Arguments

object	A fitted model object, either of class merMod from the lme4 package or merModLmerTest from the lmerTest package.
ev.perc	A real number between 0 and 1 indicating the proportion of the theoretical distribution beyond which values are considered extreme values (possible outliers). Defaults to .001.
robust	Whether to use robust mean and standard deviation estimates for normal distribution
distr	A character string given the assumed distribution. Passed on to testDistribution. Defaults to “normal”.
standardized	A logical whether to use standardized residuals. Defaults to TRUE generally where possible but may depend on method.
...	Additional arguments, not currently used.

Value

A logical (is.modelDiagnostics) or a modelDiagnostics object (list) for as.modelDiagnostics and modelDiagnostics.

Examples

library(JWileymisc)
sleep[1,1] <- NA
m <- lme4::lmer(extra ~ group + (1 | ID), data = sleep)

md <- modelDiagnostics(m, ev.perc = .1)
md$extremeValues
class(md)

plot(md)

data(aces_daily, package = "JWileymisc")
m <- lme4::lmer(PosAff ~ STRESS + (1 + STRESS | UserID), data = aces_daily)
md <- modelDiagnostics(m, ev.perc = .1)

#  gm1 <- lme4::glmer(cbind(incidence, size - incidence) ~ period + (1 | herd),
#    data = lme4::cbpp, family = binomial)
# modelDiagnostics(gm1) ## should be an error

rm(m, md, sleep)

multilevelTools documentation built on March 13, 2020, 2:07 a.m.