Skip to contents

Assumption Checks for Residuals

Source: R/residuals.R
residuals.Rd

res_norm(), res_fit(), and res_box() provide diagnostic plots to check model assumptions at the within-group level for linear mixed-effects models fitted with lme4::lmer(). These checks are applicable to all valid lmer model structures (no shorthand syntax), including complex nested and crossed random structure.

Usage

res_norm(model)

res_fit(model)

res_box(model, group_var)

Arguments

model

A fitted model object from lme4::lmer(). Currently not supporting formula that contains shorthand ., such as y ~ ..

group_var

A character value specifying the name of the grouping variable. Must be a valid random intercept in model and a valid variable in the data used.

Value

A ggplot object.

Checking Assumptions

  • res_norm() generates a quantile-quantile (QQ) plot of the Pearson residuals to assess normality within groups.

  • res_fit() plots Pearson residuals against fitted values to detect funnel shapes or mean-variance trends that suggest transformation issues.

  • res_box() creates a boxplot of residuals grouped by a specified random intercept grouping variable to visually inspect heteroscedasticity.

These plots help detect violations of within-group assumptions in mixed-effect models (normality, homoscedasticity, independence), which are central to valid inference in hierarchical models. Between-group assumption checks are not included here, due to the various natures of complex random structures.

Examples

library(lme4)
library(ggplot2)

model <- lmer(math ~ math_old + cltype + (cltype | school_id), data = star)

res_norm(model)

res_fit(model)

res_box(model, "school_id")