Fit a Linear Mixed-Effects Model With Safe Fallback

Fits a linear mixed model via [lme4::lmer()], catching non-convergence, singular fit warnings, or missing package conditions and gracefully falling back to a fixed-effects linear model ([stats::lm()]).

Usage

fit_lmer_safely(formula, data, warn_on_fallback = TRUE, ...)

Arguments

formula: Model formula, typically containing random effects (e.g. `y ~ x + (1 | id)`).
data: Data frame containing variables in `formula`.
warn_on_fallback: Logical. Emit a warning when falling back to `lm`. Default `TRUE`.
...: Additional arguments passed to [lme4::lmer()] or [stats::lm()].

Value

A list of class `"cbe_lmer_safely"` containing:

model: The fitted `merMod` or `lm` object
method: Character indicating `"lmer"` or `"lm"`
is_singular: Logical indicating whether the mixed model had a singular fit
formula: The formula used for the final fit

Details

Replaces inline tryCatch / fallback copies across longitudinal reports.

Examples

df <- data.frame(
  id = rep(1:5, each = 4),
  time = rep(1:4, times = 5),
  y = rnorm(20)
)
fit <- fit_lmer_safely(y ~ time + (1 | id), data = df)
#> boundary (singular) fit: see help('isSingular')
#> Warning: `lme4::lmer()` produced a singular fit; falling back to `stats::lm()`.
fit$method
#> [1] "lm"