Plot fixed effects across optimisers

Visualises how the fixed-effect estimates of a mixed model vary across the optimisers tried by lme4::allFit(). It offers a quick check that a model has settled on a stable solution: tight clusters of points indicate agreement between optimisers, whereas scatter signals a fragile fit.

Usage

optimizer_fixef_plot(
  x,
  intercept = TRUE,
  select_terms = NULL,
  interaction = c("times", "asterisk", "colon", "space"),
  labels = NULL,
  number_optimizers = TRUE,
  free_y = TRUE,
  ncol = NULL,
  point_size = 2,
  palette = NULL,
  y_lab = "Fixed effect",
  title = NULL
)

Arguments

x

Either the object returned by lme4::allFit(), or a data frame with one row per optimiser-by-term combination (columns such as optimizer, term and value/estimate).

intercept

Whether to keep the intercept panel. Defaults to TRUE.

select_terms

Optional character vector of terms to display (the intercept is always kept when intercept = TRUE).

interaction

Passed to format_terms() for the panel titles.

labels

Optional named character vector renaming the term panels, e.g. c(conditionunrelated = "Unrelated priming"). When x is a raw lme4::allFit() object, names are prettified to the effect (variable) name automatically (e.g. conditionunrelated to condition) and these labels override that default.

number_optimizers

Whether to prefix each optimiser name with a number, so that the legend doubles as an index.

free_y

Whether to give each panel its own y-axis range. This is advisable, because the intercept and the slopes usually occupy very different scales.

ncol

Number of facet columns. If NULL, chosen automatically.

point_size

Point size.

palette

Colours for the optimisers; defaults to depictr_palette().

y_lab, title

Axis label and plot title.

Value

A ggplot2::ggplot object.

Details

Each fixed effect occupies its own panel, with its own y-axis. The function also accepts a plain data frame, so it can be used without 'lme4'.

Examples

# Without lme4, build the input data frame directly:
set.seed(1)
df <- expand.grid(
  optimizer = c("bobyqa", "Nelder_Mead", "nlminbwrap"),
  term = c("(Intercept)", "rainfall", "fertiliser")
)
df$value <- c(5, 5.01, 4.99, 0.3, 0.31, 0.29, -0.2, -0.18, -0.21)
optimizer_fixef_plot(df)


# \donttest{
if (requireNamespace("lme4", quietly = TRUE)) {
  m <- lme4::lmer(life_satisfaction ~ stress + (1 | region),
                  data = wellbeing_survey)
  af <- lme4::allFit(m)
  optimizer_fixef_plot(af)
}
#> bobyqa : [OK]
#> Nelder_Mead : [OK]
#> nlminbwrap : [OK]
#> nloptwrap.NLOPT_LN_NELDERMEAD : [OK]
#> nloptwrap.NLOPT_LN_BOBYQA : [OK]

# }