Diagnostic plots of a regression model

A teaching companion, not a decision tool. Every verdict these panels illustrate is already a row in the table's own footer, for every model column, with no plotting package installed (see the stats argument of tab_reg()). This function exists to show what a violation looks like.

One call diagnoses every model in the table: one titled grid per model, drawing the panels its own family allows. Pass a tab_reg() table — the data it was built from is usually found on its own — or a fitted model directly.

Usage

reg_check_plots(
  x,
  data = NULL,
  check = "auto",
  predictors = NULL,
  ncol = NULL,
  facet_ncol = NULL,
  theme = NULL,
  lang = NULL,
  max_points = 2000L,
  nbins = 10L,
  conf = 0.95,
  seed = 20260810,
  ...
)

Arguments

x

A tab_reg() table, or a fitted model (lm / glm / svyglm / polr / multinom / svyolr).

data

The data frame or survey::svydesign the table was built from. Usually unnecessary: a table records the name it was called with, and when that name still holds data of the same size, it is used — otherwise the call stops rather than draw the wrong model. Give data explicitly when the table was built from an expression rather than a named object (tab_reg(gss |> dplyr::filter(...), ...)), or when the name has since changed. Ignored with a bare model.

check

Which panels to draw. "auto" (default) draws the panels that decide something the footer cannot say in one number — linearity, residuals, normality, influence, and proportionality for an ordinal outcome. "all" adds dispersion and collinearity, whose footer row is normally enough. Or name them: any of "linearity", "residuals", "normality", "dispersion", "influence", "collinearity", "proportionality" — the same words the footer rows and tab_reg()'s stats argument use.

predictors

Optional: restrict the linearity panel to these continuous predictors.

ncol

Number of panel columns in the assembled grid (default: as square as it can be, 3 at most).

facet_ncol

Number of facet columns inside a panel (default: 2 for linearity, 4 for proportionality).

theme

"light", "dark", or a black-and-white publication palette ("print_ready" and friends). Defaults to options("tabxplor.theme"), like the table exporters.

lang

Language of the titles and captions ("en", "fr", ...). Defaults to options("tabxplor.lang").

max_points

Thin the raw-point layers to about this many observations; statistics and verdicts are always computed on the full data.

nbins

Bins of the linearity panel's observed curve (default 10).

conf

Confidence level of the Q-Q band. Default 0.95.

seed

Seed of the randomised quantile residuals (NULL for a fresh draw each time).

...

Unused, for future extension.

Value

Invisibly, the assembled gtable — or, with several models, the named list of them, one per model, all drawn on the current graphics device.

See also

tab_reg() and its stats argument (the same checks as footer rows), and forest_plot() for the RESULTS – its opposite contract: it reads the finished table and never re-fits, where a model check always must.

Examples

# \donttest: building a multi-panel ggplot grid costs a few seconds of CPU.
# \donttest{
d <- forcats::gss_cat |>
  dplyr::mutate(married = factor(dplyr::if_else(marital == "Married",
                                                "Married", "Not married")))
if (requireNamespace("ggplot2", quietly = TRUE) &&
    requireNamespace("gridExtra", quietly = TRUE)) {
  t <- tab_reg(d, "married", c("race", "age"), family = "binomial")
  reg_check_plots(t)
}

# }