Skip to contents

Tidy a fitted gllvmTMB model

Source: R/methods-gllvmTMB.R
tidy.gllvmTMB_multi.Rd

Returns a tibble (or data.frame) of either the fixed-effect coefficient table, the random-effects variance / covariance terms, or the ordinal threshold cutpoints. Mirrors the tidy.sdmTMB() API but augmented for the additional covstructs and the gllvmTMB-native ordinal_probit() / ordinal_logit() families.

Usage

# S3 method for class 'gllvmTMB_multi'
tidy(
  x,
  effects = c("fixed", "ran_pars", "cutpoint"),
  conf.int = FALSE,
  conf.level = 0.95,
  ...
)

Arguments

x

A fit returned by gllvmTMB().

effects

One of "fixed" (default), "ran_pars", or "cutpoint". The "cutpoint" class returns the ordinal cutpoints (one row per (trait, threshold) pair); it is empty for fits with no ordinal_probit() / ordinal_logit() traits. (Earlier releases lumped the cutpoints into "ran_pars" as a categorisation hack — see NEWS.)

conf.int

Whether to add conf.low / conf.high columns.

conf.level

Confidence level for the CI.

...

Currently unused.

Value

A data.frame. effect = "fixed" rows include a link column reporting each trait's link function ("identity", "probit", "log", "logit", …). effect = "cutpoint" rows carry the ordinal-probit thresholds. For a non-unit weighted objective, fixed-effect point estimates remain available but std.error is NA, an inference_status column explains the boundary, and conf.int = TRUE fails because no sandwich interval is certified.