Wald tests for fixed coefficients — wald_test • gamlss.longitudinal

wald_test() provides a small reporting-friendly hypothesis-test surface for fitted gamlss.longitudinal models. It uses the same variance-covariance route as summary.gamlss.longitudinal() and confint.gamlss.longitudinal(), so numerical-Hessian tests should be reported as approximate.

Usage

wald_test(
  object,
  terms = NULL,
  L = NULL,
  rhs = 0,
  joint = FALSE,
  method = "analytical",
  ...
)

Arguments

object: A fitted gamlss.longitudinal object.
terms: Optional coefficient names, formula-term names such as "mu.treatment", coefficient-name prefixes, or numeric indices. When L is NULL, these select coefficients for individual tests or a joint test.
L: Optional contrast matrix. Columns must either be named with coefficient names or have one column per fixed coefficient in model order.
rhs: Null-hypothesis value. Either a scalar or one value per tested row.
joint: Logical; when TRUE, test selected terms jointly. Contrast matrices supplied through L are always tested jointly.
method: Variance-covariance method passed to vcov.gamlss.longitudinal().
...: Additional arguments passed to vcov.gamlss.longitudinal().

Value

An object of class gamlss_longitudinal_wald_test.