test_dispersion: Test for over- or underdispersion

In tulpa: Templated Unified Library for Posterior Approximation in Bayesian Hierarchical Models

Test for over- or underdispersion

Description

Compares the variance of observed data to the variance expected under the fitted model (via simulation). Ratio > 1 = overdispersion; < 1 = underdispersion. Equivalent to DHARMa::testDispersion().

Usage

test_dispersion(object, ...)

## Default S3 method:
test_dispersion(
  object,
  observed = NULL,
  nsim = 250L,
  seed = 123L,
  alternative = c("two.sided", "greater", "less"),
  ...
)

Arguments

object	A fitted model with simulate() method
...	Passed to methods.
observed	Observed response vector (optional – extracted from fit)
nsim	Number of simulations (default 250)
seed	Random seed (default 123)
alternative	"two.sided", "greater", or "less"

Value

An htest object with dispersion ratio and p-value

Examples

set.seed(1)
d <- data.frame(x = rnorm(80))
d$y <- rpois(80, exp(0.3 + 0.5 * d$x))
fit <- tulpa(y ~ x, data = d, family = "poisson", mode = "laplace")
test_dispersion(fit, nsim = 50)

tulpa documentation built on Oct. 19, 2026, 5:09 p.m.