Plot Posterior Predictive Check

Compares posterior predictive replicated statistics with fitted expected proportions or scores computed from posterior mean parameters. DPMirt does not currently store the original response matrix in dpmirt_fit, so this is a fitted-proportion check rather than a direct observed-data PPC. Requires ggplot2.

Usage

dpmirt_plot_pp_check(
  fit,
  stat = c("prop_correct", "total_score"),
  n_rep = 50,
  ...
)

Arguments

fit

A dpmirt_fit object from dpmirt.

stat

Character. Summary statistic for comparison: "prop_correct" (default) or "total_score".

n_rep

Integer. Number of replicated datasets. Default 50.

...

Currently unused.

Value

A ggplot object.

See also

plot.dpmirt_fit, dpmirt_diagnostics

Other visualization: dpmirt_plot_caterpillar(), dpmirt_plot_clusters(), dpmirt_plot_density(), dpmirt_plot_density_compare(), dpmirt_plot_dp_density(), dpmirt_plot_icc(), dpmirt_plot_info(), dpmirt_plot_items(), dpmirt_plot_parameter_trace(), dpmirt_plot_trace(), dpmirt_plot_wright_map(), plot.dpmirt_estimates(), plot.dpmirt_fit(), plot.dpmirt_sim()

Examples

if (FALSE) { # \dontrun{
sim <- dpmirt_simulate(200, 20, model = "rasch", seed = 42)
fit <- dpmirt(sim$response, model = "rasch", prior = "normal",
              niter = 5000, nburnin = 1000, seed = 123)
dpmirt_plot_pp_check(fit)
} # }