Plot what the priors say the data should look like — plot_prior

The picture a prior check is read through. "density" draws each prior-predictive draw as its own thin line, so the spread between draws stays visible where a single pooled density would hide it; "histogram" is for a discrete response, where a density is misleading; "statistic" plots the summary table instead of the distribution.

Usage

plot_prior_check(
  x,
  type = c("density", "histogram", "statistic"),
  observed = TRUE,
  draws = 50,
  facet_by = "response",
  ...
)

Arguments

x: A bmmtools_prior_check object from prior_check().
type: "density", "histogram" or "statistic".
observed: Overlay the observed response from the data the check was run over. Dropped without comment when the model's response column is not in that data, since a prior check over a design skeleton has nothing observed to show. Ignored by "statistic".
draws: Prior-predictive draws to show per set, capped at what the object holds.
facet_by: A column of x to make panels from, or NULL. "response" by default, so a model with several observables does not share an axis.
...: Not used.

Value

A ggplot object.

Examples

plot_prior_check(prior_check_sdt_yn, type = "histogram")

plot_prior_check(prior_check_sdt_yn, type = "statistic")