Ancillary results from high-level validation

Posterior medians and credible intervals for R and k from the simulation-based calibration which produced sbc_quants.

data(sbc_quants)

Format

data.frame where each row summarizes posterior estimates and MCMC quality for an analysis of a simulated dataset. All 1,000 analyses are included, regardless of convergence.

r_true

The true simulating value of R.

k_true

The true simulating value of k.

r_point

The posterior median value of R.

k_point

The posterior median value of k.

r_low

The posterior 2.5th percentile of R.

r_high

The posterior 97.5th percentile of R.

k_low

The posterior 2.5th percentile of k.

k_high

The posterior 97.5th percentile of k.

min_ess

The minimum (across model parameters) n_eff value in the MCMC run.

max_rhat

The maximum (across model parameters) Rhat value in the MCMC run.

num_low_bfmi

Result of rstan::get_low_bfmi_chains() on the MCMC run.

num_divergent

Result of rstan::num_divergent() on the MCMC run.

num_max_treedepth

Result of rstan::num_max_treedepth() on the MCMC run.

Source

Generated in R using nbbp.

Examples

if (FALSE) { # \dontrun{
data(sbc_quants)
plot(sbc_ests$r_true, sbc_ests$r_point)
abline(a = 0, b = 1)
} # }