A precomputed prior_check() result, shipped so that summary() and
plot_prior_check() can be tried without compiling a model. Two
prior sets are compared for bmm::sdt_yn() with a random intercept on
d and criterion: bmm's defaults (default) and the defaults with
normal(0, 1) priors on both between-subject SDs (narrow_sd). The
data are 20 simulated subjects with 50 trials per stimulus class; 100
prior-predictive draws are kept per set, from 2 chains of 1000
iterations. Seed 2026.
Format
A bmmtools_prior_check tibble with the columns prior,
response, statistic and value, carrying the draws, the prior
summaries of both fits, the data and the model as attributes.
Examples
summary(prior_check_sdt_yn)
#> # A tibble: 5 × 5
#> response statistic default narrow_sd difference
#> <chr> <chr> <dbl> <dbl> <dbl>
#> 1 hits floor_rate 0.392 0.152 -0.241
#> 2 hits ceiling_rate 0.420 0.146 -0.274
#> 3 hits q50 30 23 -7
#> 4 hits q90 50 50 0
#> 5 hits q95 50 50 0