How pp_check "calculate" probability observation y when link function is bernoulli

[xf15]

sorry that this is probably a dumb question

i understand that when link function is identity or lognormal, observation is just observation, like this pic from
Gabry, J. , Simpson, D. , Vehtari, A. , Betancourt, M. and Gelman, A. (2019), Visualization in Bayesian workflow. J. R. Stat. Soc. A , 182: 389-402. doi:10.1111/rssa.12378

but i don’t get how pp_check “calculate” probability observation y when link function is bernoulli
like this pic from my doing prior predictive check on my own data

if someone simply point me to things i should read that would be helpful, thanks

[franzsf]

Different visualizations are more or less helpful for different models. Try ‘pp_check(yourModel, “xyz”)’ to get a list of the options outside of the default density overlay. For bernoulli, “stat” and its grouped variant work OK at seeing how well the model estimates probabilities.