Labeled coords and dims in hierarchical group setting with different sizes

Now checking the results and for example if I select one participant who I know only belongs to experiment group 1 e.g.

coords = {"experiment": [2],"participants": [1]};
az.plot_posterior(trace, coords=coords)

there should be no posteriors because participant 1 has no observations for experiment 2, but strangely there are? So I assume it is not working correctly…?

Similarly:

coords = {"participants": [1]};
az.plot_posterior(trace, coords=coords);

outputs two parameters for each experiment’s mu_a, mu_i, mu_l and their sigmas. Why is that when participant 1 did not even have values for experiment 2?