I think you are right that the interpretation is reversed if the location of the computed data averaged prior is not correct. For over and under dispersion I think it should not matter. Visually with the same data:
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Draws and simulations in the SBC example in the Stan's user guide | 4 | 590 | October 27, 2020 | |
| Comparing theory and stan implementation on sbc | 7 | 956 | October 29, 2021 | |
| SBC with histogram/ecdf vs treating credible intervals as confidence intervals | 6 | 1168 | March 9, 2022 | |
| Extending Rstan's Simulation Based Calibration | 4 | 653 | July 25, 2024 | |
| Simulation-based calibration case study with RStan | 31 | 3054 | July 30, 2019 |