Page 257 - Demo

Appendices255APSISOnly one out of 457 observations had a k-value just over 0.7, indicating that this observation had relatively large effect on the posterior distribution. However, because it concerned only one observation, it is unlikely to heavily influence model stability.Chapter 4Posterior predictive checkPosterior predictive checks based on the posterior predictive distribution indicated good model fit for all three experiments (Figure 2).Figure 2 – Predictive posterior distribution of 50 simulated datasets based on the model (in grey) with the distribution of the original data plotted on top (in black) for (A) Experiment 1, (B) Experiment 2, and (C) Experiment 3.PSISFor the three main models, all k-values were < 0.5, indicating that no extremely influential observations were found.Tom Roth.indd 255 08-01-2024 10:42