Figure 1.

Results from the synthetic Gaussian case study. (A) Posterior mean (solid lines) with 95% CI (dashed lines) for ${\sigma }_u^2$ (black and open squares) and ${\sigma }_e^2$ (gray and closed squares) from the Gaussian synthetic case study against the value of ${\sigma }_u^2$ used in the simulations (together with ${\sigma }_e^2=1-{\sigma }_u^2$). The power of a model selection test using ΔDIC > 10 as limit is plotted as x and solid lines. The power is estimated using the simulation approach described in the text. (B) Boxplots of simulated values of ΔDIC against the value of ${\sigma }_u^2$ used in the simulations (together with ${\sigma }_e^2=1-{\sigma }_u^2$). The values of ΔDIC from the synthetic case study are plotted as stars. ΔDIC equal to 10 is indicated by a horizontal line.