Figure 2:

The logarithm of the approximated Bayes factors based on the unit information prior $\text{log\hspace{0.17em}}\left({\widehat{B}}^{UI}\right)$ (dotted line), and the local unit information prior $\text{log\hspace{0.17em}}\left({\widehat{B}}^L\right)$ (solid line) of M₁ : θ₂ > θ₁ > 0 versus $M_u:\left({\theta }_1,{\theta }_2\right)\in {ℝ}^2$ (left panel), of M₁ versus the complement model M₂ (middle panel), and of M₁ versus M₀ : θ₂ = θ₁ = 0 (right panel). The criteria are plotted for n = 20 as a function of a, where $\left({\widehat{\theta }}_1,{\widehat{\theta }}_2\right)=\left(a,2a\right)$.