Figure 4.

The effect of P(D) on post-test probability functions obtained by the LR method. Post-test probability functions P(D|x) of diagnostic value x, computed at different P(D) of 0.10 (dash-dotted lines), 0.50 (solid lines) and 0.90 (dashed lines). The three curves at each pre-test probability are obtained from the Monte Carlo experiment shown also in Figure 3 (scenario 1, 200 fictitious data sets) using the LR method, employing data sets with P(D) of 0.10 (red), 0.50 (black) and 0.90 (blue). Note that all three curves with the same colour run “parallel” to each other; i.e., they are obtained using the same slope parameter [mean value of the 1000 estimates for α1 at the respective P(D)]. The little arrows denote the crossing regions of the three curves obtained with the three different LR functions (see text for explanation).