Skip to main content
Figure 1 | BMC Medical Education

Figure 1

From: The standard error of measurement is a more appropriate measure of quality for postgraduate medical assessments than is reliability: an analysis of MRCP(UK) examinations

Figure 1

In a Monte Carlo analysis, a simulated group of 10,000 candidates take an examination with a true mean of 50%, a true SD of 10%, a true reliability of 0.9, and a pass mark of 60%. Figure 1a shows the candidates' marks on the first attempt (horizontal axis), with the pass mark shown as the vertical dashed grey line, the failing candidates shown in red and the passing candidates shown in black. All of the simulated candidates then take the examination again, and their marks on that second occasion are shown on the vertical axis, with the horizontal dashed line showing the same pass mark as was used on the first occasion. Figure 1b is restricted to the 1565 candidates who passed the examination on the first assessment, and shows the marks they obtained when they took the examination for the second time (horizontal axis), and then again on taking it for a third time (vertical axis). Once again the notional pass mark of 60% is indicated by the vertical and horizontal grey dashed lines.

Back to article page