A graphical organization of the results is produced by plotting the proportion of hits (on the y-axis) versus the proportion of false alarms (on the x-axis) for each stimulus intensity. The result is a set of receiver operating characteristic curves (ROC curves).

The diagonal line (brown) is a reference level for random responses. Usually subjects perform at better than this random level. Curves in the space below the diagonal would show performance that is systematically worse than random. (This probably means an error in the experimental design, or that the subject did not understand the instructions.) Each curve is created from the hit and false alarm rates for a given sensitivity (d'), such as size of target. This example has three sensitivity levels; higher sensitivities are bowed farther from the diagonal (blue), toward perfect performance. As perfection is usually impossible, the graph also shows the possible trade-off between missing signals and generating false alarms (response bias, in green). Shifting right along the curve shows what increase in proportion of false alarms to expect for a given increase in hit rate.