Skip to main content
. Author manuscript; available in PMC: 2019 Oct 1.
Published in final edited form as: Cancer Epidemiol. 2018 Aug 10;56:83–89. doi: 10.1016/j.canep.2018.07.014

Table 1:

Reference distribution for the sensitivity corresponding to 90% specificity estimated with the empirical ROC when calculated with data for 40 cases and 160 controlsa. The reference distribution is used to determine exact P-values and was generated by 40,000 randomly chosen enumerationsb of ranks for 200 subjects with the first 40 labelled as cases.

r Probability that the estimated sensitivity ≥ r
0.000 1.000000
0.025 0.976575
0.050 0.892075
0.075 0.739125
0.100 0.549825
0.125 0.367025
0.150 0.218200
0.175 0.119050
0.200 0.059175
0.225 0.027675
0.250 0.012025
0.275 0.004850
0.300 0.001750
0.325 0.000550
0.350 0.000225
0.375 0.000125
0.400 0.000025
a

Smallest increment for realized values of the estimated sensitivity is 0.025 = 1/40 where 40 is the number of cases.

b

Smallest increment for probability is 0.000025 = 1/40000 where 40000 is the number of random rank enumerations.