There are 500 cases from a defined population that have been subjected to diagnostic Test A,

and Test B.

In Test A, 80 people test result was positive and 420 people test result was negative. Suppose

that 100 actually positive cases and 400 actually negative cases were ultimately found in the

population studied and that the diagnostic test to be evaluated yielded 60 true positive (TP)

decisions, 380 true negative (TN) decisions.

In Test B, 110 people test result was positive and 390 people test result was negative. Suppose

that 100 actually positive cases and 400 actually negative cases were ultimately found in the

population studied and that the diagnostic test to be evaluated yielded 75 TP decisions, 365 TN

decisions.

a) Summarize the data in a separate “decision matrix” for each test and calculate Accuracy for

Test A and Test B.

b) Compare the accuracy of Test A and B and tell which test is best and which test is worst.

c) If accuracy measure can’t help completely answer question b), please draw simple ROC

curves to determine the superior test.