How to: ROC curve

1. Enter ROC data

From the Welcome or New table dialog, choose the one-way tab, and then choose a scatter plot. If you are not ready to enter your own data, choose the sample ROC data.

Enter diagnostic test results for controls into column A and patients in column B. Since the two groups are not paired in any way, the order in which you enter the data in the rows is arbitrary. The two groups may have different numbers of subjects.

2. Create the ROC curve

From the data table, click on the toolbar, and then choose Receiver-operator characteristic curve from the list of one-way analyses.

In the ROC dialog, designate which columns have the control and patient results, and choose to see the results (sensitivity and 1-specificity) expressed as fractions or percentages. Don't forget to check the option to create a new graph.

Note that Prism doesn't ask whether an increased or decrease test value is abnormal. Instead, you tell Prism which column of data is for controls and which is for patients, and it figures out automatically whether the patients tend to have higher or lower test results.

3. View the graph