Risk is a challenging concept to interpret. This is especially true for rare events, where our intuition often leads us astray. Unfortunately, people must often make critical decisions based on these poorly understood statistics. This interactive visualization aims to help communicate how basic statistics such as the false positive rate, false negative rates, and disease incidence rate interact to determine how best to interpret the results from predictive tests.
Suppose that you tested positive for a rare disease. Does that mean you are truly sick? It depends! How often does the test flag a healthy person as sick (the false positive rate)? How often does the test flag a sick person as healthy (the false negative rate)? How rare is the disease (the disease prevalence)?
The visualization includes 1,000 squares, each representing a single patient within a group of 1,000 people. The red squares represent patients who are sick and tested positive for their disease. The white squares are those who are healthy and tested negative. In combination, therefore, the red and white squares represent those for whom the test was accurate.
In contrast, the orange squares represent sick patients that the test missed. These are the False Negatives. The blue squares represent health patients who incorrectly tested positive for the disease. These are the False Positives. As you can see, with the default settings below, a majority of those testing positive for a disease are actually healthy even though the test has a low error rate!
Edit the values below to experiment on your own. Simply change one or more of the rates and press <ENTER> to update the visualization.