Do I have the disease? (Solution)
Classic bayes-rule question. Instead of using the formula, let’s solve it using an example.
Let’s start with a population of 1000 people. 2% of people have the disease, so that’s 20 people.
Yes Disease | No Disease |
---|---|
20 | 980 |
If everyone took the test, 95% of the 20 people with the disease would be positive; that’s 19 people. 10% of the 980 people without the disease would be positive; that’s 98 people.
Yes Disease | No Disease | Total | |
---|---|---|---|
Test Positive | 19 | 98 | 117 |
Test Negative | 1 | 882 | 883 |
Total | 20 | 980 | 1000 |
You take the test and you’re positive. Given that information, we now only care about the “Test Positive” row. In that row, 19 out of a total of 117 people have the disease. So the chance you have it is \(19/117 \approx 0.16%\).