Our previous article covered the cases where you shouldn't use the arithmetic mean. This time we introduce the median. The median is certainly a correct and useful calculation, but not in every case. To understand the median, let's look at what happened in 1982 to Stephen Jay Gould, a Harvard University professor and biologist. When Gould learned he had cancer and talked to his doctor, the doctor told him nothing more than that it was "a rare disease." Gould then looked into it himself and found that people with this cancer lived 8 months on average. Reading this, he was shocked and thought, "So that's why they wouldn't tell me anything more." But that average was a median. In other words, 50% of the people with this disease lived less than 8 months, and the other 50% lived longer than 8 months. With this cancer, though, people who had beaten it once the 8 months had passed tended to live for decades. The median works well for data with outliers, but it is a poor fit for data with a spread this wide.