Mystified, you look at a scatterplot of the data and see a pattern in the results indicating a clear relation. They look like an upside-down U, as stress goes up, so does performance until a point where performance goes back down as stress continues to increase. What limitations of r, the Pearson correlation, does this example illustrate