What does it mean for data on a scatter plot to have no association? How does that affect the line of best fit?