Suppose that you have a set of data that produces a correlation of r = -0.67 between graduate GPA and the number of errors made in a psychological report. You’re curious what this correlation looks like if you just looked at the data for people who scored in the “A” range on GPA. What is likely to happen to the correlation when you look at just these individuals?
Its magnitude will decrease
Its sign will change
Its magnitude will stay the same
Its magnitude will increase