A professor tells his class that he knows their second exam score without their having to take the test. He tells them that the second exam score can be predicted from the first with this equation:

Predicted second exam score = 5 + 0.75 (first exam score)

This tells us that the deterministic part of the regression model that predicts second exam score on the basis of first exam score is a straight line. What factors might contribute to the random component? In other words, why might a student’s score not fall exactly on this line?