Question: Perch and bass. Example 2.8 (page 91) gives data from an experiment in ecology. Figure 2.4 is the scatterplot of proportion of perch eaten by bass against the number of perch in a pen before the bass were let in. There is a roughly linear pattern. The least-squares line for predicting proportion eaten from initial count of perch is

(a) When 10 more perch are added to a pen, what happens to the proportion that are eaten (according to the line)? Explain your answer.

(b) If there are no perch in a pen, what proportion does the line predict will be eaten? Explain why this prediction is nonsense. What is wrong with using the regression line to predict y when x = 0? You see that the intercept, though it is needed to draw the line, may have no statistical interpretation if x = 0 is outside the range of the data.