For Exercises 15–24, use the data and the regression equations that you calculated in the Section 13.1 exercises.

Question 13.99

21. World Temperatures. Refer to the data from Exercise 17.

  1. Construct a 90% prediction interval for the high temperature for a randomly chosen city with a low temperature of 30 degrees.

    743

  2. Suppose we were asked to provide a prediction interval for the high temperature for a randomly chosen city with a low temperature of zero degrees. Explain what the danger would be in constructing such a prediction interval.

13.2.21

(a) Lower bound: 22.1568 degrees (Minitab: 22.1564); Upper bound: 64.5317 degrees (Minitab: 64.5322) (b) The range of the low temperatures in the data set is from 7°C to 70°C, inclusive. Therefore, a low temperature of 0°C is outside of the range of our data set, and any predictions or estimates using the regression equation for 0°C represent extrapolation. Extrapolation should be avoided if possible because the relationship between the variables may no longer be linear outside the range of .