Statistics 101-106, Fall 1998

Pollard homework sheet 11

Due: Thursday 3 December

Lecture 11 used information extracted from the April 1990 Consumer Reports. I borrowed the data set from the Splus statistical analysis program (the program that I use routinely).

• The full data set, and the subset described in the lecture.
• A slightly modified version of the Splus help page for the car data, giving descriptions of the variables.

You may work with the data for only the four countries discussed in class, if you wish.

1. Regress logPrice on the two predictors logHP and German (the indicator for cars from Germany). Repeat the exercise, with each of the other country indicators replacing the indicator for Germany in turn.
2. Why does Minitab not complain about redundant variables in part (i)?
3. Interpret the output from each of the four regressions from part (i).
4. Are there any other variables in the data set that might offer a better explanation for the apparent Country effect? (You might try regressions with Type or Reliability included amongst the predictors.)
5. Are there any variables in the data set that capture the idea of a luxury car, or a car that has a good reason for being expensive? (I fear I know too little about cars to give a good answer to this question.)
6. Based on what you can determine from the data, answer the question: Are German cars really more expensive than comparable cars from other countries? How confident are you of your answer? Would your conclusion apply to a comparison between all cars on the market, or are there any caveats that need to be attached?