Assignment Seven, Quiz on Friday Oct. 31st

Bring a printout of your unedited list file to the quiz. Don't write on the printout, except for your name and student number. Remember, using any part of someone else's command file constitutes academic dishonesty.

First read pages 12 through 16 in Part 2 of the class notes. This material will also be covered in lecture.

For this assignment you will be using the SAT data of assignment 2. Your command file should start by reading grades.dat and providing variable labels. Then use the regression procedure (see p. 15 of the notes) to predict grade point average from verbal SAT and math SAT (two independent variables). Be ready to answer questions like the following (see the table on page 16).

1. Considered simultaneously, are math SAT and verbal SAT significantly related to first-year GPA?
2. If verbal SAT is held constant, is math SAT significantly related to first-year GPA? Is the relationship (regardless of whether it is significant) positive or negative?
3. Controlling for math SAT, is verbal SAT significantly related to first-year GPA? Is the relationship (regardless of whether it is significant) positive or negative?
4. What proportion of the variation in first-year GPA is explained by verbal and math SAT together?

There are other interesting questions that are not part of this assignment, and will not be on the quiz (is that clear enough?). For example, how does the answer to the last question compare to the squared correlation between total SAT score and GPA? And what would happen if you were to do a multiple regression including total SAT as well as verbal & math (3 IVs)? If you want to try this as an experiment (do not bring the experimental printout to the quiz), do it 2 ways in 2 separate regression commands. On the variables= list, try putting total score first, and then try putting it last. How does this affect the output? Can you guess what is happening?