PSY340: Statistics for the Social Sciences

These are problems from Labs 6 and 7. Re-interpret your answers (or do it them again for added practice) but use the information from the various hypothesis tests (make sure to state your Null and alternative hypotheses).

Part I:

1) Suppose that you (and your study group) want to predict how well you will do on the final exam based on your midterm exam score. Your professor provides you with some data from past semesters and you determine that the regression equation is:
predicted final exam = 40 + .5(midterm).
For each of the following students in your study group predict their final exam grade.
Brad 30, Jennifer 40, Angelina 60, Billy Bob 70, and Johnny Lee 90

2) Consider the following dataset:

Hours studied	Test Grade
0	52
10	95
6	83
8	71
6	64

Using this data (a) make a scatterplot, (b) compute the least squares regression line, (c) compute the SS_error

1) a) Compute the regression equation for college gpa (criterion variable) and math SAT (predictor). What is the slope? What is the intercept? Write out the regression equation.
b) How much of the variance in college gpa is accounted for math SAT?
c) Predict John Smith's college gpa based on his math SAT score of 520.

2) Do the same (parts a & b) as in 1 but with verbal SAT.
c) Predict Sarah Doe's college gpa based on her verbal SAT score of 520.

3) Which SAT score is a better predictor of college gpa?

4) (a) Predict college gpa based on math SAT.
(b) Predict college gpa based on verbal SAT.
(c) Predict college gpa based on math and verbal SAT
(d) Why does adding verbal SAT to the math SAT increase r-squared by such a small amount?

5) Consider models predicting college gpa based on high school math, science, and english. Which combination of factors do you belive lead to the "best" model for predicting college gpa? Explain your rationale.