Module 3, Week 1, gretl problem set 5

Use gretl and the "Georgia Public High Schools" dataset to answer the following questions.

You are interested in the relationship between SAT scores and teacher's salary.

Only use observations where the school type is "regular".

1. Produce a scatterplot between the variables "sat_total" and "tch_avg_salary".

2. Calculate descriptive statistics for the variables of interest.

3. Calculate the correlation coefficient between the variables of interest.

4. Estimate a simple linear regression model using least squares where "sat_total" is the dependent variable and "tch_avg_salary" is the predictor.

a) Write out the estimated regression equation.

b) Interpret the estimated slope coefficient.

c) Is the estimated slope coefficient statistically significant? How do you know?

d) Calculate a 95% confidence interval for the estimated slope coefficient.

e) Interpret the model's R-squared.

f) Predict average SAT score for a school where teachers have an average salary of

$60,000.

g) Calculate a prediction interval for the prediction asked for in part f) of question 4.

5. Produce the scatterplot asked for in question 1 but now include the estimated linear regression equation.

6. Give an example of another variable that is likely an important predictor of SAT scores.