Homework assignment 1
Course: Introductory Econometrics
Lecturer: Dmytro Vikhrov
Due date: October 28, 2014.
Instructions. Use wage1.gdt dataset for this assignment (File → Open data →
Sample file... → Wooldridge tab). You can work in groups, but each student must
submit their own answers.
Problem 1
There are four occupation types in the dataset. You need to have one dummy variable
for each occupation. Since there are only three dummy variables, you have to create a
fourth one. For this go to Add → Define new variable... and type the following code:
series base occ = (profocc != 1 && clerocc != 1 && servocc != 1)
To make sure that it worked, create another variable:
series tot = base occ + profocc + clerocc + servocc
If you did everything correctly, then variable tot should be a column of ones.
Problem 2
1. Regress wage on four occupation dummies (from Problem 1), without an intercept.
That is, run the following regression:
wageij = β1Di,base + β2Di,prof + β3Di,cler + β4Di,serv + ij.
What do ˆβ1, ˆβ2, ˆβ3 and ˆβ4 mean?
2. Construct 95% confidence intervals for mean wages by occupation. Interpret the
meaning of each confidence interval.
Problem 3
1. Using GRETL (Tools → Test statistic calculator → 2 means ), perform t-test on
the equality of means for each pair of occupations (six pairs). Fill in the table on
page 2. In each cell write: H0 and HA, test statistics and p-value. Also, indicate
which hypothesis you reject / do not reject. Use help button in the Test statistic
calculator window if you need help on syntax.
2. Test the equality of variances for each occupation pair and fill in the table on page 3.
TESTS ON MEANS
Servocc Clerocc Base_occ Profocc
Servocc
Clerocc
Base_occ
Profocc
TESTS ON VARIANCES
Servocc Clerocc Base_occ Profocc
Servocc
Clerocc
Base_occ
Profocc