School of Economics and Business Administration

University of Navarra
Academic year: 2024/25
Econometrics I
Problem Set III: Ch. 4

NOTE: Please remember that problem sets do not count for the final grade, so it is not
needed to hand in the solutions for this problem set. However, it is highly recommended
trying to solve the questions before the practice class.

PROBLEMS

1. Suppose that a researcher, using wage data on 200 randomly selected male workers
and 240 female workers, estimates the OLS regression:

W ˆage = 10.73 + 1.78 ∗ M ale,

(0.16) (0.29)
2
R = 0.09, SER = 3.8

where W age is measured in dollars per hour and M ale is a binary variable that is
equal to 1 if the person is a male and 0 if the person is a female. Define the wage
gender gap as the difference in mean earnings between men and women.

(a) What is the estimated gender gap?

(b) Is the estimated gender gap significantly different from 0? To do so, compute
the p-value for testing the null hypothesis that there is no gender gap.
(c) Construct a 95% confidence interval for the gender gap.
(d) In the sample, what is the mean wage of women? Of men?
(e) Another researcher uses these same data but regresses Wages on Female, a
variable that is equal to 1 if the person is female and 0 if the person a male.
i. Complete the following equation, where γ̂0 and γ̂1 are the estimated OLS
coefficients on the regression of W ages on the dummy variable F emale:

W ˆage = γ̂0 + γ̂1 ∗ F emale,

R2 = , SER =

ii. Without further calculations, can we know the SE of γ̂0 and γ̂1 ? Explain.

1
2. In the 1980s, Tennessee conducted an experiment in which kindergarten students
were randomly assigned to “regular” and “small” classes and given standardized
tests at the end of the year. (Regular classes contained approximately 24 students,
and small classes contained approximately 15 students). Suppose, in the population,
the standardized tests have a mean score of 925 points and a standard deviation of
75 points. Let SmallClass denote a dummy variable equal to 1 if the student is
assigned to a small class and equal to 0 otherwise. A regression of T estScore on
SmallClass yields:
ˆ
T estScore = 918 + 13.9 ∗ SmallClass,
(1.6) (2.5)
2
R = 0.01, SER = 74.6

(a) Do small classes improve test scores? By how much? Is the effect large?
Explain.
(b) Is the estimated effect of class size on test scores statistically significant? Carry
out a test at the 5% level.
(c) Does SLR.3 plausibly hold for this regression? Explain.
(d) About the variance of the error term.
i. Do you think that the regression errors are plausibly homoskedastic?
Explain.
ii. SE(β̂1 ) was computed using the robust-to-heteroskedasticty formula. Sup-
pose the regression errors were instead homoskedastic. Would this affect
the validity of the confidence interval constructed at the beginning of the
exercise? Explain.

3. 25 workers in a factory were asked to evaluate their satisfaction with their own work,
on a scale from 0 to 10. In addition, the total number of missing days during the
last year for these employees were also obtained. With this data, the OLS regression
line obtained was:
ŷ = 13.6 − 1.2X

The following information is also available: X̄ = 6, ni=1 (Xi − X̄)2 = 130, ni=1 (Yi −
P P
Ȳ )2 = 210. Let suppose that the estimated model satisfies all the assumptions of
the simple linear regression model with normal errors.

(a) Test, using α = 0.01, the null hypothesis that the degree of satisfaction with
your own work does not have an effect on absenteeism.
(b) A particular employee has a degree of satisfaction with her own work equal 4.
Determine the distribution of the random variable indicating the missing days
at work in a year for this employee, and compute the unbiased estimates for
the mean and the variance of this random variable.

2
4. Analyse whether the following statements are true or false. If they are true, prove
them and if they are false, justify the reason why.

(a) In a simple linear regression model, if we have a large sample and the null
βˆ1 −1
hypothesis β1 = 1 is true, then the approximate distribution of SE is normal
(βˆ1 )
standard, even if the error term is not normally distributed.
(b) Let Xi denote a dummy variable, and consider the regression Yi = β0 +β1 Xi +ui .
Let Ȳ0 denote the sample mean for observations with X = 0, and let Ȳ1 denote
the sample mean for observations with X = 1. Then, β̂0 = Ȳ0 and β̂1 = Ȳ1 .
(c) A researcher has two independent samples of observations on (Yi , Xi ). To
be specific, suppose Yi denotes earnings, Xi denotes years of schooling, and
the independent samples are for men and women. Write the regression for
men as Yi,m = β0,m + β1,m Xi,m + ui,m and the regression for women as Yi,w =
β0,w +β1,w Xi,w +ui,w . Let β̂1,m denote the OLS estimator constructed using the
sample of men, β̂1,w denote the OLS estimator constructed from the sample of
women, and SE(β̂1,m ) andr SE(β̂1,w ) denote the corresponding standard errors.
 2  2
Then, SE(β̂1,m − β̂1,w ) = SE β̂1,m + SE β̂1,w .