PUBPOL 5310
Fall 2025
I would like to give some examples of some questions I’ve asked in the past, on courses that are related to PUBPOL 5310. These are not perfect review questions for our exam.
And the topics covered (and order of topics covered) were different in these prior years. So the questions below are not fully representative of an exam. But they give some sense of some of the types of questions I’ve asked.
A few notes:
• In the older exams, I was having students use calculators to perform. calculations in those exams; this won’t be needed for our exam. I may have you perform very simple (integer) calculations; if so, these will be do-able by hand.
• These are just cut-and-paste examples from old exams, selected to overlap with
topics that we have covered so far. This isn’t meant to represent the length our exam will take.
• Don’t worry about the exam #s or points #s. They are not meaningful given this cut- and-paste exercise.
--
1. [8 points] For each of the following examples state whether the data are numerical or categorical. Also state whether the data are cross-section, time series, or panel data
Numerical or Categorical? Cross-section, Time
series, or Panel?
|
Data on Sales this quarter by each of 23
Sales Representatives
|
|
Data on whether the S&P 500 stock
market index rose or fell each day, for all
trading days in 2018
|
|
Data on mortality rates in each US State
from 1978-2010
|
Data on political party affiliation, for a sample of eligible voters in 2020.
3. Consider the following output from Stata, summarizing information on a variable
“yrseduc” which contains the number of years of schooling for 65,685 adults sampled from the US population in 2012.
|
yrseduc
|
|
|
Percentiles
|
Smallest
|
|
|
1%
|
6
|
0
|
|
|
5%
|
10
|
0
|
|
|
10%
|
12
|
0
|
Obs 65685
|
|
25%
|
12
|
0
|
Sum of Wgt. 65685
|
|
50%
|
14
|
|
Mean 14.07761
|
|
|
|
Largest
|
Std. Dev. 2.718555
|
|
75%
|
16
|
20
|
|
|
90%
|
18
|
20
|
Variance [omitted by Doug]
|
|
95%
|
18
|
20
|
Skewness -.06566363
|
|
99%
|
20
|
20
|
Kurtosis 5.261059
|
3.1 [3 points] What is the variance of yrseduc in this sample?
3.2 Use the output to assess whether the data are skewed, and if so in what direction. Give two specific measures to justify your conclusion.
Question 3 [4 points] We estimate the regression GDP_pc = b1 + b2 ln(X), where GDP_pc is total real GDP divided by the population, and X is the number of individuals working in the high-tech sector, how do you interpret b2? Be as precise as possible
Question 5 [26 points] When a newborn is classified as “at risk,” doctors will perform extra checks and interventions. You are interested in measuring the financial costs of these interventions, and so you have collected a data set where the unit of observation is an infant. First you run a regression of “Hospital Expenditures” on a dummy variable for “at risk” . (mean hospital expenditures per child are $85,000, with a std. deviation of $15,000) Your regression results are:
|
Hospital Expenditures
|
=
|
79500
(10,000)
N = 63
|
+ 20,000 * AT_RISK (5,000)
R-squared = 0.07
|
a. [6pts] What are the average expenditures for a child “not at risk”? What are the average expenditures for a child “at risk”?
You now run a multivariate regression which also controls for birthweight and get the following:
Hospital Expenditures = 79,500 + 10,000 * AT_RISK - 30 * Birthweight
(12,000) (4,000) (10)
N = 63 R-squared = 0.27
b. [5pts] What are the predicted expenditures for a child who is “not at risk” and has a birthweight of 2000 grams?
c. [4 pts] Comparing two babies with the same at risk status but where one has a higher birthweight than the other by 500 grams, would expenditures be higher or lower for the heavier baby? By how much?
d. [8 pts] Interpret the coefficient on AT_RISK (b2=10,000). Be as precise as possible
Question 5. Below are the results from a regression of automobile price on characteristics for a sample of 63 new car models. Car price (PRICE) is measured in dollars. Power (POWER) is a measure of how powerful the car’s engine is. Miles Per Gallon (MPG) is a measure of fuel efficiency. DOMESTIC is an indicator variable which equals 1 if the manufacturer was based in the USA. The numbers in parentheses are standard errors.
PRICE = 20000 + 375 POWER + 210 MPG - 5000 DOMESTIC R2=.35
(1.5) (100) (20) (30)
(a) (3 points) Explain in words what the coefficient in front of MPG is telling us in the equation.
(b) (2 points) Explain in words what the coefficient in front of DOMESTIC is telling us in the equation.
True/False. Choose the best answer. You do not need to show your work. (2 points each)
3. The OLS estimator minimizes the sum of squared vertical deviations of actual points from the regression line.
1. When sxy > 0, it must be the case that b2 > 0.
Multiple Choice. Choose the best answer. You do not need to show your work. (2 points each)
1. A covariate is the same thing as:
a. a test statistic.
b. a dependent variable.
c. a regressor.
d. an estimator.
1. Which of the following measures the spread of a random variable x?
(a) The median.
(b) The mode.
(c) The standard deviation.
(d) The 75th percentile.
8. In a bivariate regression of Y on X the estimated slope is 5 and the estimated intercept is 10. For a particular observation the actual X is 5. What is the predicted value of Y?
a) -5
b) 5
c) 20
d) 35
e) None of the above
3. Which of the following data series are least likely to be positively skewed?
a. A cross-section of income data
b. A cross-section of data on height
c. Payouts from lottery tickets
d. House prices
e. They are all positively skewed