1. Binary variables 0.

A) are generally used to control for outliers in D. There are slopes in dummy regression
your sample. models.
B) can take on more than two values. 7. A paired data set has
C) exclude certain individuals from your sample. n=5
D) can take on only two values Σ x = 15; slope = 1.9
2. The reason why estimators have a sampling Σ y = 27
distribution is that Σ xy = 100
A) economics is not a precise science. Σ x^ 2 = 55.
B) individuals respond differently to incentives. What is the intercept of the regression line?
C) in real life you typically get to sample many A. -0.3
times. B. 27
D) the values of the explanatory variable and the C. 0.3
error term differ across samples D. 20
3. Which of the following is false? 8. A set of paired data has n = 30, b = 1.45,
A. The error term incorporates all the factors
responsible for the difference between the “i” th
district’s average test score and the value and . A 95% confidence
predicted by the regression line. interval for the slope of the least squares’
B. The error term contains all the other factors regression line for all pairs in the population is
besides X that determine the value of Y, for a constructed. The degree of freedom is (linear
specific observation, “i” regression with single independent var)
C. β1 and βo are the parameters of the linear A. 30
regression line. B. 28
D. The SSR express the total variation of Y. C. 27
4. The sample average of the OLS residuals is D. 29
A) some positive number since OLS uses
squares.
B) zero. 9. Which of the following is true?
C) unobservable since the population regression I. The best regression line will be defined as the
function is unknown. one with the minimum sum of squared errors or
D) dependent on whether the explanatory SSR.
variable is mostly positive or negative II. TSS= SSR + ESS
5. The slope estimator, β1, has a smaller standard III. If R or linear correlation is 0.69, it means that
error, other things equal, if the variability of Y can be explained by the
A) there is more variation in the explanatory regression equation based on X.
variable, X.
B) there is a large variance of the error term, u. A. I only C. I and II only
C) the sample size is smaller. B. II and III only D. I, II and III
D) the intercept, βo, is small
6. Which of the following is false? 10. Which of the following is not an assumption of
A. A joint hypothesis is a hypothesis that the error in the regression model
imposes two or more restrictions on the (homoskedasticity)?
regression coefficients. A. Independent
B. Using t-stat for testing the joint Ho can be B. Normally Distributed
unreliable. C. Mean of at least 0
C. In the Dummy Regression Model, β1 is the D. Have a constant variance regardless of the
difference in population means for the value of X
population Di = 1 relative to the population Di =
11. Which of the following is true about constant.
Multicollinearity in Regression Models? C. Under, homoskedasticity, the variance of u
A. It happens when independent variables are does not depend on the regressor.
correlated with combination of other D. Heteroskedasticity is a special case of
independent variables, may create duplication Homoscedasticity.
of info. 16. The normal approximation to the sampling
B. The overall F test is still valid. distribution of ˆβ is powerful because___
C. The hypothesis testing related to individual A) many explanatory variables in real life are
coefficients is not valid. normally distributed.
D. All of the above B) it allows econometricians to develop methods
for statistical inference.
12. SER is the _______ C) many other distributions are not symmetric.
A. Estimator of the standard deviation of the D) is implies that OLS is the BLUE estimator for
error term, u. β1.
B. One that measures the spread of the 17. Which of the following statements are true
observations around the regression line. I. The larger the population, the lesser the
C. A and B variance of the OLS estimator
D. None of the above II. If the 3 least square assumptions hold and if
the error is homoscedastic, the OLS is has the
13. Which of the following is not part of the least smallest variance, conditional on X.--- Basically
squares assumption? BLUE (Best Linear Conditionally Unbiased
A. The explanatory variable in regression model is Estimator)
normally distributed. III. OLS estimator has normal sampling
B. The error term has a conditional mean of 0, distribution when the sample size is large.
given X A. I only C. I and III only
C. (Xi, Yi), i=1, …., n are Independently and B. I and II only D. I, II, III
identically distributed. 18. Given the following, solve the expected salary
D. Large outliers are unlikely. difference between an academic and a
professional.
14. Given a regression line when X is a binary Y= βo+ β1Di
variable, D= 1 if professional, 0 if academic
R^2= 0.037 A. The professional has β1 more expected salary
SER= 18.70 than the academic
se of Bo = 1.3 B. The professional has βo less expected salary
se of B1= 1.8 than the academic
Equation: Y= 650+ 7.4D C. The academic has βo less expected salary
n =420 than the professional
What is the result of your hypothesis testing at D. The academic has β1 more expected salary
95% confidence level? than the professional
A. Reject Ho
B. Accept Ho.
C. Do not Reject Ho
D. None of the Above

15. Which of the following is false?

A. Whether the errors are homoscedastic or
heteroskedastic, the OLS estimator is unbiased,
consistent, and asymptotically normal.
B. Homoskedasticity happens when u, given Xi, is
 C. A and B
D. None of the above
23. Which of the following is false under hypothesis
testing?
A. If the p value is 0.05 or less, reject the Ho.
B. If the calculated t value (absolute value) is less
than the absolute value of t value from table, do
not reject Ho.
C. If the computed F value is more than the
benchmark F value, reject the Ho.
D. If the F value is significant, it does not mean
all indv variables are significant.
19. 24. Given the ff
Which of the following is true? Yi = β0 + β1 ( X1) + β2 (X2) +ui
A. T refers to the ESS Which of the following is true?
B. The Total Sum of Squares is U A. B1 is the partial effect of X1 on Y
C. T + U = V B. B2 is the full effect of X2 on Y
D. None of the above C. A and B
20. Which of the following is true? D. None of the Above
A. MSE or Mean Squared Error tends to amplify 25. Given the ff, which of the following is true?
the impact of outliers on the model’s accuracy. Score = 698.9 − 2.28 × STR
B. The problem with R² is that it keeps on Score = 686.0− 1.10 × STR − 0.65 × PctEL A. The
decreasing as you increase the number of coefficient on STR falls from −2.28 to −1.10
variables, regardless of the fact that the new when we hold constant the effect of percent
variable is actually adding new information to English learners.
the model. B. The first estimate showed omitted variable
C Omitted Variable Bias occurs when the bias because it reflected both the effect of a
omitted variable is not a determinant of the change in the Student-Teacher Ratio and the
dependent variable. omitted effect of more English learners.
D. None of the Above C. The Percent English learners tends to lower
21. Given the following information, what is the the district test score (holding constant
difference between the salary of an assistant classroom size)
and an associate? D. All of the above
Salary = 65.25+41.71Assistant+45.32Associate+ 26. Given the following equation, find the change in
72.11Fulltime Y (1% increase in X in I and III, 1 unit increase in
A. The associate gets 45.32 more X in II)
B. The associate gets 3.61 more I. Y= 25+3lnIncome = 0.03 unit increase
C. The assistant gets 3.61 more II. ln(Sales) = 2.5+0.05Products = 5% increase
D. The assistant gets 41.71 less. III. ln(Phones)= 20+0.5ln(Copper) = 0.5%
increase

22. Which of the following is true?

A. To determine the estimator coefficients, one
must minimize SSR
B. Percentage of English Learners in the US and
the percentage of district residents who are first
gen immigrants in the US are examples of
perfect multicollinearity.
27. Given the equation, if the person is Chinese, the 30. Which of the following is true about the
effect of getting married on Y is ___ interactions between 2 continuous variables?
Y= Bo+ B1(D1)+ B2(D2)+ B3(D1 x D2) A. The interaction term allows the effect of a
D1 = 1 if Chinese, 0 if American unit change in X1 to depend on X2 and vice
D2 = 1 if Married, 0 if Single versa.
Y = ln (Rice Consumption) B. Given the equation:
A. B2+B3 Yi= Bo+B1 (X1) +B2(X2)+B3 (X1 x X2)+u ,
B. B3 The effect on Y when there is a change in X1,
C. B2 while X2 is constant is (B1+ B3(X2) ) Δ X1
D. B1+B2+B3 C. In nonlinear regression, the slope of the
28. Given the equation, if the person uses a popular regression function, depends on the
Samsung phone, the effect of getting married on value of one or more of the independent
Y is ___ variables
Y= 200+ 50(D1)+ 5(D2)+ 10(D1 x D2) D. All of the above
D1 = 1 if Apple, 0 if Samsung 31. Which of the following is true?
D2 = 1 if Married, 0 if Single A. Adjusted R^2 cannot be used to compare two
Y = ln (Mobile Data Consumption) regression models if their dependent variables
A. 15 are different.
B. 5 B. R^2 is the fraction of the sample variance of Y,
C. 65 explained by the regressors (Xi)
D. 10 C. SSR usually decreases when a new regressor is
29. Given the following binary-continuous added.
regression model, max each model with the D. R^2 measures how much prediction error is
description eliminated when we use Least Sq Regression
Y= ln(Salary) E. All of the above
X1 = Work exp in years F. None of the above
Di =1 if college grad, if not = 0 ; 32. Given the equation below, what is the difference
I. Y= Bo+B1 (X1)+B2 (Di) + u in the effect of changing the % mineral content
II. Y= Bo +B1(X1)+ B2(Di)+ B3(X1 xD2)+u by 1 unit given 2mL of water and 10mL of
III. Y=Bo+B1(X1)+B2(X1 x Di)+u water, on ln (purity of gold) or Y.
A. effect of an addt’l year of work exp is the same Y= 6+ 1.5PoMC-2mlW+0.5(PoMC x mlW)
for college grads and non-grads due to same
slope; takes into account the difference of the A. If mlW= 10, a 1 unit increase in on mineral
effect of additional year of work for college and content percentage increases Y by 1.5 units
non-grads; same base salary regardless if college more than if mlW= 2
grad or not if no prior exp. B. If mlW = 2, a 1 unit increase on mineral
B. takes into account the difference of the effect content percentage increases Y by 1.5 units
of additional year of work for college and hs more than if mlW =10
grads; same base salary regardless if college C. If mlW =10, a 1 unit increase on mineral
grad or not if no prior exp; effect of an addt’l content percentage increases Y by 4 units more
year of work exp is the same for college grads than if mlW = 2
and non-grads due to same slope D. None of the Above
C. effect of an addt’l year of work exp is the
same for college grads and non-grads due to
same slope; same base salary regardless if
college grad or not if no prior exp; ; takes into
account the difference of the effect of additional
year of work for college and hs grads.
D. None of the above.