ECONOMETRICS Đinh Văn Thành - K55 KTQT

st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

CHAPTER 1: INTRODUCTION TO ECONOMETRICS

Question 1: What is the main analytical tool that Econometrics uses?

A. Regression analysis C. Cause and effect analysis
B. Correlation Analysis D. Both A, B and C

Question 2: The first step to take when researching an econometric problem is:
A. Build the model C. Study documents, theorem, theory related to the research problem
B. Collect the data D. Estimate the model parameters

Question 3: Which type of data does not show the change of the variables over time?
A. Cross-section data C. Panel data
B. Time series data D. None of these above

Question 4: The difference between economic model and econometrics model is:
A. The economic model exists only in theory
B. The econometric model is built independently on the economic model
C. Economic models cannot use data
D. The econometric model has disturbing factors

Question 5: Research on the relationship between salary and education level of Northern provinces
in May 2018. What kind of data is used?
A. Cross-section data C. Panel data
B. Time series data D. None of these above

Question 6: Which of the following is true about the correlation coefficient?

A. It is an indicator to evaluate the tight of a single linear correlation relationship
B. Determine the strength of the relationship from which to choose the primary or secondary cause for the
research phenomenon.
C. Determine the specific direction of the relationship (positive - negative).
D. All of these above are correct

Question 7: Which of the following statements is incorrect:

A. r → - 1: X and Y are closely linearly related.
B. r > 0: X and Y are positively linearly related
C. r → 0: X and Y are loosely linearly related.
D. r = 0: X and Y are independent

Question 8: Regression analysis

A. shows the strength and weakness of relationships among variables B. is a one-way analysis
C. does not distinguish roles between variables D. contains causal relationship

Question 9: Which of the following statements is incorrect:

A. Regression relation includes causal relationship
B. Relation among economic variables is statistical relation

https://www.facebook.com/thanh.d.0715 1
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

C. Correlation relations do not distinguish the roles of variables

D. Regression relation is the dependence of one variable on one or more other variables

Question 10: The independent variable is also called:

A. The explained variable C. The unexplained variable
B. The explanatory variable D. All are incorrect

Question 11: The factor Y in regression analysis is:

A. The influencing factor
B. The influenced factor
C. Both the influencing factor and the influenced factor
D. All are incorrect

Question 12: How many types of data are there for econometric research:
A. 1 type B. 2 types C. 3 types D. 4 types

Question 13: In regression analysis, which is not the basis for determining the role of independent
variables and dependent variables?
A. Theoretical basis and research hypothesis
B. Theoretical basis
C. Practical experience of the researcher
D. Research hypothesis

https://www.facebook.com/thanh.d.0715 2
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

CHAPTER 2: TWO-VARIABLE REGRESSION MODEL

Questions about regression model

Question 1: Residual is:
A. The variation between real value and expected value
B. The variation between real value and mean value
C. The variation between mean value and expected value
D. None of those above

Question 2: In OLS estimation, which factor gets minimun value?

A. Residual B. Residual squared C. Intercept D. Slope

Question 3: Given these following models:

(1) Yi = β̂ 0 + β̂ 1 Xi + ui ̂ i = β̂ + β̂ Xi + ui
(2) Y 0 1
̂ i = β̂ + β̂ Xi
(3) Y 0 1
Which model of these above is correctly performed:
A. (1) B. (2) C. (3) D. (1) and (3)

Question 4: Which of the following equations is not a linear regression?

A. Y = β0 + β1 X1 + ui C. lnY = β0 + β1 lnX1 + ui
B. Y = β0 + β1 lnX1 + ui D. Y = β0 + β21 X1 + ui

Question 5: Given the model: Y ̂ = β̂ + β̂ X. Regress the model with OLS method; i = 1, 2, …, n.
1 2
Estimation of β̂ is unbiased, which means:
2
A. E(β̂ 2 ) = β2 C. E(β2 ) = β̂ 2
B. β̂ has BLUE properties.
2
D. E(β2 ) = 0

̂ i = β̂ + β̂ Xi , i = 1, 2, …, n. Given var(X) = 2 cov(X,Y), if the model is

Question 6: Given the model: Y 1 2
estimated by OLS method then the estimated β̂ is: 2
A. 2 B. 0,5 C. 0,2 D. 0

Question 7: Given the regression model Yi = β1 + β2Xi + ui:

A. β1 is the intercept, β2 is the slope, ui is the residual
B. β1 is the slope, β2 is the intercept, ui is the residual
C. β1 is the intercept, β2 is the slope, ui is random error
D. β1 is the slope, β2 is the intercept, ui is random error

̂𝐢 = 20 + 0,75Xi. The value of the residual at the

Question 8: Given a sample regression model 𝐘
observation X = 100, Y = 90 is:
A. 10 B. -5 C. 5 D. Not determined

Question 9: What does the model residuals represent in practice:

A. Wrong function form C. The theory is not enough
B. Missing data D. Both A, B and C are correct

https://www.facebook.com/thanh.d.0715 3
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

Question 10: A linear regression model is possible to have:

A. One dependent variable and many independent variables
B. Many dependent variables and many independent variables
C. One independent variable and many dependent variables
D. All of these above

̂ i = 120 + 2,5Xi. The value of residual at the observation

Question 11: Given the regression model: Y
with X = -16, Y = 100 is:
A. 60 B. 20 C. -20 D. Not enough information to calculate

Question 12: Which of the following is a linear regression model:

A. Yi = exp(β1 + β2Xi + ui) C. Yi = β1 + β22 Xi + ui
B. ln(Yi) = β1 + β2/Xi + ui D. Yi = ln(β1) + β2Xi + ui

Question 13: Error is:

A. The difference between the actual value and the estimated value
B. The difference between the average value and the estimated value
C. The difference between the sample regression model and the population
D. All are wrong

Question 14: Which equation is not linear regression?

A. ln 𝑒𝑖2 = 𝛽1+ 𝛽2 ln 𝑋𝑖 + 𝑣𝑖 C. | 𝑒𝑖 | = √𝛽1 + 𝛽2 𝑋𝑖2 + 𝑣𝑖

1
B. 𝑒𝑖2 = 𝛽1+ 𝛽2 ln𝑋 2 + 𝑣𝑖 D. | 𝑒𝑖 | = 𝛽1 + 𝛽2 + 𝑣𝑖
√ 𝑋𝑖

Questions about the assumptions of the OLS method

Question 15: In OLS estimation:
A. RSS is minimized B. ESS is minimized C. R2 is minimized D. All of these above

Question 16: Which of these following ideas is proved by Gauss-Markov theory?

A. OLS method gives unbiased estimated coefficients
B. OLS method gives smallest variance estimations of coefficients
C. OLS method only gives estimations of variance coefficients that are minimized when the estimations
are linear and unbiased
D. OLS method guarantees that errors are normally distributed

Question 17: The properties of OLS estimation are:

A. Linear, Unbiased, Best C. Unbiased
B. Linear, Unbiased D. Linear, Efficient

Question 18: If the hypothesis of the conditional expectation of the non-zero random error is satisfied,
we have:
A. The estimated parameters are linear, unbiased, and best @KXU3IU\GXOGTIK3HKZ]KKT3
B. The independent variables are not correlated with the random error in the model [O3GTJ3>O3

C. The overall regression line is horizontal

https://www.facebook.com/thanh.d.0715 4
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

D. The exact values of the parameters in the model can be determined

Questions about variance and standard error

Question 19: Which of the following is not a way to improve the accuracy of OLS estimates in a
simple regression model?
A. Reduce the sample size C. Divide both sides of the equation by the sample variance
B. Take the logarithm of the dependent variable value D. None of the above three methods

Question 20: Company X has a portfolio with two investments A and B, of which investment A
accounts for 40%. We get the yield variance of the two investments as var(RA) = 0,09, var(RB) = 0,16,
and the covariance cov(RA,RB) = 0,24. The standard deviation of the return on a portfolio is:
A. 0,433 B. 0,1872 C. 0,552 D. 0,743

Question 21: OLS estimations are unbiased when:

A. Heteroscedasticity happens B. Expected random error differ from
C. Model is lack of variables D. The sample is not random

Question 22 – 23 using this following information:

Assuming the correct model showing the relationship between education (educ - years of training, units in
years) and wages (wage - hourly wages, in dollar) as wage = β1 + β2educ + ui
Question 22: Which of these following statements is true:
A. The unbiased property of OLS estimators means E(β̂ 1 ) = β2, E(β̂ 2 ) = β1
B. The least squares method for best unbiased unlinear estimates
C. Independent variables with a large degree of variability (or ∑(Xi - X ̅ )2) and large sample size are good
signs because they improve the accuracy of the estimated parameters.
D. None of these above

Question 23: Assume that we estimate the above model through a data sample and obtain the
following estimates: 𝛃̂ 𝟏 = 3 and 𝛃
̂ 𝟐 = 1,5. Which is not a prediction from the obtained model?
A. As the number of years of training increases by 1 year, the average hourly wage increases by $1.5
B. A person without training is paid an average of $3 per hour
C. A person with 16 years of training is paid $27 per hour in average
D. For every increase by 1.5 years of training, the average hourly wage increases by $1

Question about coefficient of determination

Question 24: Coefficient of determination estimates:
A. Correlation between X and Y C. Covariance between X and Y
B. The amount of variation in the value of the dependent variable Y D. Ratio between ESS and TSS

Question 25: The estimated slope of the single regression model is 0 when:
A. 0 < R2 < 1 B. R2 = Y̅ C. R2 = 0 D. R2 > ESS/TSS
X狤3UL3`KXU3SKGTY3ZNGZ3ZNKXK3OY3TU3XKRGZOUTYNOV3HKZ]KKT3ZNK3XKMXKYY3GTJ3GTJ3ZNK3XKMXKYYUX3

]NGZYUK\KX3OK3ŜϽ狤3#3狢3
Question 26: Given R2 = 1. Which of these followings are true:
(ii) All data points lie on a straight line
(ii) All residuals are equal to 0
https://www.facebook.com/thanh.d.0715 5
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

(iii) Total variation of y around the mean value is explained by the model
(iv) The fitted line is horizontal for all explanatory variables
A. (ii) and (iv) B. (i) and (iii) C. (i), (ii) and (iii) D. All of these above

Question 27: Using data including 100 observations, we can estimate the following model: Yi = 8 +
5Xi + ei . Given that R2 = 0,86. Correlation coefficient between X and Y is:
A. 0.86 C. -0.86
B. 0.927 D. Can’t determine with given information

Question 28: When estimated a regression model using a data set of 50 observations, we obtain RSS
= 120, ESS = 356. The value of the determination coefficient R2 is:
A. 0.252 B. 0.748 C. 0.337 D. None of these above

Question 29: Coefficient of determination indicates:

A. Percentage of the mean value of the dependent variable is explained by the independent variables
B. Percentage of the mean value of the dependent variable is explained by the residuals
C. Percentage of the mean value of the dependent variable is explained by the estimators of independent
variables
D. None of these above

Question 30: The correlation coefficient is -0.8, then the change in Y explained by X is:
A. 80% B. 64% C. 92% D. Cannot be calculated

Questions about the statistical properties of OLS estimates

Question 31: The dependent variable Y in regression analysis is:
A. Random B. Given C. Non-random D. Both A, B and C are correct

Question 32: The dependent variable has:

A. Poisson distribution C. Standard distribution
B. Student distribution D. Bernoulli distribution

Question 33: The sample mean of the variables in the model:

A. Has a constant error variance C. Coincides with the population mean
B. Lies on the sample regression line D. Cannot be determined

Questions about the measure units

Question 34: Estimating a regression model we have the following results: Yi = 8 + 5Xi + ei . If we
change the ratio of the independent variable X by multiplying X by 0,5, the new intercept and slope
will be:
A. 4 and 2,5 B. 8 and 2,5 C. 8 and 10 D. 16 and 10

https://www.facebook.com/thanh.d.0715 6
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

CHAPTER 3: MULTI-VARIABLE REGRESSION MODEL

Questions about regression model

Question 1: Which is the correct form of an polulation regression model?
A. Yi = β1 + β2X2i + β3X3i C. Yi = β1 + β2X2i + β3X3i + ui
B. E(Y) = β1 + β2X2i + β3X3i + ui D. ̂
Yi = β1 + β2X2i + β3X3i + ui

Question 2: After estimating the model, we SRF = Y ̂ = 12,1 + 2,4X1 – 1,8X2 . Now the residual at the
observation with the coordinate Y = 25,19, X1 = 9, X2 = 6 is:
A 22,9 B. 2,29 C. 12,49 D. None of these above

Question 3: Given the regression model: Yi = β1 + β2X2i + β3X3i + ui với i = 1, 2, 3, ..., 98. The degree
of freedom of the model is:
A. 96 B. 95 C. 94 D. 93

Question 4: Estimate a regression model with 3 independent variables using sample data of 150
observations. The degree of freedom of this model is:
A. 148 B. 147 C. 146 D. 145

Question about coefficient of determination

Question 5: Which of the following is true when using adjusted R^2?
A. Adjusted R2 always increases when adding variables to the model
B. Adjusted R2 often results in more statistically significant variables and more statistically insignificant
variables for multivariate models
C. Adjusted R2 cannot be compared for models with different numbers of explanatory variables
D. Adjusted R2 cannot be compared for models with different explanatory variables

Question 6: Which of the following is true about the coefficient of determination:

(i) Indicates that the variation in the sample of independent variables is explained by the dependent variable
(ii) Can be negative when the model has no intercept
(iii) A higher coefficient of determination indicates a more efficient model
(iv) To choose which model is better, the coefficient of determination is always used for all situations
A. (i), (ii) and (iii) are correct C. Only (ii) is correct
B. (iii) and (iv) are correct D. (ii) and (iv) are correct

Given the following linear regression equation, where Y is the revenue (hundred million VND), X is the
selling price (million VND), Z is the advertising cost (million VND), answer 2 questions 7, 8:
̂ = 5.821 - 2.124X + 2.678Z
Y
t (9.39) (-59) (12.81)
se (0.62) (0.036) (0.209) R2 = 0.9

Question 7: “R2 = 0.9” means:

A. 90% of average sales are explained by changes in selling prices
B. 90% of average sales are explained by changes in sales prices and advertising costs
C. 90% of average sales are explained by changes in advertising costs
https://www.facebook.com/thanh.d.0715 7
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

D. None of these above

Questions about function forms

Question 8: Based on the estimation results, we can explain that :
A. Excluding the impact of advertising costs, if the selling price increases by 1 million VND, the average
revenue will increase 212.4 million VND.
B. Excluding the impact of advertising costs, if the selling price drops by 1 million VND, the average
revenue will increase 212.4 million VND.
C. None of these above

Question 9: In the regression model lnQi = 2.25 – 0,7lnPi + 0,02Yi where Pi is price and Qi is the
quantity demanded for a certain good, Yi is the available income. What is the significance of the
coefficient 0,7?
A. If the price increases by 1%, the quantity of demand will be 0,007% lower on average while the other
variables remain the same
B. If the price increases by 1% the quantity of demand will be 70% lower on average while the other
variables remain the same
C. If the price increases by 1% the quantity of demand will be 0,7% lower on average while the other
variables remain the same
D. None of these above

Question 10: Assume the following estimates of wage estimate for age as follow: Wage = 953,754 +
131,37age – 1,67age2 . Assume the independent variables are statistically significant. This model:
A. Follows the law of diminishing margins C. None of these above
B. Does not follow the law of diminishing margins

Question 11: (continued from Question 10) If the employee is 30 years old, the age increases by 1
unit, the average salary change is:
A. 31.17 B.131.37 C.3.34 D. None of these above

Question 12: Given the regression model Y = 4,4243 + 0,3892ln𝑿𝟏 + 0,7327ln𝑿𝟐 + e. Which of the
following statements is true?
A. If X1 decreases by 1 unit, Y increases by an average of 0.3892%, holding all other factors constant.
B. If X1 decreases by 1%, Y increases by an average of 0.3892%, holding all other factors constant.
C. If X1 decreases by 1%, Y decreases by an average of 0.3892 units, holding all other factors constant.
D. If X1 decreases by 1%, Y decreases by an average of 0.3892/100 units, holding all other factors constant.

Question 13: Which model describes the unstable marginal effect of the independent variable on the
dependent variable?
1
A. Y = β1 + β2 X + β3 X3i ui C. lnY = β1 + β2 lnX2i + β3 X3i + ui
2i

B. Y = β1 + β2 X2i + β3 X22i + ui D. Y = β1 + β22 X2i + β3 lnX3i + ui

Questions about variance and standard error

Question 14: The greater the correlation between independent variables in a multiple regression
model:
https://www.facebook.com/thanh.d.0715 8
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

A. The more inaccurate the OLS estimates will be

B. The less likely the sample will represent the population
C. The less likely random error expectation with the condition X is different from 0
D. There is no correct answer

Question 15: The variance of the estimates is smaller when

A. The sample size is smaller C. The sample size is larger
B. The relationship between the independent variables is stronger D. All of the above are incorrect

Questions from regression results

The information below is used for questions 16-20:
A researcher wants to investigate how a country's consumption depends on their inflation, population and
income.
In which:
c: real expenditure per capita ($)
inf: inflation rate (%)
pop: population (thousand people)
y: real income per capita ($)

Model 2: OLS, using observations 1959-1995 (T = 37)

Dependent variable: c

Coefficient Std. Error t-ratio p-value

const -2385.46 708.058 .............. 0.0019 ***
inf .................. 6.00019 -1.5528 0.1300 ***
y 0.589597 0.0479538 12.2951 <0,0001 ***
pop 0.0250876 0.00615279 4.0774 0.0003 ***

Mean dependent var 11328.65 S.D. dependent var 2505.241

Sum squared resid 341024.4 S.E. of regression ...............
R-squared ............... Adjusted R-squared 0.998353
F(3, 33) 7277.008 P-value(F) 1.32e-46
Log-likelihood -221.3834 Akaike criterion 450.7667

Question 16: The significance of the regression coefficient of variable y is:

A. The real average income increases by $ 1, the mean c-value increases by $ 0.589597, ceteris paribus
B. The real average income increases by $ 1, the mean c-value increases by 0.589597%, ceteris paribus
C. The real average income increases by 1%, the mean c-value increases by $ 0.589597, ceteris paribus
D. None of these above

Question 17: The significance of the regression coefficient of pop is:

A. Population increases by 1%, then average real expenditure increases by $ 0.0250876, ceteris paribus
B. Population increases by a thousand people, then average real expenditure increases by $ 0.0250876
C. Population increases by a thousand people, then average real expenditure increases by 0.0250876%,
ceteris paribus

https://www.facebook.com/thanh.d.0715 9
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

D. Population increases by a thousand people, then average real expenditure increases by $0.0250876,
ceteris paribus

Question 18: The standard error of the regression line is:

A. 2505,514 B. 18,52815 C. 101,6566 D. 2614,949

Question 19: The coefficient of determination of the model is:

A. 0.998491 B. 0.843914 C. 0.756981 D. 0.974773

Question 20: The significance of the coefficient of determination in the model is

A. 99.8491% of the mean of the dependent variable is explained by the independent variables
B. 99.8491% of the change of the dependent variable is explained by the random factors
C. 99.8491% of the change of the dependent variable is explained by the independent variables
D. 99.8491% of the mean of the dependent variable is explained by the random factors

The information below is used for questions 21-24:

Regress the electricity production cost (unit: million USD) by power output (thousand kWh), labor rental
(USD / year), cost price (unit: USD) and fuel price (USD per 1 million BTU). Regression results are as
follows:

Question 21: The residual sum of squares is:

A. 29147.0952 B. 28501.4815 C. 21859.4027 D. 21992.5929

Question 22: The estimated sum of squares is:

A. 852001.954 B. 488104.481 C. 369519.501 D. 182049.859

Question 23: The coefficient of determination of the model is:

A.0.9420 B. 0.9438 C. 0.9571 D. 0.9753

Question 24: The meaning of the coefficient of the variable Price_Lab is:
A. If labor cost increases by 1 USD, the average electricity production cost increases by 0.0036183 million
B. If labor cost increases by 1 USD, the average electricity production cost increases by 0.0036183 million
USD when other factors remain unchanged.
https://www.facebook.com/thanh.d.0715 10
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

C. If labor cost increases by 1%, electricity production cost increases by 0.0036183% when other factors
remain unchanged.
D. If labor cost increases by 1 USD, the average electricity production cost increases by 0.0036183%

https://www.facebook.com/thanh.d.0715 11
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

CHAPTER 4: STATISTICAL INFERENCE

General questions
Question 1: Which of the following quantities always takes a non-negative value?
A. t B. F C. cov(X, Y) D. R2

Question 2: Which of the following statements is FALSE regarding using p-value for tests
A. p-value is only used for 2-sided test
B. p-value is the exact significance level of the test
C. p-value is the level of marginal significance - where we will decide between rejecting or not rejecting
the null hypothesis.
D. If p-value is known, we can make statistical inference without checking statistical tables

Question 3: Significance level is 5%. Small p - value (less than 1%) indicates evidence:
A. Reject the null hypothesis C. Accept the null hypothesis
B. The t-statistic is less than 2.58 D. None of these above

Questions about one coefficient hypothesis testing

Question 4: The null hypothesis of the coefficient’s significance test is:
A. i = 0 B. i  0 C. i  0 D. i = 0

Question 5: To test whether one independent variable has more impact on the mean of the dependent
variable than the other, the counter hypothesis is:
 = j   j   j
A. i B. i C. i D. Đáp án khác

Question 6: Given the following regression results:

wage Coef. Std. Err. t P>|t| [95% Conf. Interval]

age 20.16885 3.923236 5.14 0.000 12.46944 27.86826

black -182.6703 37.03254 -4.93 0.000 -255.3472 -109.9934
educ 55.56984 5.638564 9.86 0.000 44.50407 66.63561
_cons 2065.76 152.7506 13.52 0.000 1765.985 2365.536

Given cov( black , educ ) = 37.590. The value of se( black − educ ) is:
A. 1328.022 B. 38.450 C. 37.459 D. 36.442

Used the following information for questions 7 - 12

Use sample data to estimate the dependence of wage (units: dollars) on years of experience (unit: years),
years of education (units: years), indicators IQ and the number of siblings in the house (unit: person) having
the following result:

https://www.facebook.com/thanh.d.0715 12
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

Question 7: The value of the coefficient of exper is:

A. 17.1867 B. 29.9401 C. 12.9249 D. 16.5304

Question 8: The variance of educ has a value of:

A. 7.105801 B. 40.04911 C. 50.49241 D. 20.19869

Question 9: The t-statistic value in the two-sided test of the variable iq is:
A. 7.02 B. 5.00 C. - 1.96 D. -3.81

Question 10: Is gaining 1 year of experience more effective than increasing 1 year of schooling in
improving wages? Given cov(β2 , β3 ) = 9.068, critical value c = 1.64
A. More effective B. No more effective C. Cannot determine

Question 11: Is the number of siblings in the house significant?

A. Statistically significant at the 5% level
B. Statistically significant at the 10% level
C. Statistically significant at 5% but not at the 10% level
D. Statistically insignificant at both 5% and 10% levels

Question 12: The meaning of the slope of the variable educ is:
A. When other factors remained unchanged, a 1-year increase in school attendance results in a 56,8143%
increase in average wages.
B. When other factors remained unchanged, a 1-year increase in school attendance results in a 56,8143
USD increase in average wages.
C. When other factors remained unchanged, years of schooling increased by 1%, resulting in a 56,8143%
increase in wages.
D. When other factors remained unchanged, years of schooling increased by 1%, resulting in a 56,8143
USD increase in wages.

Question 13: Given the linear regression equation as follow:

https://www.facebook.com/thanh.d.0715 13
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

Y = 0,5334391 + 0,0798558X1 + ei
se 0,1358994
t 15,46
The t-statistic value of β1 and se(β̂ 2 ) is:
A. 3,93 and 0,0051642 C. 7,76 and 2,9692015
B. 5,35 and 4,8302067 D. 3,93 and 8,8319481

Question 14: The variance of the estimated coefficient increases, the t-statistic will:
A. Decrease B. Increase C. stay unchanged D. Can’t conclude

Question about multi-coefficients testing

Question 15: Testing the model’s significance is to determine:
A. Whether all coefficients, including intercept, are zero concurrently
B. Whether the model is accurately estimated or not
C. Whether all the coefficients, excluding intercept, are concurrently zero
D. Whether we need to add additional independent variables to the model

Question 16: To test the significance of the regression model with 4 independent variables, the null
hypothesis is:
A. H0: R2 = 0 C. H0: β1 = β2 = β3 = β4 = β5 = 0
B. H0: β̂ 2 = β̂ 3 = β̂ 4 = β̂ 5 = 0 D. H0: β2 = β3 = β4 = β5 ≠ 0

Question 17: Given the following model: Yi = β0 + β1 X1 + β2 X2 + β3 X3 + ui . After running the

regression we obtain R2 (1) = 0,977. It is suggested that X2 , X3 should be excluded from the model.
The model Yi = β0 + β1 X1 + ui gives the R2 (2) = 0,958. The number of observations is 24. To test
whether X2 , X3 should be included in the model or not, compute the F-statistic equal to:
A. F = 9,09 B. F = 8,26 C. F = 16,52 D. None of these above

Question 18: With the result of Question 17, at the 5% level of significance, critical value c = 3,4928,
we can conclude that:
A. X2 , X3 cannot be excluded from the model C. None of these above
B. X2 , X3 can be excluded from the model

Questions about confidence interval

Question 19: The confidence interval has properties:
A. The left-sided confidence interval is used to estimate the maximum value of the regression coefficient
B. The right-sided confidence interval is used to estimate the minimum value of the regression coefficient
C. The bigger the interval is, the smaller the significance is
D. All of these above

Question 20: With reduced significance level, confidence interval widths will tend to:
A. Increase B. Decrease C. Stay unchanaged D. All of these above

Question 21: With increased confidence level, confidence intervals will tend to:
A. Increase B. Decrease C. Stay unchanaged D. All of these above
https://www.facebook.com/thanh.d.0715 14
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

Summary questions
Used the following information for questions 22-24
The model is estimated using 142 observations
Y = -12,3 + 0,32X2i + 0,043X3i – 0,005X4i + 0,0035X5i + ui
Se 6,89 0,15 0,078 0,0047 0,0022
p-value 0,07644 0,03468 0,58234 0,28928 0,11292 R2 = 0,0395

Question 22: What can we say about the coefficient of X5i ? (considering 2-sided test)
A. X5i increases by 1 unit then Y increases by 0,0035 on average while the others X stay the same. But this
is not statistically significant at the 5% level.
B. X5i increases by 1 unit then Y decreases by 0,0035 on average while the others X stay the same. This is
statistically significant at the 5% level.
C. X5i increases by 1 unit then Y increases by 0,0035 on average and this is statistically significant at the
1% significance level.
D. The coefficient X5i is statistically significant at the 5% level but not at the 1% level.

Question 23: The model now includes the variables X2i and X3i . We get R2 = 0,0387. Value of F-
statistic for the hypothesis test H0 : β2 = β3 = 0 is:
A. 32,821 B. 0,0570 C. 0,0808 D. Not enough information to calculate

Question 24: What can we say about the coefficient of X2i (considering the 2-sided test):
A. It is statistically significant at both 5% and 1% significance levels.
B. It is statistically significant at 1% significance level
C. It is statistically significant at the 1% level but not at 5% level
D. It is statistically significant at the 5% level but not at 1% level

Used the following information for questions 25 - 29

Using sample data to estimate the dependence of an infant's weight bwghtlbs (unit: pound) on the number
of cigarettes smoked per day of mothers cigs (unit: cigarettes), the order in which they were born parity,
the family income faminc (unit: thousand USD) and the number of school years of the father fatheduc (unit:
year), we have the following results:

https://www.facebook.com/thanh.d.0715 15
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

Question 25: The regression coefficient of the parity variable is valued as:
A. 0,74019 B. 0,11582 C. 0,81921 D. 0,30812

Question 26: The t-statistic of the fatheduc variable is:

A. 4,23 B. -1,58 C. 1,27 D. 3,64

Question 27: The value of the model's degree of freedom is:

A. 1186 B. 1187 C. 1188 D. 1189

Question 28: The confidence interval for the faminc variable is:
A. -0,8591043; 3,8467592 C. -1,7653926; 3,0142989
B. 4,5827013; 7,1048513 D. -0,0013838; 0,0074048

Question 29: The model's coefficient of determination is equal to:

A. 0,0345 B. 0,9655 C. 0,0378 D. 0,5629

https://www.facebook.com/thanh.d.0715 16
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

CHAPTER 5: THE ASSUMPTION VIOLATIONS OF THE MODEL

General questions
Question 1: Which problem can Robust regression fix?
A. Multicollinearity B. Heteroskedasticity C. Autocorrelation D. None of these above

Question 2: Given the function as follows: Yi = β0 + β1 X1 + β2 X2 + β3 X3 + ui

VIF X1 16.55
X2 38.86
X3 45.09
Does the model have any problems?
A. Autocorrelation C. Disturbance without standard distribution
B. Multicollinearity D. No problems

Question 3: Which of the following problems is found only in the multiple regression model?
A. Heteroskedasticity B. Multicollinearity C. Autocorrelation D. Omitted-variable bias

Question 4: The Breusch – Godfrey test is used to detect:

A. Multicollinearity C. Structural stability of regression models
B. Heteroscedasticity D. Autocorrelation

Question about Multicollinearity

Question 5: Multicollinearity occurs when:
A. The model has many explanatory variables
B. The dependent variable is represented by a linear combination of the remaining variables
C. 1 explanatory variable is represented by a linear combination of the remaining variables
D. None of these above

Question 6: Which case is not a sign of multicollinearity?

A. 𝑅 2 is large and t is small
B. When a variable is deleted, the regression results will change greatly
C. The confidence intervals of the regression coefficient estimates are narrow
D. The regression coefficient estimates are not significant

Question 7: A model has a correlation matrix table as follows:

X1 X2 X3
X1 1
X2 0.84243 1
X3 0.94565 0.93457 1
The model has a sign of:
A. Multicollinearity B. Autocorrelation C. Heteroskedasticity D. None of these above

Question 8: Which following method is not used to fix Multicollinearity?

A. Omitt 1 explanatory variable C. Change 2 explanatory variables to 1 ratio
B. Change data to logarithm D. Collect more data for each variable

https://www.facebook.com/thanh.d.0715 17
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

Question 9: Which of the following does not result from imperfect multicollinearity?
A. OLS estimators are sensitive to changes in data
B. Variance and Standard Error of estimator is big
C. Ineffective forecast
D. Small t statistic may result in the conclusion of rejecting H0 without enough bases.

Question 10: One of the evidences of multicollinearity is:

A. There’s not so many changes in estimators when removing a dependent variable from the model
B. The value of t corresponding important variables are huge
C. Variance matrix – covariance consists of small values
D. None of these above

Question 11: The consequences of multicollinearity do not cause:

A. The variance of the estimated coefficients is large
B. The variance of the estimated coefficients is small
C. The signs of the estimated coefficients may be wrong
D. Adding or removing variables that are collinear with other variables will change the sign or the
magnitude of the estimates in the model

Questions about Heteroscedasticity

Question 12: In the auxiliary model of Park test, the value used as the dependent variable is:
A. ei B. ln(e2i ) C. e2i D. No correct answer

Question 13: In the White test, Chi-squared statistic is 1,339 with the corresponding p-value of 0,512,
using the significance level of 0,05. We conclude that:
A. There exists heteroskedasticity C. Not enough information to conclude
B. There exists homokedasticity

Question 14: In the White test, if the original model has 2 independent variables, how many
independent variables are there in the auxiliary model:
A. 5 B. 11 C. 6 D. None of these above

Question 15: Suppose that estimating the model Yi = β1 + β2X2i + β3X3i + ui but the variance of ui is
proportional to X22 . To find the GLS estimate we regress:
A. Y based on an intercept, 1/X2 and X3/X2 C. Y/X2 based on an intercept, 1/X2 and X3/X2
B. Y/X2 based on 1/X2 and X3/X2 D. Can’t determine due to unknown ratio between var(ui) and X22

Question 16: Given regression equation Yi = β0 + β1 X1 + β2 X2 + β3 X3 + ui . Using White test to detect

heteroskedasticity, how many parameters are there in the auxiliary model?
A. 6 B. 7 C. 9 D. 10

Question 17: Given White test with a significance of 10%, does the model consist of
heteroskedasticity?

https://www.facebook.com/thanh.d.0715 18
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

A. No B. Yes C. Can’t conclude

Question 18: Method of using graph to detect heteroskedasticity is:

A. Draw graph of Y by each independent variable sequentially
B. Draw graph of residuals by independent variable sequentially
C. Draw graph of squared residuals by each independent variable sequentially
D. Both B and C are correct

Question 19: Dependent variable of White test is:

A. ln(e2i ) B. e2i C. |ei | D. σ2

Question about Autocorrelation

Question 20: In the Durbin - Watson test, if d- statistic is in the range dU to 4 - dU, the conclusion is:
A. There is positive autocorrelation C. There is no autocorrelation
B. There is negative autocorrelation D. Have not yet concluded

Question 21: What type of data that does NOT need to test for disturbance’s autocorrelation?
A. Time series data B. Cross-sectional data C. Panel data D. No correct answer

Question 22: If there is a negative autocorrelation, which of the following values would be suitable
for d statistics of Durbin – Watson test?
A. Nearly 0 B. Nearly 1 C. Nearly 2 D. Nearly 4

Question 23: Given the model: Yt = β1 + β2 Xt + Yt-1 + ut . To test for autocorrelation problem, we use:
A. Durbin-Watson statistic B. BG test C. d-statistic or BG test

Question 24: Which of the following conditions are required for Durbin-Watson test to be usable?
(i) The regression function must have an intercept
(ii) The independent variables are not randomly taken
(iii) There is no lag dependent variable in the model
(iv) There is no lag independent variable in the model
A. (i), (ii) and (iii) B. (i) and (ii) C. All of those D. (i), (ii) and (iv)

Question 25: What is NOT a cause of autocorrelation problem?

A. Lags B. Coweb phenomenon C. Exotic observations D. All of these above

Question 26: Assume the model has a first degree autocorrelation with an autocorrelation coefficient
ρ = 0,4. When fixed, Xt should be transformed into:
A. 0,4Xt B. 0,6Xt C. Xt – 0,4Xt-1 D. 0,6Xt – 0,4Xt-1

Question 27: Given the result of Breusch-Godfrey test. With the confidence level of 95%, does the
model have autocorrelation?
https://www.facebook.com/thanh.d.0715 19
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

A. No B. Yes C. Can’t conclude

Question 28: If the d-statistic of Durbin-Watson test approaches 0, the value of primary correlation
coefficient:
A. Approaches 0 B. Approaches 1 C. Approaches 1 or -1 D. Approaches -1

Question 29: If a model contains autocorrelation, this error:

A. Makes the OLS estimators biased C. Makes OLS estimation impossible
B. Makes the variance of OLS estimators not the smallest D. Both A and B are correct

Question 30: Secondary correlation random variance means that random variance ut at present is a
linear function of:
A. ut-1 B. ut-2 C. u2t-2 D. ut-1 and ut-2

Question 31: Given Durbin-Watson test: dL = 1,221; dU = 1,42; d= 1,928. Does the model have 1st
level autocorrelation?
A. Yes B. No C. Can’t conclude

Question 32: Use the Durbin - Watson test for a model with an intercept coefficient and 3 variables
X with 100 observations d = 3,6. Given dL = 1,613; dU = 1,736. Which of the following conclusions is
correct?
A. There is positive autocorrelation C. There is no autocorrelation
B. There is negative autocorrelation D. Cannot conclude

Questions about other violations

Question 33: Use Ramsey's Reset test when:
A. Model with multiple variables
B. Model with missing variables, with data on variable Z
C. Model with missing variables, with no data on variable Z
D. Statements A and C are correct

Question 34: In the RESET test with a significance level of 5%, we have the result Prob > F = 0,038,
concluding:
A. The model has omitted variables C. Not yet concluded
B. The model has no omitted variables

Question 35: RESET test belongs to:

A. chi-squared test B. F-test C. t-test D. None of these above

Question 36: Given the result of RESET test. With significance level α = 5%, does the model leave
out any variables?

https://www.facebook.com/thanh.d.0715 20
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

A. No B.Yes C. Can’t conclude

Question 37: There is the test result of the normal distribution of disturbance with Chi2 = 15,05 and
p-value respectively 0,0005. We can conclude that:
A. Disturbace has a normal distribution C. None of these above
B. Disturbace doesn’t have normal distribution

Question 38: How to fix the problem of disturbance without standard distribution?
A. Add independent variables C. Increase the sample size
B. Change the model D. None of these above

Question 39: Given the result of Skewness/Kurtoris test as following. With a significance level α =
5%, does disturbance have normal distribution?

A. No B. Yes C. Can’t conclude

Question 40: Which of the following is not a consequence of non-normally distributed noise?
A. Inefficient parameter estimation
B. Statistical values do not follow the corresponding rules
C. Statistical inference is unreliable when sample size is small
D. The regression line is horizontal for all values of the independent variables

https://www.facebook.com/thanh.d.0715 21
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

CHAPTER 6: REGRESSION WITH INDEPENDENT DUMMY VARIABLES

Question 1: What roles can dumy variables play in a model?

A. Only independent variables C. Can either be independent variables or dependent variables
B. Only dependent variables D. No right answer

Question 2: In a model that has a qualitative variable with the scope of m, we’ll use____ dummy
variables.
A. m - 1 B. m C. m +1 D. All of the above answers are wrong

Question 3: Model Yi = β1 + β2 Xi + β3 D + ui with D as a dummy variable shows the effect of dummy

variable on:
A. Slope B. Intercept C. Both A, B are right D. Both A, B are wrong

Question 4: Which model shows the effect of the qualitative variable on both the slope and the
intercept?
A. Yi = β1 + β2 Xi + β3 D + ui C. Yi = β1 + β2 Xi + β3 D + β4 Xi D + ui
B. Yi = β1 + β2 Xi + β3 Xi D + ui D. None of them.

Question 5: The regression of wage has these qualitative variables: sex (male, female); education level
(highschool, university); type of business (state-owned, private, joint venture). The amount of dummy
variables used are:
A. 3 B. 4 C. 5 D.7

The below information is used for Q6 and Q7

Using a model data set to estimate the dependence of wage (unit: dollar) on workers’ age, years of schooling
educ (unit: year), sex male (equal 1 if worker is male, equal 0 if is female) and interaction variable
experblack (calculated by exper*black, in which exper is years of experience(unit: year),black is skin color
of worker (equal 1 if dark skinned, equal 0 if light skinned)) we have the results:

Question 6: Which is the interpretation of the coefficient of variable male:

A. The wage of male worker is 568,2624 USD higher than that of female worker.
https://www.facebook.com/thanh.d.0715 22
 ECONOMETRICS Đinh Văn Thành - K55 KTQT
st st
Class: august | Facebook Time: 1 period, 1 semester, 2526

B. The wage of male worker is 568,2624 USD higher than that of female worker, given that other factors
unchanged.
C. The average wage of male worker is 568,2624 USD higher than that of female worker.
D. Average wage of male worker is 568,2624 USD higher than that of female worker, given that other
factors unchanged.

Question 7: A model with only dummy variables cannot test the error:
A. Multi-Collinearity C. Omitted-variable bias
B. Autocorrelation D. None of the above

Question 8: An individual's income in Budopia depends on his ethnicity and some other factors that
can be measured quantitatively. If there are 5 ethnic groups in Budopia, how many dummy variables
should be included in the regression equation to determine income in Budopia?
A. 6 B. 4 C. 5 D. 1

https://www.facebook.com/thanh.d.0715 23