ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 1 of 20 pages

CONFIDENTIAL

QUEEN'S UNIVERSITY AT KINGSTON

Department of Economics

CONFIDENTIAL
turn in exam
question paper

ECONOMICS 351* - Fall Term 2008

Introductory Econometrics

Fall Term 2008 MID-TERM EXAM: ANSWERS M.G. Abbott

DATE: Monday October 27, 2008.

TIME: 80 minutes; 1:00 p.m. - 2:20 p.m.

INSTRUCTIONS: The exam consists of FOUR (4) questions. Students are required to
answer ALL FOUR (4) questions.
Answer all questions in the exam booklets provided. Be sure your student
number is printed clearly on the front of all exam booklets used. Your
name is optional.
Do not write answers to questions on the front page of the first exam
booklet.
Please label clearly each of your answers in the exam booklets with the
appropriate number and letter.
Please write legibly.
Tables of percentage points of the t-distribution and F-distribution are
given on the last two pages of the exam.
This exam is CONFIDENTIAL. This question paper must be submitted
in its entirety with your answer booklet(s); otherwise your exam will not
be marked.

MARKING: The marks for each question are indicated in parentheses immediately
above each question. Total marks for the exam equal 100.

GOOD LUCK!

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 1 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 2 of 20 pages
CONFIDENTIAL

QUESTIONS: Answer ALL FOUR questions.

All questions pertain to the simple (two-variable) linear regression model for which the
population regression equation can be written in conventional notation as:

Yi = β 0 + β1X i + u i (1)

where Yi and Xi are observable variables, β 0 and β 1 are unknown (constant) regression
coefficients, and ui is an unobservable random error term. The Ordinary Least Squares (OLS)
sample regression equation corresponding to regression equation (1) is

Yi = βˆ 0 + βˆ 1X i + û i (i = 1, ..., N) (2)

where β̂ 0 is the OLS estimator of the intercept coefficient β 0 , β̂1 is the OLS estimator of the
slope coefficient β 1 , u$ i is the OLS residual for the i-th sample observation, and N is sample size
(the number of observations in the sample).

(15 marks)
1. Show that the OLS slope coefficient estimator β̂1 is a linear function of the Yi sample
values. Stating explicitly all required assumptions, prove that the OLS slope coefficient
estimator β̂1 is an unbiased estimator of the slope coefficient β1 .

ANSWER to Question 1:

(5 marks)
• Show that the OLS slope coefficient estimator β̂1 is a linear function of the Yi sample
values.

∑ xy ∑ x (Y − Y ) ∑ xY Y ∑i x i
βˆ 1 = i i 2 i = i i i 2 = i i2i −
∑i x i ∑i x i ∑i x i ∑ i x i2
∑ i x i Yi
= because ∑ i x i = 0 (5 marks)
∑ i x i2
xi
= Σ i k i Yi where k i ≡ .
Σ i x i2

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 2 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 3 of 20 pages
CONFIDENTIAL

ANSWER to Question 1 -- continued:

(10 marks)
• Stating explicitly all required assumptions, prove that the OLS slope coefficient estimator
β̂1 is an unbiased estimator of the slope coefficient β 1 .

(1) Substitute for Yi the expression Yi = β 0 + β1X i + u i from the population regression
equation (or PRE). (5 marks)

βˆ 1 = ∑ i k i Yi
= ∑ i k i (β 0 + β1X i + u i ) since Yi = β 0 + β1X i + u i by assumption (A1)
= ∑ i (β 0 k i + β1k i X i + k i u i )
= β 0 ∑ i k i + β1 ∑ i k i X i +∑ i k i u i
= β1 + ∑ i k i u i , since ∑ i k i = 0 and ∑ i k i X i = 1

(2) Now take expectations of the above expression for β̂1 conditional on the value Xi of
X: (5 marks)

E(βˆ 1 ) = E(β1 ) + E[∑ i k i u i ]

= β1 + ∑ i k i E(u i X i ) since β1 is a constant and the k i are nonstochastic
= β1 + ∑ i k i 0 since E(u i X i ) = E(u i ) = 0 by assumption (A2)
= β1

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 3 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 4 of 20 pages
CONFIDENTIAL

(15 marks)
2. Give a general definition of the F-distribution. Starting from this definition, derive the F-
statistic for the OLS slope coefficient estimator β̂1 . State all assumptions required for the
derivation.

ANSWER to Question 2:
(3 marks)
• General definition of the F-distribution: Consider the two random variables V1 and V2 such
that

(1) V1 ~ χ2[m1]
(2) V2 ~ χ2[m2]
(3) V1 and V2 are independent.

Then the random variable

V1 m1
F = ~ F[m1, m2]
V2 m 2

where F[m1, m2] denotes the F-distribution with m1 numerator degrees of freedom and m2
denominator degrees of freedom.

(2 marks)
• Error Normality Assumption: The random error term ui is normally distributed with mean 0
and variance σ 2 ; that is,

[ ]
u i X i ~ N 0, σ 2 for all i OR u i is iid as N 0, σ 2[ ]
• Use three implications of the error normality assumption to derive F(βˆ 1 ) :

[ ]
1. βˆ 1 ~ N β1 , Var(βˆ 1 ) where Var (βˆ 1 ) = N
σ2
= N
σ2
(1 mark)
∑ x i2 ∑ (Xi − X ) 2

i =1 i =1

Implications of normality of β̂1 :

βˆ 1 − β1
Z(βˆ 1 ) = ~ N[0, 1]
se(βˆ 1 )

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 4 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 5 of 20 pages
CONFIDENTIAL

ANSWER to Question 2 -- continued:

(βˆ 1 − β1 ) 2 (βˆ 1 − β1 ) 2 (βˆ 1 − β1 ) 2 ∑i x i2

[ Z(β1 )] =
ˆ 2
= 2 = ~ χ 2 [1] (1) (1 mark)
Var (β1 )
ˆ σ ∑i x i 2
σ 2

( N − 2) σˆ 2 σˆ 2 χ 2 [ N − 2]
2. ~ χ 2 [ N − 2] ⇒ ~ (2) (1 mark)
σ 2
σ2 ( N − 2)

3. The estimators β̂1 and σ̂ 2 are statistically independent

OR (1 mark)
The statistics Z(βˆ 1 ) and σˆ 2 σ 2 are statistically independent.

♦ The F-statistic for β̂1 . The F-statistic for β̂1 is therefore the ratio of (1) to (2): (6 marks)

(
Z(βˆ 1 )
F(β1 ) = 2 2
ˆ
2
)
σˆ σ

=
(βˆ 1 )(
2
)
− β1 ∑ i x i2 σ 2
σˆ 2 σ 2

=
(βˆ 1 )(
2
− β1 ∑ i x i2 )
σˆ 2

=
(βˆ − β1
1 )
2

σˆ ∑ i x i2
2

=
(βˆ 1 − β1 )
2

since σˆ 2 ∑ i x i2 = Vâr (βˆ 1 ).

Vâr (βˆ 1 )

‰ Result: The F-statistic for β̂1 takes the form

F(βˆ 1 ) =
(βˆ − β ) = (βˆ − β )2 2

~ F[1, N − 2] .
σˆ (∑ x )
1 1 1 1
2
Vâr (βˆ )
i
2
i 1

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 5 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 6 of 20 pages
CONFIDENTIAL

(32 marks)
3. A researcher is using data for a sample of 3240 female employees 25 years of age and over to
investigate the relationship between employees’ hourly wage rates Yi (measured in dollars
per hour) and their age Xi (measured in years). The population regression equation takes the
form of equation (1): Yi = β 0 + β1X i + u i . Preliminary analysis of the sample data produces
the following sample information:
N N N
N = 3240 ∑ y i2 = 78434.97 ∑ x i2 = 25526.17 ∑ x i yi = 3666.426
i =1 i =1 i =1
N N N
∑ Yi = 34379.16 ∑ X i = 96143.00 ∑ Yi2 =
i =1
443227.1
i =1 i =1
N N N
∑ X i2 = 2878451.0 ∑ X i Yi = 1023825.0 ∑ û i2 = 77908.35
i =1 i =1 i =1

where x i ≡ X i − X and y i ≡ Yi − Y for i = 1, ..., N. Use the above sample information to

answer all the following questions. Show explicitly all formulas and calculations.

(10 marks)
(a) Use the above information to compute OLS estimates of the intercept coefficient β 0
and the slope coefficient β 1 .

∑ x y 3666.426
• βˆ 1 = i i 2 i = = 0.143634 = 0.1436 (5 marks)
∑i x i 25526.17

• βˆ 0 = Y − βˆ 1X
N N
∑ Yi 34379.16
∑ Xi 96143.00
i =1 i =1
Y = = = 10.61085 and X = = = 29.67377
N 3240 N 3240

Therefore

βˆ 0 = Y − βˆ 1X = 10.610852 − (0.143634)(29.67377) = 10.610852 − 4.262162 = 6.34869

= 6.349
(5 marks)

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 6 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 7 of 20 pages
CONFIDENTIAL

(4 marks)
(b) Interpret the slope coefficient estimate you calculated in part (a) – i.e., explain in words
what the numeric value you calculated for β̂ 1 means.

Note: β̂ 1 = 0.1436. Yi is measured in dollars per hour, and Xi is measured in years.

The estimate 0.143634 of β1 means that a 1-year increase (decrease) in age Xi is
associated on average with an increase (decrease) in hourly wage rate equal to 0.1436
dollars per hour, or 14.36 cents per hour.

(4 marks)
(c) Calculate an estimate of σ 2 , the error variance.

N
RSS = ∑ û i2 = 77908.35; N − 2 = 3240 − 2 = 3238
i =1

RSS
∑ û i2
77908.35 77908.35
σˆ 2 = = i=1 = = = 24.060639 = 24.0606 (4 marks)
N−2 N−2 3240 − 2 3238

(6 marks)
(d) Compute the value of R 2 , the coefficient of determination for the estimated OLS
sample regression equation. Briefly explain what the value you have calculated for R 2
means.

(4 marks)
N N

ESS
∑ yi2 − ∑ û i2 78434.97 − 77908.35 526.62
i =1 i =1
R2 = = N
= = = 0.006714
TSS 78434.97 78434.97
∑ yi2
i =1

OR
N

RSS
∑ û i2 77908.35
i =1
R2 = 1 − =1 − N
=1− = 1 − 0.993286 = 0.006714
TSS 78434.97
∑ yi2
i =1

(2 marks)
Interpretation of R2 = 0.006714: The value of 0.006714 indicates that 0.67 percent of
the total sample (or observed) variation in Yi (hourly wage rates) is attributable to,
or explained by, the regressor Xi (age) or the sample regression function.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 7 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 8 of 20 pages
CONFIDENTIAL

(8 marks)
(e) Calculate the sample value of the t-statistic for testing the null hypothesis H0: β1 = 0
against the alternative hypothesis H1: β1 ≠ 0 . (Note: You are not required to obtain or
state the inference of this test.)

(βˆ − β1 )
• t-statistic for β̂ 1 is t (βˆ 1 ) = 1 (1) (2 marks)
sê(βˆ 1 )

σˆ 2 24.060639
• From part (a), β̂ 1 = 0.143634; Vâr (βˆ 1 ) = N
= = 0.000942587.
25526.17
∑ x i2
i =1

• The estimated standard error of β̂ 1 is:

sê(βˆ 1 ) = Vâr (βˆ 1 ) = 0.000942587 = 0.03070158 = 0.0307016

• Calculate the sample value of the t-statistic (1) under H0: set β1 = 0, β̂ 1 = 0.143634
and sê(βˆ ) = 0.0307016 in (1) above.
1

(βˆ − β1 ) 0.143634 − 0 0.143634

t 0 (βˆ 1 ) = 1 = = = 4.678388 = 4.68 (6 marks)
sê(β1 )
ˆ 0.0307016 0.0307016

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 8 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 9 of 20 pages
CONFIDENTIAL

(38 marks)
4. You have been commissioned to investigate the relationship between the birth weights of
newborn females and the number of prenatal visits to a physician or midwife that their
mothers made during pregnancy. The dependent variable is bwghti, the birth weight of the i-
th newborn female, measured in grams. The explanatory variable is pnvisitsi, the number of
prenatal visits of the i-th newborn’s mother during pregnancy, measured in number of visits.
The model you propose to estimate is given by the population regression equation

bwght i = β0 + β1 pnvisitsi + u i .

Your research assistant has used 857 sample observations on bwghti and pnvisitsi to estimate
the following OLS sample regression equation, where the figures in parentheses below the
coefficient estimates are the estimated standard errors of the coefficient estimates:

bwght i = 3199.02 + 14.1219 pnvisitsi + û i (i = 1, ..., N) N = 857 (3)

(65.6909) (5.36347) ← (standard errors)

(8 marks)
(a) Perform a test of the null hypothesis H0: β1 = 0 against the alternative hypothesis H1:
β1 ≠ 0 at the 1% significance level (i.e., for significance level α = 0.01). Show how you
calculated the test statistic. State the decision rule you use, and the inference you would
draw from the test. What would you conclude from the results of the test?

H0: β1 = 0
H1: β1 ≠ 0 a two-sided alternative hypothesis ⇒ a two-tailed test

βˆ 1 − β1 (βˆ 1 − β1 ) 2
• Test statistic is either t (β1 ) =
ˆ ~ t[ N − 2] or F(β1 ) =
ˆ ~ F[1, N − 2] .
sê(βˆ 1 ) Vâr (βˆ 1 )

• β̂1 = 14.1219; sê(βˆ 1 ) = 5.36347; (

Vâr (βˆ 1 ) = sê(βˆ 1 ) )
2
= 28.76681

• Calculate the sample value of either the t-statistic or the F-statistic under H0:
set β1 = 0, β̂1 = 14.1219, sê(βˆ 1 ) = 5.36347, and Vâr (βˆ 1 ) = 28.76681.

βˆ − β 14.1219 − 0.0 14.1219

t 0 (βˆ 1 ) = 1 1 = = = 2.63298 = 2.63
sê(βˆ 1 ) 5.36347 5.36347
or (3 marks)

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 9 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 10 of 20 pages
CONFIDENTIAL

ANSWER to Question 4(a) -- continued:

(βˆ 1 − β1 ) 2 (14.1219 − 0.0) 2 199.42806

F0 (β1 ) =
ˆ = = = 6.9326 = 6.93
Vâr (βˆ 1 ) 28.76681 28.76681

• Null distribution of t 0 (βˆ 1 ) is t[N − 2] = t[857 − 2] = t[855]

• Null distribution of F0 (βˆ 1 ) is F[1, N − 2] = F[1, 857 − 2] = F[1, 855]

Decision Rule: At significance level α, (2 marks)

• reject H0 if F0 (βˆ 1 ) > Fα [1, 855] or t 0 (βˆ 1 ) > t α 2 [855] ,

i.e., if either (1) t 0 (βˆ 1 ) > t α 2 [855] or (2) t 0 (βˆ 1 ) < − t α 2 [855] ;

• retain H0 if F0 (βˆ 1 ) ≤ Fα [1, 855] or t 0 (βˆ 1 ) ≤ t α 2 [855] ,

i.e., if − t α 2 [855] ≤ t 0 (βˆ 1 ) ≤ t α 2 [855] .

Critical values of t[855]-distribution or F[1, 855]-distribution: in t-table, use df = ∞.

• two-tailed 1 percent critical value = t α 2 [855] = t0.005[855] = 2.5816 = 2.58 (1 mark)

= t0.005[∞] = 2.576 = 2.58

Critical values of F[1, 855]-distribution: in F-table, use denominator df = ∞.

• 1 percent critical value = Fα [1, 855] = F0.01[1, 855] = 6.6646 = 6.66

= F0.01[1, ∞] = 6.63

Inference: (1 mark)

♦ At 1 percent significance level, i.e., for α = 0.01,

t 0 (βˆ 1 ) = 2.63 > 2.58 = t0.005[855] ⇒ reject H0 vs. H1 at 1 percent level.

F0 (βˆ 1 ) = 6.93 > 6.63 = F0.01[1, 855] ⇒ reject H0 vs. H1 at 1 percent level.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 10 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 11 of 20 pages
CONFIDENTIAL

ANSWER to Question 4(a) -- continued:

♦ Inference: At the 1% significance level, the null hypothesis H0: β1 = 0 is rejected in

favour of the alternative hypothesis H1: β1 ≠ 0.

Conclusion implied by test outcome: (1 mark)

Rejection of the null hypothesis β1 = 0 against the alternative hypothesis β1 ≠ 0 means
that the sample evidence favours the existence of a relationship between the number
of prenatal visits of mothers and the birth weights of their female newborns.

Question 4(a) – Alternative Answer -- uses confidence interval approach

• The two-sided (1 − α)-level, or 100(1 − α) percent, confidence interval for β1 is:

βˆ 1 − t α / 2 [ N − 2] sê(βˆ 1 ) ≤ β1 ≤ βˆ 1 + t α 2 [ N − 2] sê(βˆ 1 )
βˆ 1L ≤ β1 ≤ βˆ 1U

• Required results and intermediate calculations:

N − K = 857 − 2 = 855; β̂1 = 14.1219; sê(βˆ 1 ) = 5.36347

1 − α = 0.99 ⇒ α = 0.01 ⇒ α/2 = 0.005: t α 2 [ N − 2] = t0.005[855] = 2.576 (1 mark)

t α 2 [ N − 2] sê(βˆ 1 ) = t 0.005[855] sê(βˆ 1 ) = 2.576(5.36347) = 13.816299

• Lower 99% confidence limit for β1 is: (2 marks)

βˆ 1L = βˆ 1 − t α 2 [ N − 2] sê(βˆ 1 ) = βˆ 1 − t 0.005[855] sê(βˆ 1 )

= 14.1219 − 2.576(5.36347) = 14.1219 − 13.816299 = 0.305601 = 0.3056

= 14.1219 − 2.617(5.36347) = 14.1219 − 14.036201 = 0.0696086 = 0.08570
= 14.1219 − 2.58(5.36347) = 14.1219 − 13.837753 = 0.284147 = 0.2841

• Upper 99% confidence limit for β1 is: (2 marks)

βˆ 1U = βˆ 1 + t α 2 [ N − 2] sê(βˆ 1 ) = βˆ 1 + t 0.005[855] sê(βˆ 1 )

= 14.1219 + 2.576(5.36347) = 14.1219 + 13.816299 = 27.938199 = 27.9382

= 14.1219 + 2.617(5.36347) = 14.1219 + 14.036201 = 28.158101 = 28.1581
= 14.1219 + 2.58(5.36347) = 14.1219 + 13.837753 = 27.959653 = 27.9597

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 11 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 12 of 20 pages
CONFIDENTIAL

Question 4(a) – Alternative Answer (continued)

• Two-sided 99% confidence interval for β1 is therefore:

0.3056 ≤ β1 ≤ 27.9382 or 0.08570 ≤ β1 ≤ 28.1581 or 0.2841 ≤ β1 ≤ 27.9597

• Decision Rule: At significance level α, (1 mark)

• reject H0 if the hypothesized value b1 of β1 specified by H0 lies outside the two-sided

(1−α)-level confidence interval for β1, i.e., if either
(1) b1 < βˆ 1 − t α 2 [855] sê(βˆ 1 ) or (2) b1 > βˆ 1 + t α 2 [855] sê(βˆ 1 ) .

• retain H0 if the hypothesized value b1 of β1 specified by H0 lies inside the two-sided

(1−α)-level confidence interval for β1, i.e., if
βˆ 1 − t α 2 [855] sê(βˆ 1 ) ≤ b1 ≤ βˆ 1 + t α 2 [855] sê(βˆ 1 ) .

Inference: (1 mark)

♦ At 1 percent significance level, i.e., for α = 0.01,

b1 = 0 < 0.3056 = βˆ 1L = βˆ 1 − t 0.005[855] sê(βˆ 1 )

⇒ reject H0 vs. H1 at 1 percent level.

♦ Inference: At the 1% significance level, the null hypothesis H0: β1 = 0 is rejected in

favour of the alternative hypothesis H1: β1 ≠ 0.

Conclusion implied by test outcome: (1 mark)

Rejection of the null hypothesis β1 = 0 against the alternative hypothesis β1 ≠ 0 means
that the sample evidence favours the existence of a relationship between the number
of prenatal visits of mothers and the birth weights of their female newborns.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 12 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 13 of 20 pages
CONFIDENTIAL

(8 marks)
(b) Perform a test of the proposition that newborn females’ birth weights are positively
related to their mother’s number of prenatal visits during pregnancy, i.e., that an increase
in the number of prenatal visits increases the birth weights of newborn females. Use the
1 percent significance level (i.e., α = 0.01). State the null hypothesis H0 and the
alternative hypothesis H1. Show how you calculated the test statistic. State the decision
rule you use, and the inference you would draw from the test.

ANSWER to Question 4(b):

Null and Alternative Hypotheses:

H0: β1 = 0 (1 mark)
H1: β1 > 0 ⇒ a right-tailed t-test (1 mark)

βˆ 1 − β1
• Test statistic is t (βˆ 1 ) = ~ t[ N − 2] ; β̂1 = 14.1219; sê(βˆ 1 ) = 5.36347.
sê(βˆ )
1

• Calculate the sample value of the t-statistic under H0:

set β1 = 0, β̂1 = 14.1219, and sê(βˆ 1 ) = 5.36347.

βˆ − β 14.1219 − 0.0 14.1219

t 0 (βˆ 1 ) = 1 1 = = = 2.63298 = 2.63 (2 marks)
sê(βˆ 1 ) 5.36347 5.36347

• Null distribution of t 0 (βˆ 1 ) is t[N − 2] = t[857− 2] = t[855]

Decision Rule: At significance level α, (2 marks)

• reject H0 if t 0 (βˆ 1 ) > t α [855] ,

• retain H0 if t 0 (βˆ 1 ) ≤ t α [855] .

Critical value of t[855]-distribution: from t-table, use df = 120 or df = ∞ or any

number in between.

• right-tail 1 percent critical value = t0.01[855] = 2.331 = 2.33 (1 mark)

= t0.01[855] = 2.358 = 2.36
= t0.01[∞] = 2.326 = 2.33

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 13 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 14 of 20 pages
CONFIDENTIAL

ANSWER to Question 4(b) -- continued:

Inference: (1 mark)

♦ At 1 percent significance level, i.e., for α = 0.01,

t 0 (βˆ 1 ) = 2.63 > 2.33 = t0.01[855] ⇒ reject H0 vs. H1 at 1 percent level.

♦ Inference: At the 1% significance level, the null hypothesis H0: β1 = 0 is rejected

in favour of the alternative hypothesis H1: β1 > 0.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 14 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 15 of 20 pages
CONFIDENTIAL

(14 marks)
(c) Compute the two-sided 95% confidence interval for the intercept coefficient β0 . Use
this two-sided 95% confidence interval for β0 to test the hypothesis that the mean birth
weight of newborn females whose mothers made no prenatal visits to a physician or
midwife equals 3,000 grams. State the null hypothesis H0 and the alternative hypothesis
H1. State the decision rule you use, and the inference you would draw from the test.

ANSWER to Question 4(c):

• The two-sided (1 − α)-level, or 100(1 − α) percent, confidence interval for β1 is

computed as

βˆ 0 − t α / 2 [ N − 2] sê(βˆ 0 ) ≤ β0 ≤ βˆ 0 + t α 2 [ N − 2] sê(βˆ 0 ) (2 marks)

where

• βˆ 0 L = βˆ 0 − t α 2 [ N − 2] sê(βˆ 0 ) = the lower 100(1 − α)% confidence limit for β0

• βˆ 0 U = βˆ 0 + t α 2 [ N − 2] sê(βˆ 0 ) = the upper 100(1 − α)% confidence limit for β0
• t α 2 [ N − 2] = the α/2 critical value of the t-distribution with N−2 degrees of
freedom.

• Required results and intermediate calculations:

N − K = 857 − 2 = 855; β̂0 = 3199.02; sê(βˆ 0 ) = 65.6909

1 − α = 0.95 ⇒ α = 0.05 ⇒ α/2 = 0.025: t α 2 [ N − 2] = t0.025[855] = 1.963

= t0.025[855] = 1.980
= t0.025[∞] = 1.960

t α 2 [ N − 2] sê(βˆ 0 ) = t 0.025 [855] sê(βˆ 0 ) = 1.963(65.6909) = 128.95124

• Lower 95% confidence limit for β0 is: (3 marks)

βˆ 0 L = βˆ 0 − t α 2 [ N − 2] sê(βˆ 0 ) = βˆ 0 − t 0.025 [855] sê(βˆ 0 )

= 3199.02 − 1.963(65.6909) = 3199.02 − 128.95124 = 3070.069 = 3070.07

= 3199.02 − 1.980(65.6909) = 3199.02 − 130.06798 = 3068.952 = 3068.95
= 3199.02 − 1.960(65.6909) = 3199.02 − 128.75416 = 3070.266 = 3070.27

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 15 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 16 of 20 pages
CONFIDENTIAL

ANSWER to Question 4(c) -- continued:

• Upper 95% confidence limit for β0 is: (3 marks)

βˆ 0 U = βˆ 0 + t α 2 [ N − 2] sê(βˆ 0 ) = βˆ 0 + t 0.025 [855] sê(βˆ 0 )

= 3199.02 + 1.963(65.6909) = 3199.02 + 128.95124 = 3327.971 = 3327.97

= 3199.02 + 1.980(65.6909) = 3199.02 + 130.06798 = 3329.088 = 3329.09
= 3199.02 + 1.960(65.6909) = 3199.02 + 128.75416 = 3327.774 = 3327.77

• Result: The two-sided 95% confidence interval for β0 is: [3070.07, 3327.97]

3070.07 ≤ β0 ≤ 3327.97 or
3068.95 ≤ β0 ≤ 3329.09 or
3070.27 ≤ β0 ≤ 3327.77

• Null and Alternative Hypotheses: (2 marks)

H0: β0 = 3000
H1: β0 ≠ 3000 a two-sided alternative hypothesis ⇒ a two-tailed test

• Decision Rule: At significance level α, (2 marks)

• reject H0 if the hypothesized value b0 of β0 specified by H0 lies outside the two-sided

(1−α)-level confidence interval for β0, i.e., if either
(1) b 0 < βˆ 0 − t α 2 [855] sê(βˆ 0 ) or (2) b 0 > βˆ 0 + t α 2 [855] sê(βˆ 0 ) .

• retain H0 if the hypothesized value b0 of β0 specified by H0 lies inside the two-sided

(1−α)-level confidence interval for β0, i.e., if
βˆ 0 − t α 2 [855] sê(βˆ 0 ) ≤ b 0 ≤ βˆ 0 + t α 2 [855] sê(βˆ 0 ) .

Inference: (2 marks)

♦ At 5 percent significance level, i.e., for α = 0.05,

b 0 = 3000 < 3070.07 = βˆ 0 L = βˆ 0 − t 0.025 [855] sê(βˆ 0 )

⇒ reject H0 vs. H1 at 5 percent level.

♦ Inference: At the 5% significance level, the null hypothesis H0: β0 = 3000 is rejected
in favour of the alternative hypothesis H1: β0 ≠ 3000.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 16 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 17 of 20 pages
CONFIDENTIAL

(8 marks)
(d) Perform a test of the proposition that each additional prenatal visit made by the mother is
associated on average with an increase in their newborn females’ birth weight of less
than 25 grams. Use the 5 percent significance level (i.e., α = 0.05). State the null
hypothesis H0 and the alternative hypothesis H1. Show how you calculated the test
statistic. State the decision rule you use, and the inference you would draw from the test.

ANSWER to Question 4(d):

Null and Alternative Hypotheses:

H0: β1 = 25 (1 mark)
H1: β1 < 25 ⇒ a left-tailed t-test (1 mark)

βˆ − β
• Test statistic is t (βˆ 1 ) = 1 1 ~ t[ N − 2] ; β̂1 = 14.1219; sê(βˆ 1 ) = 5.36347.
sê(βˆ 1 )

• Calculate the sample value of the t-statistic under H0:

set β1 = 25, β̂1 = 14.1219, and sê(βˆ 1 ) = 5.36347.

βˆ − β 14.1219 − 25.0 − 10.8781

t 0 (βˆ 1 ) = 1 1 = = = −2.02818 = −2.03 (2 marks)
sê(βˆ 1 ) 5.36347 5.36347

• Null distribution of t 0 (βˆ 1 ) is t[N − 2] = t[857− 2] = t[855]

Decision Rule: At significance level α, (2 marks)

• reject H0 if t 0 (βˆ 1 ) < − t α [855] ,

• retain H0 if t 0 (βˆ 1 ) ≥ − t α [855] .

Critical value of t[855]-distribution: from t-table, use df = 120 or df = ∞ or any

number in between.

• left-tail 1 percent critical value = − t0.05[855] = −1.6466 = −1.65 (1 mark)

= − t0.05[855] = −1.658 = −1.66
= − t0.05[∞] = −1.645 = −1.65

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 17 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 18 of 20 pages
CONFIDENTIAL

ANSWER to Question 4(d) -- continued:

Inference: (1 mark)

♦ At 5 percent significance level, i.e., for α = 0.05,

t 0 (βˆ 1 ) = −2.03 < −1.65 = − t0.05[855] ⇒ reject H0 vs. H1 at 5 percent level.

♦ Inference: At the 5% significance level, the null hypothesis H0: β1 = 25 is rejected

in favour of the alternative hypothesis H1: β1 < 25.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 18 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 19 of 20 pages
CONFIDENTIAL

Percentage Points of the t-Distribution

Source: Damodar N. Gujarati, Basic Econometrics, Third Edition. New York: McGraw-Hill,
1995, p. 809.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 19 of 20 pages
ECONOMICS 351* ANSWERS: Mid-Term Exam -- Fall Term 2008 Page 20 of 20 pages
CONFIDENTIAL

Selected Upper Percentage Points of the F-Distribution

Source: Damodar N. Gujarati, Basic Econometrics, Third Edition. New York: McGraw-Hill,
1995, p. 814.

ECON 351* -- Fall Term 2008: Mid-Term Exam Answers Page 20 of 20 pages