Homework Quantitative

Problem I
A) Write out the estimated regression equation for the relationship between the
variables.
Y= 0.0136+ 0.7992X1 + 0.2280X2 -0.5796 X3
Family Spending = 0.0136+ 0.7992 * Income + 0.2280* family size -0.5796 * additions to
savings
B) What can you say about the strength of this relationship?
We have or Coefficient of determination = 0.946, what we can say about the strength of the
relationship is that 94.6% of the variability of Y can be explained by the independent
variables.
C) We will carry an F-test to see if Y is significantly related to the independent
variables.
State the null hypothesis:
Ho: X1=X2=X3=0 (No relationship)
H1: At least one of the variables does not equal 0. (There is a relationship)
F-stat: 64.28
F-Observed: 3.59, Stat>Observe, we reject Ho and thus conclude that the
relationship is significant.
D) Carry out a test to see if X3 and Y are significantly related. Use a .05 level of
significance.
State the null hypothesis:
Ho: X3=0 (Not significantly related)
H1: X30 (Significantly related)
Compute T-test:
T-Stat =Coefficient / St Error= -0.5796/0.920= -0.63
T-Observed : 2.201
T-lower: -2.201
Conclusion: We fail to reject Ho which means that there is no significant relationship between
X3, additions to savings, and Y, Family spending.

E) Why would a coefficient of determination very close to 1.0 be expected here?

In our example the coefficient of variation being close to 1 means that there are 1 or more
variable that highly affects the independent variable. In our problem we have a very high

Homework Quantitative

impact of two variables in spending which is Income followed by family size, that is why our
coefficient of determination is high.

Problem 2 :
a) Coefficient of X3= 18;
It means that a sales increase by 18000 dollars for the case when advertising is used.
b) E2 = 10 -4*3 + 7*10 +18*1 = 86
sales = 86000 dollars
Problem 3:
A)
Y= 4.0928 + 10.0230 X1 + 0.1020X2 - 4.4811X3
X1: College GPA, X2: Age of individuals, X3: Gender of the individual
B)
Y 4.0928 : when all independent variables are equal to 0, the yearly income will increase by
$4092.8.
X1 : each increase by 1 unit (GPA), holding the other variables constant, the yearly income
will increase by $10,023.
X2 : Each increase by 1 unit (age of the individuals), holding the other variables constant the
yearly income will increase by $102.
X3 : Depending on the gender, 0 for female and 1 for male, ceteris paribus; the yearly income
will decrease by $4,481.1.
C)
r2adj= 1
= 1-

)(

)(

= 0.9067

Interpretation: 90.67% of the variation in yearly income is explained by the variation of

college GPA, the age of the individual, and the gender of the individual.
D)
1. Variable 1: GPA
H0:
(No Linear Relationship)

Homework Quantitative

H1:

(Linear Relationship)

d.f= 6 and T= 2.4469

We reject H0.
We conclude that there is a linear relationship between yearly income and the GPA taking
into consideration the other independent variables.
Age of individuals :
H0:
(No Linear Relationship)
H1:

With d.f=

(Linear Relationship)

T= 2.4469
We fail to reject H0.

Therefore,

Conclusion : There is no linear relationship between yearly income and the age of the
individuals, taking into consideration the other independent variables.
2. Gender of individuals
H0:
(No Linear Relationship)
H1:

(Linear Relationship)

d.f= 6 , T= -2.4469
We reject H0.
Conclusion : There is a linear relationship between yearly income and the gender of the
individuals, taking into consideration the other independent variables.
b. F-test:
H0:

( there is no linear relationship.)

H1: at least one of the independent variables affects Y ( there is a linear relationship.)
Ftest=

d.f1= k= 3 and d.f2= 6 F= 6.6

Homework Quantitative

Ftest > 6.6 We reject H0.

Conclusion: there is a linear relationship between the independent variables and the
dependent variable: at least one of the independent variables is related to yearly income.

The model is adequate for this analysis because all the independent variables are significant
and the results of the F test also shows that the model is significant overall.
Problem 4:
1) Create dummy variables for time of the day:
Since time of the day has three levels: morning, afternoon, and evening, the number of
levels is k=3. Therefore, we have k-1=2 dummy variables with afternoon being the default
category.
X2 = Morning = {1 if true, 0 otherwise}
X3 = Evening = {1 if true, 0 otherwise}
2) Regression model of the percentage of successful calls on the other variables:
Using Excel, we have found the following results:
SUMMARY
OUTPUT
Regression Statistics
Multiple R
0,902353639
R Square
0,81424209
Adjusted R Square
0,779412482
Standard Error
2,521881528
Observations
20
ANOVA
df
Regression
Residual
Total

Intercept
Experience
(months)

3
16
19

Coefficients
11,12694712
0,442777496

Significance
SS
MS
F
F
446,0418169 148,6806 23,37787 4,32427E-06
101,7581831 6,359886
547,8
Standard
Error
1,369141847

t Stat
P-value Lower 95% Upper 95%
8,12695 4,52E-07 8,224496062 14,0293981

0,09500362 4,660638 0,000261 0,241378819 0,64417617

Homework Quantitative

Morning

1,789959694

Evening

-7,730277798

1,418728226 1,261665 0,225153 1,217609791 4,79752917

1,40306177 -5,50958 4,76E-05 10,70463588 4,75591971

Thus, the regression equation is:

Interpretation of r: 81.42% of the variability of Y is explained by the interpreted

variables.
Interpretation of adjusted r: 77.94% of the variability of Y is explained by the
interpreted variables taking into consideration the sample size and number of
independent variables.
Interpretation of b0: If all the independent variables are equal to 0, Y is equal to
11.12694712% of successful calls.
Interpretation of b1: If X1 increases by 1 month, Y will increase by 0.442777496
months.
Interpretation of b2: If X2 is located in the front and X1 and X3 are kept constant, Y
will increase by 1.789959694.
Interpretation of b3: If X3 is located in the front and X1 and X2 are kept constant, Y
will decrease by 7.730277798.
For the variable (experience):
H0: 1=0 (there is no relationship between successful calls and experience)
H1: 10 (there is relationship between successful calls and experience)
=

=4.78

The critical values:

We have df (degrees of freedom) = 16, with /2 = 0.025
Then t = 2.1199

> 2.1199 then we reject H0. So, there is an evidence of the relationship between
successful calls and experience.
For the variable (morning):
H0: 2=0 (there is no relationship between successful calls and morning calls)
H1: 20 (there is relationship between successful calls and morning calls)
=

= 1.2623

Homework Quantitative

-2.1199
< 2.1199 then we fail to reject H0. So, there is no evidence of the
relationship between successful calls and morning calls.

For the variable (evening):

H0: 3=0 (there is no relationship between successful calls and evening calls)
H1: 30 (there is relationship between successful calls and evening calls)
=

= -5.50

< -2.1199 then we reject H0. So, there is an evidence of the relationship
between successful calls and evening calls.

3) Residual analysis:
H0: There is no linear relationship between the percentage of successful calls
and both experience and time of the day.
H1: There is a linear relationship between the percentage of successful calls and
both experience and time of the day.

Experience (months)
Residual Plot
Residuals

10
5
0
-5

10

15

Experience (months)

20

25

Homework Quantitative

Morning Residual Plot

Residuals

10
5
0
0

0.2

0.4

-5

0.6

0.8

1.2

Morning

Evening Residual Plot

6
Residuals

4
2
0

-2 0

0.5

1.5

Evening

-4

From these graphs, no plot has a pattern. Therefore, we conclude that the linear model is
appropriate.
Problem 5:
a. Unweighted aggregate price index:
( )

( )

( )

b. Laspeyres index:

( )

( )

( )

( )

( )

c. Paasche index:

( )

( )

( )

( )

( )

d. Fisher index:

( )

( )

e. Compare the results of (a), (b), (c), (d), and (e) and indicate the index which is
rational and why:

Homework Quantitative

According to the law of demand, the quantity demanded increases if the price
decreases and vice versa. Therefore, we need an estimation of that price. In fact, if we
compare the results, we can see that Fisher index is between Paasche and Laspeyres
indices. The reason behind that is that Paasche index tends to over-estimate the level
of the price, while Laspeyres index tends to under-estimate it. This is why we Fisher
index, which takes geometric means of the other two indices, to help overcome this
problem. Therefore, the rational index is Fisher index.
f. What is the shortcoming of the indexes computed in the previous questions:
The shortcoming of Laspeyres lies in the over-stating inflation towards which it is biased,
while Paasche is an understating inflation index because we do not have a fixed standard
quality. On the other hand, Fisher index tends to solve this biasing issue, but does not help
determine the link between the price change and inflation and how much is attributed to
that changing quality since the base year doesnt have a fixed price.

Problem 6 :

Years
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988

Agricultural producer price, wheat ($

per metric ton)
59,28
72,33
79,04
76,48
84,58
90,5
96,9
121,99
142,57
213,45
186,68
260,03
257,52
309,04
383,83
344,33
216,83
254,52
244,03
248,45
243,84
259,23
258,86

MA(3)
70,2166667
75,95
80,0333333
83,8533333
90,66
103,13
120,486667
159,336667
180,9
220,053333
234,743333
275,53
316,796667
345,733333
314,996667
271,893333
238,46
249
245,44
250,506667
253,976667
255,556667

Exponential
smoothing

code
59,28
61,89
73,672
78,528
78,1
85,764
91,78
101,918
126,106
156,746
208,096
201,35
259,528
267,824
323,998
375,93
318,83
224,368
252,422
244,914
247,528
246,918
259,156

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

Homework Quantitative

248,58
276,62
269,91
316,24
293,59
259,71
340,74
266,18
263,46
255,09

1989
1990
1991
1992
1993
1994
1995
1996
1997
1998

261,353333
265,036667
287,59
293,246667
289,846667
298,013333
288,876667
290,126667
261,576667

256,804
254,188
275,278
279,176
311,71
286,814
275,916
325,828
265,636
261,786

1)
450
400
350
300
250
Series1

200
150
100
50
0
1960

2) See excel sheet.

3)See Excel sheet.
4)

1970

1980

1990

2000

24
25
26
27
28
29
30
31
32
33

Homework Quantitative

Variable X 1 Trend line

450
400
350
300
Y

250
200

150

Prvisions pour Y

100
50
0
1

9 11 13 15 17 19 21 23 25 27 29 31 33
Variable X 1

Y=11.57X1
Y : Wheat Price
X1 : years code
5) Plot Y and Yhat on the same graph
450
400
350
300

Agricultural producer
price, wheat ($ per
metric ton)

250
200

Exponential

150
100
50
0
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33

6) Graphically is the model selected fits well to your data?

The model selected does not perfectly fit to our data.