Roger Gallagher owns a used car lot that deals solely used Corvettes.

He wants to develop a
regression model to help predict the price he can expect to receive for the cars he owns. He
collected the data found in the file regression data describing the mileage, model year, presence of
a T-top, and selling price of a number of cars ha has sold in recent months. Let Y present the selling
price X1, the mileage X2, the model year, X3, the presence (or absence) of a T-top.

a. If Roger wants to use a simple linear regression function to estimate the selling price of a car,
which X variable do you recommend he use?

After running a simple linear regression to each variable, the results are shown as:

i) I.V. – Selling Price, Y ii) I.V. – Selling Price, Y

D.V. – Mileage, X1 D.V. – Model Year, X2

m
er as
co
eH w
o.
rs e
ou urc
o
aC s

2
r 0.8444
vi y re

iii) I.V. – Selling Price, Y

D.V. – Presence of a T-top, X3
ed d
ar stu
is
Th
sh

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 The mileage, X1 data has the highest R2 value 0.8444. Therefore, we would recommend using
the mileage data to estimate the selling prices, if Roger Gallagher would want to use simple
regression analysis.

b. Determine the parameter estimates for the regression function represented

by: Ŷ = b0 + b1Xi + b2X2i

What is the estimated regression function?

Ŷ = -206653 – 155.64X1 + 118.79X2

R2 0.8485
Adjusted R 2 0.8182
Standard Error 2775.5872
Y Intercept -206653.0407

m
er as
Signifcance Tests of Slopes
slope (b ) SE (b ) t-stat p-value Confidence Interval 95%

co
Mileage -155.64715 51.3435 -3.0315 0.0126 -270.0476 -41.2467

eH w
Year 118.78938 229.2148 0.5182 0.6156 -391.9329 629.5117

o.
rs e
ou urc
Does X2 help to explain the selling price of the cars if X1 is also in the model? What might be the
reason for this?
o
aC s

In this case, X2 does not help to explain the selling price because the mileage and year have too
vi y re

much collinearity showing -0.88.

Correlation Matrix: Collinearity Check

Mileage Year
ed d

Mileage -0.884024639
ar stu

Year -0.884024639
is

c. Set up a binary variable (X3i) to indicate whether or not each car in the sample was a T-top.
Th

Determine the parameter estimates for the regression function represented by:

Ŷ = b0 + b1Xi + b3X3i
sh

Ŷ = 28794 – 177.48X1 + 2886.84X3

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 R2 0.8973
Adjusted R 2 0.8768
Standard Error 2285.2024
Y Intercept 28793.9191

Signifcance Tests of Slopes

slope (b ) SE (b ) t-stat p-value Confidence Interval 95%
Mileage -177.47620 19.7738 -8.9753 0.0000 -221.5351 -133.4173
T-Top 2886.84351 1272.2759 2.2690 0.0466 52.0362 5721.6508

Does X3 help to explain the selling price of the cars if X1 is also in the model? Explain.

Correlation Matrix: Collinearity Check

Mileage T-Top
Mileage -0.037745111

m
T-Top -0.037745111

er as
co
Yes. When the T-top sample was added to mileage data for multiple regression analysis, because

eH w
the two data has significantly less collinearity, R2 is 0.877, the highest result among all other

o.
regression analysis, and p-value shows statistical significance.
rs e
ou urc
d. According to the previous model, on average, how much does a T-top add to the value of a
car?
o

On average, T-Top adds to approximately $2886.84.

aC s

Signifcance Tests of Slopes

vi y re

slope (b ) SE (b )
Mileage -177.47620 19.7738
T-Top 2886.84351 1272.2759
ed d

Determine the parameter estimates for the regression function represented

ar stu

e.

by: Ŷ = b0 + b1Xi + b2X2i + b3X3i

is

Ŷ = -299860 – 144.78X1 – 164.77X2 + 3001.55X3

Th
sh

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 R2 0.9050
Adjusted R 2 0.8734
Standard Error 2316.2058
Y Intercept -299860.3004

Signifcance Tests of Slopes

slope (b ) SE (b ) t-stat p-value Confidence Interval 95%
Mileage -144.78222 43.1020 -3.3591 0.0084 -242.2857 -47.2787
Year 164.76527 192.3060 0.8568 0.4138 -270.2611 599.7917
T-Top 3001.55003 1296.4679 2.3152 0.0458 68.7358 5934.3642

f. Of all the regression functions considered here, which do you recommend Roger use?

The multiple regression functions with mileage and data has the highest R2, therefore, the
regression function “Ŷ = b0 + b1Xi + b3X3i , Ŷ = 28794 – 177.48X1 + 2886.84X3” should be
recommended.

m
er as
co
eH w
o.
An accounting firm that specializes in auditing mining companies collected the regression data
rs e
file describing the long-term assets and long- term debt of its 12 clients.
ou urc
g. Prepare a scatter plot of the data. Does there appear to be a linear relationship between
these variables?
o
aC s
vi y re
ed d
ar stu
is
Th
sh

Yes. Long term assets and long-term debt have a relatively strong, positive linear relationship.

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 h. Develop a simple linear regression model that can be used to predict long-term debt
from long-term assets. What is the estimated regression equation?
Regression Equation: Predicted Y = 0.8121X-0.784

Slope 0.8121
Intercept -7.84

i. Interpret the value of � 2?

2 0.8497
r
The long term assets and the long term debt regression model has 85% of variability.
Approximately 85% of the variability of a mining company’s long-term debt can be explained by
its long-term assets.

m
j. Suppose that the accounting firm has a client with total assets of $50,000,000. Construct an

er as
approximate 95% confidence interval for the amount of long-term debt the firm expect this client

co
eH w
to have.

o.
Prediction Interval for Single Observation
1-a if X
rs e y-hat ± t * SE = Interval
ou urc
95% 50.00 32.77 ± 9.940 = 22.826 42.706

The expected confidence intervals would be between $22.8 million to $42.7 million. As seen
o

above, a single client with assets of $50,000,000 would be expected to have long-term debt of
$22.83 million to $42.71 million. The average client with assets of $50 million would be
aC s
vi y re

expected to have long-term debt of $29.99 million to $35.54 million.

ed d

Chapter 11 Time Series Forecasting

ar stu

5. Compute the two-period and four-period moving average predictions for the data set.
is
Th

a. Prepare a line graph comparing the moving average predictions against the original data.
sh

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 m
er as
co
eH w
o.
rs e
ou urc
o

b. Do the moving averages tend to overestimate or underestimate the actual data? Why?
aC s
vi y re

The moving averages tend to underestimate the actual data. The moving averages look backward
in time. When there is a clear upward trend, as shown above, there is a lag which causes the
moving averages to underestimate the actual data. The moving averages model is more
appropriate for stationary data.
ed d
ar stu
is
Th
sh

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 c. Compute forecasts for the next 2 years using the two-period and four-period moving average
techniques.
MOVING AVERAGES

Time Data 2-Period 4-Period

1 283
2 288
3 336 285.5
4 388 312.0
5 406 362.0 323.8
6 412 397.0 354.5
7 416 409.0 385.5
8 435 414.0 405.5
9 428 425.5 417.3
10 435 431.5 422.8

m
er as
11 462 431.5 428.5

co
12 452 448.5 440.0

eH w
13 474 457.0 444.3
14 476 463.0 455.8

o.
rs e
15 497 475.0 466.0
ou urc
16 487 486.5 474.8
17 523 492.0 483.5
18 528 505.0 495.8
o

19 532 525.5 508.8

aC s

20 552 530.0 517.5

vi y re

21 542.0 533.8
22 542.0 533.8
ed d

8. Create an exponential smoothing model that minimizes the MSE for the data set. User Solver
ar stu

to determine the optimal value of α.

a.What is the optimal value of α.

is

α=1
Th

= 1
sh

MSE = 498.05

n= 20

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 b. Prepare a line graph comparing the exponential smoothing predictions against the
original data.

m
er as
co
eH w
o.
rs e
ou urc
o

c. What are the forecasts for the next 2 years using the technique?
aC s
vi y re

21 -- 552.00
22 -- 552.00
ed d

9. Use regression analysis to fit a linear trend model to the data set.
ar stu

a. What is the estimated regression functions?

Equation: y=12.2286x+312.10
is

b. Interpret the � 2 value for your model.

Th

0.9194
r
2
sh

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:58 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
 c. Prepare line graph comparing the linear trend predictions against the original data.

m
er as
co
eH w
o.
d. What are the forecasts for the next 2 years using this technique?
rs e
ou urc
if x = y-hat =
21 568.90
o

if x = y-hat =
22 581.13
aC s
vi y re

e. Fit a quadratic trend model to these data. What is the estimated regression function?

Equation: Y= -0.39907�2 + 20.6895� + 281.3719

ed d

f. Compare the adjusted-�2 value of this model to that of the linear trend model. What is
ar stu

implied by this comparison?

By comparing multiple regression results from the simple regression method, the
is

adjusted �2 has increased to 0.9388 and MSE has decreased to 348. The multiple
Th

regression was able to reduce the forecasting errors.

2 0.9453
R
sh

Adjusted R2 0.9388

This study source was downloaded by 100000805625772 from CourseHero.com on 04-07-2021 21:35:59 GMT -05:00

https://www.coursehero.com/file/53380264/Text-problems-solveddocx/
Powered by TCPDF (www.tcpdf.org)