EXCEL Output

The output obtained from EXCEL is shown below. You will observe that with some entries
have been wiped out. Fill-in these entries so you will know how the numbers in the EXCEL
output are connected.

Regression
Statistics

Multiple R 0.8908

R Square 0.7935

Adjusted R 0.7900
Square

Standard 3.0400
Error

Observations 60

ANOVA

df SS MS F Significance
F

Regression 1 2059.8551 2059.8551 226.73 0.0000

Residual 58 536,0144 9.0849

Total 59 2595.8695

Coefficients Standard t Stat P-value Lower 95% Upper 95%

Error

Intercept 4.7034 0.8518 5.5215 0.0000 2.9983 6.4085

Advertising 2.0481 0.1372 14.9295 0.0000 1.7685 2.3227

P a g e 1 | 21
Lower 99.0% Upper 99.0%

2.4347 6.9720

1.6828 2.4135

P a g e 2 | 21
Answer the following questions

1. What is the hypothesized regression model?

Profits= 0 + 1 Adv + 

2. What is the estimated least squares regression model?

Profits^ =

3. What is the interpretation of the slope coefficient?

4. Give a 99% confidence interval for the slope coefficient. Interpret the 99% confidence
interval? What does it tell you?

(1.6828; 2.4135)

5. Is the slope coefficient statistically significantly different from zero? How can you tell?

Yes because p-value is less than alpha

P a g e 3 | 21
6.. A supervisor claims that an incremental expenditure of $1000 in Advertising would result
in an average increase of at least $1500 in profits. Test this claim at a significance level of
1%. Set up the hypotheses and perform a 5-step hypothesis testing procedure.

Slope of the line (1)=1500/1000=1.5

TS= (2.0481-1.5)/0.1372

7. What is the numerical value of the correlation coefficient between the dependent variable
and the explanatory variable? What does it tell you?

P a g e 4 | 21
8. What percentage of the variation in the dependent variable (or Y) is explained by the
explanatory variable (or X)?

9. What is the F-value in the ANOVA table of the EXCEL output? What information does this
F-value provide you? Is this statistically significant at say a significance level of 1%?

10. What is your forecast for the average profits when the firm spends $35,000 on
Advertising?

Average profit (Y^)= 4.7 + 2.04*35 =

P a g e 5 | 21
 CASE
USED CAR PRICES

Consider the following data set. It contains data on the asking Price in $ and the odometer
reading (Miles) of forty randomly selected used 2007 Ford Explorer vehicles. Use EXCEL and
perform a comprehensive regression analysis and provide answers to the questions that
follow. EXCEL file: Data Set – Used Car Prices

P a g e 6 | 21
Miles Price Miles Price

37,880 $21,988 46,731 $19,389

37,152 $21,995 25,693 $20,990

41,675 $20,995 36,437 $22,973

30,329 $20,495 34,717 $19,495

47,332 $19,999 53,010 $18,490

43,441 $18,995 38,707 $19,890

57,779 $16,454 30,982 $18,290

69,973 $19,448 66,977 $22,995

37,626 $20,989 45,826 $18,996

20,591 $22,490 27,816 $23,922

40,781 $21,995 41,920 $18,991

39,403 $19,955 35,228 $20,995

55,786 $15,950 61,026 $17,995

64,367 $18,988 43,754 $18,961

50,829 $17,982 43,453 $16,997

31,353 $21,995 19,776 $22,995

41,450 $19,941 53,327 $18,646

P a g e 7 | 21
67,980 $21,999 39,801 $18,241

81,469 $18,999 60,977 $17,495

62,358 $17,950 56,538 $17,950

P a g e 8 | 21
EXCEL Output

SUMMARY OUTPUT

Regression Statistics

Multiple R 0.465470188

R-Square 0.216662496

Adjusted
R-Square 0.196048351

Standard
Error 1756.159627

Observations 40

ANOVA

df SS MS F Significance F

Regression 1 32415027.71 32415027.71 10.51038003 0.002472975

Residual 38 117195672.2 3084096.637

Total 39 149610699.9

Standard
Coefficients Error t Stat P-value Lower 95% Upper 95%

Intercept 22789.59866 946.0987697 24.08796987 1.27729E-24 20874.32183 24704.87549

Miles -0.064363258 0.019853122 -3.241971627 0.002472975 -0.104553802 -0.024172714

Lower 99.0% Upper 99.0%

20224.19735 25354.99997

-0.118196141 -0.010530374

P a g e 9 | 21
1. What is the estimated regression model?
Price= 22789.59866 – 0.06436x Miles

2. Find the correlation coefficient between the variable of interest and the X-variable.
Correl(price, miles)= - 0.46547= - Multiple R

3. What is the slope coefficient?

-0.064363258
Is the slope coefficient statistically significant (i.e., is it significantly different from zero)?
yes
How can you tell?
Look at the p-value of slope coefficient 0.0024< alpha 0.05

What is the interpretation of the slope coefficient?

For every mile the car has run, the price will decrease by 0.0643 dollars

4. Find a 95% confidence interval for the slope coefficient and interpret it.

-0.104553802, -0.024172714

5. What percentage of the variation in the dependent variable (or Y) is explained by the
explanatory variable (or X)?
21.66

6. Is the overall regression model significant at a level of α = 0.01? How can you tell?
Yes the significant F is 0.00247< 0.01

P a g e 10 | 21
7. (a) Use the regression model (in KSTAT) and predict the average price of Explorers with
36,000 miles on their odometers. Provide a 99% confidence interval.
(b) Use the regression model (in KSTAT) and predict the price of an Explorer with 36,000
miles on its odometer. Provide a 99% prediction interval.

Prediction, using most-recent regression

constant Miles

coefficients 22789.6 -0.06436

values for prediction 36000

predicted value of 20472.52

Price

standard error of prediction 1788.07

standard error of regression 1756.16

standard error of estimated mean 336.2984

confidence level 99.00%

t-statistic 2.7116

residual degr. freedom 38

confidence limits lower 15624.07 Commented [NT1]: PI

for prediction upper 25320.98

P a g e 11 | 21
confidence limits lower 19560.63 Commented [NT2]: CI

for estimated mean upper 21384.41

P a g e 12 | 21
 CASE
REFRIGERATOR PRICE AND ENERGY COST (1)

Joe Sears, manager in charge of a brand of refrigerators is reviewing the recent report
from the product development team. The report announced a breakthrough in energy
efficient refrigeration technology. The team believes that for an additional production cost
of $50 per unit, the consumer’s annual energy costs to run the refrigerator will drop by $20.
Joe must address the decision whether this new technology should be incorporated into the
upcoming model enhancement.

One key consideration is how much the customers would be willing to pay for a more
efficient refrigerator. To get an estimate of this, Joe orders a sampling study of the
relationship between the price of a refrigerator and the annual energy cost for using the
refrigerator. The data on Price (in $) and Energy Cost (in $/year) was collected from a
random sample of 41 popular refrigerator models and it is presented in the table below. See
Data Set – Refrigerator Price (1).

Your assignment is to carry out regression analyses to address the following issue:
What is the estimate of the average change in the price of a refrigerator when its annual energy
cost decreases by $1? Given this estimate, would you suggest going ahead with new technology?
Does this estimate make sense? Explain.
1.Average Price= 300.15+ 17.14Energy Cost
2.The estimate of the average change in the price of a refrigerator decreases by $17.15
when its annual energy cost decreases by $1.
3. Does not make sense
Omitted variable bias

P a g e 13 | 21
Data Set
Price Energy cost

$980.00 40

$820.00 33

$850.00 39

$800.00 37

$900.00 39

$870.00 37

$750.00 37

$650.00 40

$600.00 36

$830.00 37

$850.00 38

$530.00 40

$750.00 44

$800.00 43

$1,200.00 40

$500.00 40

$450.00 40

$1,400.00 52

$1,650.00 49

$1,000.00 60

$1,330.00 54

$1,150.00 55

P a g e 14 | 21
 $1,400.00 55

$1,350.00 53

$950.00 58

$1,600.00 56

$2,000.00 52

$2,300.00 54

$975.00 55

$1,330.00 90

$1,250.00 44

$1,800.00 45

$1,000.00 44

$1,200.00 44

$950.00 42

$1,050.00 42

$1,900.00 43

$750.00 46

$800.00 61

$1,500.00 49

$1,400.00 53

CASE
NEWSPAPER CIRCULATION

A Newspaper Company is currently selling newspapers on six days only. It is now

considering the circulation of a Sunday edition. This is a major investment project. (Recall
P a g e 15 | 21
from finance course that, from a financial perspective, an investment project is considered
good if and only if the NPV (Net Present Value) of the project is positive.)

Data pertaining to this important marketing and investment decision are as follows.
• Projected Costs and Profits:

• Fixed cost: $2 million; Operating costs: $1 million a year.

• Expected Profits: $5 per reader per year.

• Cost of capital: 15%.

• It is necessary to know the number of copies of the Sunday edition sold to determine
the revenues and profits. However, it is not known. Call it “X” (the breakeven
quantity) and ask the question: At what circulation (-X) would the launch be
profitable, i.e., the NPV of the project is positive?

• It has been determined that X must be greater than 260,000.

The company must now predict the Circulation of its Sunday Edition to see how it compares
with the breakeven quantity, 260,000. It believes that circulation of Sunday Edition is related
to the circulation of the Daily edition. The daily circulation of the company’s daily edition is
190,000. The company has collected data on the circulation for 35 newspapers with
different levels of circulation for their daily edition and their Sunday edition.

See Data Set – Newspaper Circulation. (Numbers represent circulation in ‘000s)

Should the company launch the Sunday edition? Do a regression analysis to

analyze the data and answer the question.

P a g e 16 | 21
Data Set
Paper Sunday Daily
Des Moines Register 344.522 206.203
Philadelphia Inquirer 982.663 515.523
Tampa Tribune 408.342 321.626
New York Times 1762.015 1209.224
New York News 983.239 781.796
Sacramento Bee 338.355 273.843
Los Angeles Times 1531.526 1164.387
Boston Globe 798.297 516.981
Cincinnati Enquirer 348.743 198.832
Orange Co. Register 407.760 354.842
Miami Herald 553.479 444.580
Chicago Tribune 1133.249 733.775
Detroit News 1215.149 481.765
Houston Chronicle 620.752 449.755
Kansas City Star 423.304 288.571
Omaha World Herald 284.610 223.748
Denver Post 417.778 252.623
St. Louis Post-Dispatch 585.681 391.286
Portland Oregonian 440.923 337.671
Washington Post 1165.567 838.901
Long Island Newsday 960.307 825.512
San Francisco Chronicle 704.322 570.364
Chicago Sun Times 559.093 537.780
Minneapolis Star Tribune 685.974 412.871
Baltimore Sun 488.506 391.951
Pittsburgh Press 557.000 220.464
Rocky Mountain News 432.502 374.009
Boston Herald 235.083 355.627
New Orleans Times-Picayune 324.240 272.279
Charlotte Observer 299.450 238.554
Hartford Courant 323.084 231.177
Rochester Democratic and Chronicle 262.048 133.238
St. Paul Pioneer Press 267.781 201.860
Providence Journal-Bulletin 268.059 197.119
L.A. Daily News 202.613 185.735

P a g e 17 | 21
Regression Output
Regression: Sunday
constant Daily
coefficient 24.763467 1.35117
std error of coef 46.986673 0.09298
t-ratio 0.5270 14.5323
p-value 60.1697% 0.0000%
beta-weight 0.9300
standard error of regression 143.865
R-squared 86.49%
adjusted R-squared 86.08%
number of observations 35
residual degrees of freedom 33
t-statistic for computing
95%-confidence intervals 2.0345

P a g e 18 | 21
Analysis

 The p-value of the β-coefficient is very low. The explanatory variable is highly significant.

 𝑅2 is pretty high. The regression model explains a good amount of variation in Y.

Prediction of Sunday Edition given Daily Circulation:

The predicted value for the average circulation of our Sunday edition is = 281,486

The break-even point was = 260,000.

 Should we launch the Sunday edition of our paper based on this model’ s
forecast of Sunday Circulation?

P a g e 19 | 21
Forecasting
Prediction, using most-
recent regression

constant Daily
coefficients 24.76347 1.351173
values for prediction 190

predicted value of Sunday 281.4864

standard error of
prediction 147.6359
standard error of
regression 143.8646
standard error of
estimated mean 33.15637

confidence level 90.00%

t-statistic 1.6924
residual degr. freedom 33

confidence limits lower 31.63326

for prediction upper 531.3395

confidence limits lower 225.3739

for estimated mean upper 337.5989

P a g e 20 | 21
 Should we launch the Sunday Edition?

P a g e 21 | 21