Vous êtes sur la page 1sur 13

18010040

SAM-Assignment
ADS

Introduction:
Mary Kazinczi, the newly appointed general manager of Alfonses Department Store wanted to
be sure that the newspaper advertising was the reason behind the true increase in sales. She

Alfonsos Deppartment Store


believed that each department was competing not only with the department in the competitors

store but with the other departments in Alfonse itself. She also believed that these advertisements
transferred sales from one point in time to another and costing the store profits.
The first step in developing a policy for newspapers advertising was to understand if the store as
a whole gained from the advertising or not.
Model Building:
The data consisted of the following variables:
1. Alfonsos Sales.
2. Alfonsos advertising expenditure.
3. All store sales.
4. Other stores advertising expenditure.
In addition to this, a new variable was formed which included all other stores sales in the
corresponding weeks.
All store sales = Other Store Sales + Alfonso Sales
A multiple regression was carried out on the data to out the relationship between advertising
and sales. The regression equation that resulted from this regression was:

Roll#: 18010040

E (A.sales) = 106676.814 + 9.283 Aadvert 11.340 otherAdvert + 1.306 otherSales

Page

0 of 13

18010040
SAM-Assignment
ADS

It shows that there is a positive relationship between advertising and sales. A 1 unit increase
in Alfonsos advertising expenditure will lead to a related increase in Aflonsos sales by
9.283 units holding all the other independent variables constant.
Similarly, a 1 unit increase in other stores advertising expenditure will lead to a related
decrease in Aflonsos sales by 11.340 units holding all the other independent variables
constant.
Lastly, a 1 unit increase in other stores sales will lead to a related increase in Aflonsos sales
by 1.306 units holding all the other independent variables constant.
The value of R2 is 0.972 which means that 97.2% of the variability in Alfonsos sales in the
sample is explained by the estimated regression equation with Alfonsos advertising expenditure,
other store sales and other store advertising as the independent variables.
Significance Tests:
F-Test:
H0: 1, 2, 3 = 0
Ha: 1, 2, 3 0
Since the p-value of the F-test is less than the 0.05 level of significance, we will reject H0 and
conclude that the independent variables are significant.
t-Test:
H0: 1 = 0
Ha: 1 0
As p-value is less than the 0.05 level of significance, we will reject H0 and conclude that the 1 is
significant.
Similarly, 2 and 3 when independently tested are all significant at 0.05 level of significance.
Model Assumptions:
There are four basic assumptions of the model. It can be seen from the residual plot in exhibit 1
that the assumption of homoscedasticity does not hold which the variance of x for the average
value of y is not constant.

Page 2 of 13

18010040
SAM-Assignment
ADS

Additionally, as the values in the plot are in a scattered manner, it can be assumed that the
assumption of error values being independent holds. We can hire a statistician to solve the
problem of heteroscedasticity.

Conclusion:
It can be seen from the regression results that the store as a whole does gain from advertising
based on the positive relationship between the two variable. One alternate scenario to determine
the effectiveness of advertising was to create a dummy variable and see that how sales are
affected in the weeks that the other stores advertised however the regression results from not
substantially different with the dummy variable turning out to be insignificant.
Recommendations:
In order to determine the future of advertising, we need to carry out some additional model
regressions in order to increase the certainty that Alfonsos sales benefit from advertisement.
First, data should be collected on sales of Alfonso of all departments in weeks their
advertisement took place instead of the other. Secondly, a dummy variable could be created to
see the difference in sales from one point in time to another in order to determine whether
Alfonso is not costing the stores profit by selling items at a lower cost.

Page 3 of 13

18010040
SAM-Assignment
ADS

Exhibit 1
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT aSales
/METHOD=ENTER aAdvert othAdver otherSales
/PARTIALPLOT ALL
/SCATTERPLOT=(*ZRESID ,*ZPRED)
/RESIDUALS HISTOGRAM(ZRESID) NORMPROB(ZRESID).

Regression

Notes
Output Created

22-SEP-2016 05:43:29

Comments

Input

Active Dataset

DataSet1

Filter

<none>

Weight

<none>

Split File

<none>

N of Rows in Working Data

26

File
Missing Value Handling

Definition of Missing

User-defined missing values


are treated as missing.

Page 4 of 13

18010040
SAM-Assignment
ADS

Statistics are based on cases


Cases Used

with no missing values for


any variable used.
REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS
R ANOVA
/CRITERIA=PIN(.05)
POUT(.10)
/NOORIGIN
/DEPENDENT aSales

Syntax

/METHOD=ENTER aAdvert
othAdver otherSales
/PARTIALPLOT ALL
/SCATTERPLOT=(*ZRESID
,*ZPRED)
/RESIDUALS
HISTOGRAM(ZRESID)
NORMPROB(ZRESID).

Resources

Processor Time

00:00:00.84

Elapsed Time

00:00:01.22

Memory Required

3552 bytes

Additional Memory Required


for Residual Plots

1160 bytes

[DataSet1]

Variables Entered/Removeda
Model

Variables

Variables

Entered

Removed

Method

otherSales,
1

othAdver,
aAdvert

. Enter

a. Dependent Variable: aSales


b. All requested variables entered.
Page 5 of 13

18010040
SAM-Assignment
ADS

Model Summaryb
Model

.986a

R Square

Adjusted R

Std. Error of the

Square

Estimate

.972

.968

39095.190

a. Predictors: (Constant), otherSales, othAdver, aAdvert


b. Dependent Variable: aSales

ANOVAa
Model

Sum of Squares
Regression

Residual

Total

df

Mean Square

1154389491562.

661
33625544533.37
7
1188015036096.
038

384796497187.5
54

Sig.

251.759

.000b

22 1528433842.426

25

a. Dependent Variable: aSales


b. Predictors: (Constant), otherSales, othAdver, aAdvert

Coefficientsa
Model

Unstandardized Coefficients

Standardized

Sig.

Coefficients
B
(Constant)
1

Std. Error

106676.814

75834.982

9.283

.723

-11.340
1.306

aAdvert
othAdver
otherSales

Beta
1.407

.173

.542

12.844

.000

1.403

-.311

-8.081

.000

.048

1.203

27.476

.000

a. Dependent Variable: aSales

Page 6 of 13

18010040
SAM-Assignment
ADS

Residuals Statisticsa
Minimum

Maximum

Mean

Std. Deviation

Predicted Value

1695945.13

2549822.00

2150209.81

214885.038

26

Residual

-64338.926

60574.297

.000

36674.539

26

Std. Predicted Value

-2.114

1.860

.000

1.000

26

Std. Residual

-1.646

1.549

.000

.938

26

a. Dependent Variable: aSales

Charts

Page 7 of 13

18010040
SAM-Assignment
ADS

Page 8 of 13

18010040
SAM-Assignment
ADS

Page 9 of 13

18010040
SAM-Assignment
ADS

Page 10 of 13

18010040
SAM-Assignment
ADS

Page 11 of 13

18010040
SAM-Assignment
ADS

Page 12 of 13

18010040
SAM-Assignment
ADS

Page 13 of 13

Vous aimerez peut-être aussi