You are on page 1of 14

# For sugar level of 2, there are 5 data points which shows different sweetness level .

This implies
that there is some other variable which is influencing.
Also as sugar increases, the sweetness level increases.
As sugar level increases, the sweetness seem to increase by constant level.
There is also some other variable which is affecting sugar level at each level. ( Here 2,4,6,8,10
level of honey ,cream, butterfa are added at each level of sugar)

The other factor seem to influence sweetness level much higher than the other factors which
was influencing sugar.

Is there any other conclusion which can be inferred at such pattern?

Is the conclusion in the below three graphs inconclusive ?

No conclusion can be drawn

No conclusion can be drawn

No conclusion can be drawn

Step two
Run Regression between sweetness and other 4 independent variable
Run Regression between richness and other 4 independent variable

Variables Entered/Removed
a

Model
Variables
Entered
Variables
Removed Method
1 SUGAR,
CREAM,
HONEY,
BUTTERFA
b

. Enter
a. Dependent Variable: SWEETNES
b. All requested variables entered.

Model Summary
Model R R Square
Square
Std. Error of the
Estimate
1 1.000
a
1.000 1.000 .00000
a. Predictors: (Constant), SUGAR, CREAM, HONEY, BUTTERFA

ANOVA
a

Model Sum of Squares df Mean Square F Sig.
1 Regression 160.000 4 40.000 . .
b

Residual .000 20 .000

Total 160.000 24

a. Dependent Variable: SWEETNES
b. Predictors: (Constant), SUGAR, CREAM, HONEY, BUTTERFA

Coefficients
a

Model
Unstandardized Coefficients
Standardized
Coefficients
t Sig. B Std. Error Beta
1 (Constant) 3.078E-015 .000

.000 1.000
HONEY .400 .000 .447 218999805.346 .000
CREAM 5.515E-018 .000 .000 .000 1.000
BUTTERFA -6.893E-017 .000 .000 .000 1.000
SUGAR .800 .000 .894 437999610.691 .000
a. Dependent Variable: SWEETNES

Conclusions:
Adj R sq is 1 . Therefore 100% of the data is explained.
F and Significance shows nothing .What does it mean ?
What does the constant represent?
Conclusion from Perception database:
What affects sweetness and richness most?

Preferences DataBase:
Run ScatterPlot:
Unit Sales vs Sweetness

Conclusion :
Segment 1 prefers sweetness 4
What gives us clue to test richness square , price square and sweetness square.
Why not the product terms.
We dont use any BPV here .Why ?

Segment3 prefers sweetness 3

Segment 1 prefers price of 8

Step 2:
Run Regression :
LnUnitSales vs Price ,PriceSquare , Sweetness,SweetnessSquare,Richness ,RichnessSquare.

Regression

Notes
Output Created 17-FEB-2014 17:25:24

Input Data C:\Rice\Term3\DA_ICG_Preferences
(1)_Review.sav
Active Dataset DataSet1
Filter <none>
Weight <none>
Split File <none>
N of Rows in Working Data File 25
Missing Value Handling Definition of Missing User-defined missing values are treated as
missing.
Cases Used Statistics are based on cases with no missing
values for any variable used.
Syntax REGRESSION
/MISSING LISTWISE
/STATISTICS COEFF OUTS R ANOVA
/CRITERIA=PIN(.05) POUT(.10)
/NOORIGIN
/DEPENDENT LU1
/METHOD=ENTER RICHNESS PRICE
SweetnessSquare SWEETNESS
RichnessSquare PriceSquare.
Resources Processor Time 00:00:00.05
Elapsed Time 00:00:00.08
Memory Required 5872 bytes
Residual Plots
0 bytes

[DataSet1] C:\Rice\Term3\DA_ICG_Preferences (1)_Review.sav

Variables Entered/Removed
a

Model Variables Entered
Variables
Removed Method
1 PriceSquare,
SWEETNESS,
RichnessSquare,
RICHNESS,
SweetnessSquare,
PRICE
b

. Enter
a. Dependent Variable: LU1
b. All requested variables entered.

Model Summary
Model R R Square Adjusted R Square
Std. Error of the
Estimate
1 1.000
a
1.000 1.000 .03178
a. Predictors: (Constant), PriceSquare, SWEETNESS, RichnessSquare,
RICHNESS, SweetnessSquare, PRICE

ANOVA
a

Model Sum of Squares df Mean Square F Sig.
1 Regression 157.652 6 26.275 26008.772 .000
b

Residual .013 13 .001

Total 157.665 19

a. Dependent Variable: LU1
b. Predictors: (Constant), PriceSquare, SWEETNESS, RichnessSquare, RICHNESS, SweetnessSquare, PRICE

Coefficients
a

Model
Unstandardized Coefficients
Standardized
Coefficients
t Sig. B Std. Error Beta
1 (Constant) -4.632 .060

-76.778 .000
RICHNESS 1.406 .008 1.988 179.155 .000
PRICE .978 .011 1.477 90.416 .000
SweetnessSquare -.300 .001 -3.341 -288.483 .000
SWEETNESS 2.698 .011 2.913 251.481 .000
RichnessSquare -.100 .001 -1.889 -169.671 .000
PriceSquare -.049 .000 -1.650 -101.188 .000
a. Dependent Variable: LU1

Why are some constants positive, while some negative ? What is the implication?
In this case , what does the restricted model consider ? Just the unit vector ?
Is there anything else to learn from the regression?

Segment 2 regression:

[DataSet1] C:\Rice\Term3\DA_ICG_Preferences (1)_Review.sav

Variables Entered/Removed
a

Model Variables Entered
Variables
Removed Method
1 PriceSquare,
SWEETNESS,
RichnessSquare,
SweetnessSquare,
RICHNESS,
PRICE
b

. Enter
a. Dependent Variable: LU2
b. All requested variables entered.

Model Summary
Model R R Square Adjusted R Square
Std. Error of the
Estimate
1 1.000
a
1.000 1.000 .04207
a. Predictors: (Constant), PriceSquare, SWEETNESS, RichnessSquare,
SweetnessSquare, RICHNESS, PRICE

ANOVA
a

Model Sum of Squares df Mean Square F Sig.
1 Regression 230.351 6 38.392 21688.922 .000
b

Residual .030 17 .002

Total 230.381 23

a. Dependent Variable: LU2
b. Predictors: (Constant), PriceSquare, SWEETNESS, RichnessSquare, SweetnessSquare, RICHNESS, PRICE

Coefficients
a

Model
Unstandardized Coefficients
Standardized
Coefficients
t Sig. B Std. Error Beta
1 (Constant) -.280 .073

-3.836 .001
RICHNESS .208 .010 .262 21.621 .000
PRICE -.094 .013 -.126 -7.344 .000
SweetnessSquare -.100 .001 -1.714 -146.310 .000
SWEETNESS 1.597 .009 2.060 175.906 .000
RichnessSquare .029 .001 .501 41.531 .000
PriceSquare -.010 .001 -.301 -17.542 .000
a. Dependent Variable: LU2

How do we decide on Media Allocation?