Chapter 11 Advanced Data Analysis
-
Upload
puputmaulida -
Category
Documents
-
view
224 -
download
0
Transcript of Chapter 11 Advanced Data Analysis
-
8/2/2019 Chapter 11 Advanced Data Analysis
1/25
1- 1Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Essentials of
Marketing Research
MALHOTRA
HALLSHAWOPPENHEIM
ANAPPLIEDORIENTATION
PowerPoint to accompany
1- 1
-
8/2/2019 Chapter 11 Advanced Data Analysis
2/25
1- 2Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
PART FOUR
Chapter 11
Advanced Data Analysing
-
8/2/2019 Chapter 11 Advanced Data Analysis
3/25
11-3Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Chapter Objectives
After reading this chapter, you should be able to: Discuss the scope of the ANOVA technique. Describe one-way ANOVA. Describe n-way ANOVA and the testing of
significance.
Describe ANCOVA. Discuss MANOVA. Discuss the concepts of the correlation coefficient
and the partial correlation. Explain the nature and methods of multiple regression
analysis.
-
8/2/2019 Chapter 11 Advanced Data Analysis
4/25
11-4Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
ANOVA
Analysis of variance (ANOVA) examines thedifferences in the mean values of the dependentvariable (interval scale)associated with the effect ofthe controlled independent variable (nominal scale),
after taking into account the influence of theuncontrolled independent variables.
e.g. Do the brand evaluation of groups exposed to different
commercials vary?How do consumers intentions to buy the brand vary with
different price levels?
-
8/2/2019 Chapter 11 Advanced Data Analysis
5/25
11-5Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
One-way ANOVA
We are testing to determine the effect of in-store
promotion (X) on sales (Y).
H0: 1 = 2 = 3H1: 12 3
Descriptives
Sales
10 8.3000 1.33749 .42295 7.3432 9.2568 6.00 10.00
10 6.2000 1.75119 .55377 4.9473 7.4527 4.00 9.00
10 3.7000 2.00278 .63333 2.2673 5.1327 1.00 7.00
30 6.0667 2.53164 .46221 5.1213 7.0120 1.00 10.00
high
medium
low
Total
N Mean Std. Deviation Std. Error Lower Bound Upper Bound
95% Confidence Interval for
MeanMinimum Maximum
See Table 11.1 page 317 for the data set
-
8/2/2019 Chapter 11 Advanced Data Analysis
6/25
11-6Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
One-way ANOVA cont.
ANOVA
Sales
106.067 2 53.033 17.944 .000
79.800 27 2.956
185.867 29
Between Groups
Within Groups
Total
Sum ofSquares df Mean Square F Sig.
Multiple Comparisons
Dependent Variable: Sales
Tukey HSD
2.1000* .76884 .029 .1937 4.0063
4.6000* .76884 .000 2.6937 6.5063
-2.1000* .76884 .029 -4.0063 -.1937
2.5000* .76884 .008 .5937 4.4063
-4.6000* .76884 .000 -6.5063 -2.6937
-2.5000* .76884 .008 -4.4063 -.5937
(J) In-store promotion
medium
low
high
low
high
medium
(I) In-store promotion
high
medium
low
Mean
Difference
(I-J) Std. Error Sig. Lower Bound Upper Bound
95% Confidence Interval
The mean difference is significant at the .05 l evel.*.
Main effect (in-store promotion)
Residuals
RejectH0, themeansare not
equal
-
8/2/2019 Chapter 11 Advanced Data Analysis
7/2511-7Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
One-way ANOVA cont.
Interpretation
57.1% (ie. 2 = 106.067/185.856) of the variation insales is accounted for by in-store promotion,indicating a modest effect.
The mean sales figures are different, that is atleast one pair of means is statistically different.
All combination of means are statisticallydifferent, therefore the different levels of in-store
promotion will impact sales.
-
8/2/2019 Chapter 11 Advanced Data Analysis
8/2511-8Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
N-way ANOVA
N-way analysis of variance examines thedifferences in the mean values of the dependentvariable (interval scale)associated with the effectof more than one independent variable (nominal
scale). Enables the examination of interactions between
the factors.
-
8/2/2019 Chapter 11 Advanced Data Analysis
9/2511-9Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
N-way ANOVA
We are testing to determine the effect of in-storepromotion and couponing (X) on sales (Y).
H0: 1 = 2H1: 12
-
8/2/2019 Chapter 11 Advanced Data Analysis
10/2511-10Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
N-way ANOVA cont.
Descriptive Statistics
Dependent Variable: Sales
9.2000 .83666 5
7.6000 1.14018 5
5.4000 1.14018 5
7.4000 1.88225 157.4000 1.14018 5
4.8000 .83666 5
2.0000 .70711 5
4.7333 2.43389 15
8.3000 1.33749 10
6.2000 1.75119 10
3.7000 2.00278 10
6.0667 2.53164 30
In-store promotion
high
medium
low
Totalhigh
medium
low
Total
high
medium
low
Total
Coupon level
$20 store-wide coupon
No coupon
Total
Mean Std. Deviation N
-
8/2/2019 Chapter 11 Advanced Data Analysis
11/2511-11Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
N-way ANOVA cont.
Tests of Between-Subjects Effects
Dependent Variable: Sales
162.667a 5 32.533 33.655 .000
1104.133 1 1104.133 1142.207 .000
53.333 1 53.333 55.172 .000
106.067 2 53.033 54.862 .000
3.267 2 1.633 1.690 .206
23.200 24 .967
1290.000 30
185.867 29
Source
Corrected Model
Intercept
COUPON
INSTORE
COUPON * INSTORE
Error
Total
Corrected Total
Type III Sum
of Squares df Mean Square F Sig.
R Squared = .875 (Adjusted R Squared = .849)a.
Overall
Main effect
Main effectInteraction
Interaction isnot significantMain effect of promotion is significant
Main effect of coupon is significant
-
8/2/2019 Chapter 11 Advanced Data Analysis
12/2511-12Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
N-way ANOVA cont.
Interpretation
Higher level of in-store promotion results inhigher sales
The distribution of a storewide coupon results inhigher sales
The effect of each is independent of the other
-
8/2/2019 Chapter 11 Advanced Data Analysis
13/2511-13Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Analysis of Covariance
Examine differences in the mean values of thedependent variable related to the effect of thecontrolled independent variables.
Dependent variable [metric]
Independent variable [one categorical and one metric]
Example
To determine the effect of in-store promotion andcouponing on sales while controlling for the affluenceof clientele.
-
8/2/2019 Chapter 11 Advanced Data Analysis
14/2511-14Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Analysis of Covariance cont.
Tests of Between-Subjects Effects
Dependent Variable: Sales
163.505a 6 27.251 28.028 .000
103.346 1 103.346 106.294 .000
.838 1 .838 .862 .363
53.333 1 53.333 54.855 .000
106.067 2 53.033 54.546 .000
3.267 2 1.633 1.680 .208
22.362 23 .972
1290.000 30
185.867 29
Source
Corrected Model
Intercept
CLIENTEL
COUPON
INSTORE
COUPON * INSTORE
Error
Total
Corrected Total
Type III Sum
of Squares df Mean Square F Sig.
R Squared = .880 (Adjusted R Squared = .848)a.
Not significant
We can conclude that the affluence of the clientele does not have an effect
on the sales of the department store
-
8/2/2019 Chapter 11 Advanced Data Analysis
15/2511-15Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Multivariate Analysis of Variance(MANOVA)
Examine group differences across multipledependent variables simultaneously
Appropriate when 2 or more dependent variablesare correlated
-
8/2/2019 Chapter 11 Advanced Data Analysis
16/25
11-16Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Correlation coefficient
The correlation coefficient, r, is a statisticsummarising the strength of associationbetween two metric (interval or ratio) variables.
-1 0 +1
Strongnegative
relationship
Strongpositive
relationship
Norelationship
Examples
How strongly are sales related to advertising expenditure?
Are consumers perceptions of quality related to their perceptions of price?
-
8/2/2019 Chapter 11 Advanced Data Analysis
17/25
11-17Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Correlation coefficient cont.
Correlations
1 .179**
. .000
446 424
.179** 1
.000 .
424 425
Pearson Correlation
Sig. (2-tailed)
N
Pearson Correlation
Sig. (2-tailed)
N
Trust the website
Satisfaction with Website
Trust the
website
Satisfaction
with Website
Correlation is significant at the 0.01 level (2-tai led).**.
Positive relationshipbetween trust and
satisfaction
-
8/2/2019 Chapter 11 Advanced Data Analysis
18/25
11-18Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Regression Analysis
ANOVAb
16.598 1 16.598 13.953 .000a
501.986 422 1.190
518.584 423
Regression
Residual
Total
Model
1
Sum ofSquares df Mean Square F Sig.
Predictors: (Constant), Trust the websitea.
Dependent Variable : Satisfaction with Websiteb.
Model Summary
.179a .032 .030 1.09066
Model
1
R R Square
Adjusted
R Square
Std. Error of
the Estimate
Predictors: (Constant), Trust the websitea.
Model is significant
-
8/2/2019 Chapter 11 Advanced Data Analysis
19/25
11-19Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Regression Analysis cont.
Coefficientsa
4.102 .199 20.656 .000.143 .038 .179 3.735 .000
(Constant)Trust the website
Model
1
B Std. Error
Unstandardized
Coefficients
Beta
Standardized
Coefficients
t Sig.
Dependent Variable: Satisfaction with Websitea.
Sales = 4.1 + 0.14 Trust
Significant linear relationshipbetween satisfaction with the
website and trust in thewebsite
-
8/2/2019 Chapter 11 Advanced Data Analysis
20/25
11-20Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Partial correlation coefficientCorrelations
1 .222** .277** .373** .237** .235** .179** .286**
. .000 .000 .000 .000 .000 .000 .000
425 424 399 421 421 421 424 425
.222** 1 .162** .119* -.007 .250** .192** .268**
.000 . .001 .012 .878 .000 .000 .000
424 447 414 442 443 442 444 447
.277** .162** 1 .359** .078 .147** .106* .300**
.000 .001 . .000 .112 .003 .030 .000
399 414 417 414 415 413 415 417.373** .119* .359** 1 .087 .077 .050 .334**
.000 .012 .000 . .069 .107 .293 .000
421 442 414 444 440 440 442 444
.237** -.007 .078 .087 1 .139** .191** .136**
.000 .878 .112 .069 . .003 .000 .004
421 443 415 440 446 441 442 446
.235** .250** .147** .077 .139** 1 .327** .131**
.000 .000 .003 .107 .003 . .000 .006
421 442 413 440 441 445 441 445
.179** .192** .106* .050 .191** .327** 1 .128**
.000 .000 .030 .293 .000 .000 . .007
424 444 415 442 442 441 446 446
.286** .268** .300** .334** .136** .131** .128** 1
.000 .000 .000 .000 .004 .006 .007 .
425 447 417 444 446 445 446 450
Pearson Correla tion
Sig. (2-tailed)
N
Pearson Correla tion
Sig. (2-tailed)
N
Pearson Correla tion
Sig. (2-tailed)
NPearson Correla tion
Sig. (2-tailed)
N
Pearson Correla tion
Sig. (2-tailed)
N
Pearson Correla tion
Sig. (2-tailed)
N
Pearson Correla tion
Sig. (2-tailed)
N
Pearson Correla tion
Sig. (2-tailed)
N
Satisfaction with Website
Ease of use
Sociable
innovative
Speed of download
Customer control
Trust the website
Appearance
Satisfaction
with Website Ease of use Sociable innovative
Speed of
download
Customer
control
Trust the
website Appearance
Correlation is significant at the 0.01 level (2-tailed).**.
Correlation i s signifi cant at the 0.05 l evel (2-tailed).*.
-
8/2/2019 Chapter 11 Advanced Data Analysis
21/25
11-21Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Multiple Regression
Model Summary
.494a .245 .231 .95995
Model
1
R R Square
Adjusted
R Square
Std. Error of
the Estimate
Predictors: (Constant), Trust the website, innovative,
Ease of use, Speed of download, Sociable, Customer
control, Appearance
a.
ANOVAb
113.922 7 16.275 17.661 .000a
352.013 382 .922
465.936 389
Regression
Residual
Total
Model
1
Sum of
Squares df Mean Square F Sig.
Predictors: (Constant), Trust the website, innovative, Ease of use, Speed of
download, Sociable, Customer control, Appearance
a.
Dependent Variable: Satisfaction with Websiteb.
1
-
8/2/2019 Chapter 11 Advanced Data Analysis
22/25
11-22Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Multiple Regression cont.
Coefficientsa
.283 .450 .629 .5309.897E-02 .051 .097 1.951 .052
5.793E-02 .024 .115 2.425 .016
6.352E-02 .034 .093 1.891 .059
.220 .046 .237 4.757 .000
.181 .054 .156 3.359 .001
.117 .044 .129 2.658 .008
5.230E-02 .038 .067 1.373 .171
(Constant)Appearance
Ease of use
Sociable
innovative
Speed of download
Customer controlTrust the website
Model
1
B Std. Error
Unstandardized
Coefficients
Beta
Standardized
Coefficients
t Sig.
Dependent Variable : Satisfaction with Websitea.
2
-
8/2/2019 Chapter 11 Advanced Data Analysis
23/25
11-23Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Multiple Regression cont.
Interpretation 1 The overall model model is significant at = 0.05
H0: 1 = 2 = 3 = 4 = 5 = 6 = 7 =0H1: 1234567 0
2Testing which of the independent variables have a significantimpact on satisfaction with the website
H0: 1 = 0H1: 1 0
Ease of use, innovative website,speed of download, and
respondents perception of control are significant variables ininfluencing satisfaction rating of the website.
-
8/2/2019 Chapter 11 Advanced Data Analysis
24/25
11-24Malhotra Hall Shaw Oppenheim Essentials of Marketing Research Copyright 2004 Pearson Education Australia
Multiple Regression cont.
ANOVAb
111.267 4 27.817 30.119 .000a
376.813 408 .924
488.080 412
Regression
Residual
Total
Model
1
Sum of
Squares df Mean Square F Sig.
Predictors: (Constant), Customer control, innovative, Speed of download, Ease of
use
a.
Dependent Variable: Satisfaction with Websiteb.
Coefficientsa
.673 .422 1.595 .112
8.088E-02 .023 .159 3.522 .000
.293 .040 .319 7.238 .000
.205 .051 .178 4.040 .000
.134 .041 .148 3.266 .001
(Constant)
Ease of use
innovative
Speed of download
Customer control
Model1
B Std. Error
UnstandardizedCoefficients
Beta
StandardizedCoefficients
t Sig.
Dependent Variable: Satisfaction with Websitea.
-
8/2/2019 Chapter 11 Advanced Data Analysis
25/25
11 25
Multiple Regression cont.
In predicting satisfaction ratings of a website wecan use the following equation
Satisfaction = .673 + 0.08ease + 0.29innov. + 0.21speed + 0.13control
For every 1 unit increase in ease the satisfactionrating will increase by 0.08 units
Substituting values for each of the variables will
produce the overall satisfaction rating of thewebsite.