Anda di halaman 1dari 19

Mann-Whitney Test

H0 : there is no sig diff between males and females in rating given to quality
H1 : there is sig diff between males and females in rating given to quality
Test Statisticsa
rating given to
quality
Mann-Whitney U
Wilcoxon W

89.500
225.500

-.957

Asymp. Sig. (2tailed)


Exact Sig. [2*(1tailed Sig.)]

.338
.355b

a. Grouping Variable: gender of the respondent


b. Not corrected for ties.

Interpretation
1. As sig value is more than 5%
2. So H0 is accepted
Points to remember
1. Alternative test of mann-Whitney is independent sample t- test

Independent Samples Test


H0: there is no sig diff btwn males and females in terms of rating given to quality
H1: there is sig diff btwn males and females in terms of rating given to quality

Independent Samples Test


Levene's
Test for
Equality
of
Variances

Sig.
rating

Equal

given to

variances

quality

assumed

.822

t-test for Equality of Means

df

-.92
4

Sig.
(2tailed)

95% Confidence Interval of the


Difference
Mean

Std. Error

Difference Difference

Lower

Upper

28

.364

-.482

.522

-1.551

.587

27.751

.362

-.482

.520

-1.549

.584

Equal
variances
not

-.92
6

assumed

Interpretation
1. As sig value is more than 5% (shown by table above) for both cases H0 is accepted
2. (If it is less than 0.05 H0 is rejected and H1 is accepted)

One Way Annova


H0: there is no sig D/B age groups in terms of rating given to quality
H1: there is sig D/B age groups in terms of rating given to quality
ANOVA
rating given to quality
Sum of Squares
Between Groups

df

Mean Square

1.658

.415

Within Groups

57.042

25

2.282

Total

58.700

29

Sig.
.182

.946

Multiple Comparisons
Dependent Variable: rating given to quality
Tukey HSD
95% Confidence Interval

Mean
(I) Age of respondent

(J) Age of respondent

15-25

25-35

-.500

.872

35-45

-.567

.915

45-55

-.292

.816

>55

.033

.915

15-25

.500

.872

35-45

-.067

.915

45-55

.208

.816

>55

.533

.915

15-25

.567

.915

25-35

.067

.915

45-55

.275

.861

>55

.600

.955

15-25

.292

.816

25-35

-.208

.816

35-45

-.275

.861

.325

.861

15-25

-.033

.915

25-35

-.533

.915

35-45

-.600

.955

45-55

-.325

.861

25-35

35-45

45-55

>55
>55

Difference (I-J)

Std. Error

Sig.

.978
.971
.996
1.000
.978
1.000
.999
.976
.971
1.000
.998
.969
.996
.999
.998
.995
1.000
.976
.969
.995

Lower Bound

Upper Bound

-3.06

2.06

-3.25

2.12

-2.69

2.10

-2.65

2.72

-2.06

3.06

-2.75

2.62

-2.19

2.60

-2.15

3.22

-2.12

3.25

-2.62

2.75

-2.25

2.80

-2.21

3.41

-2.10

2.69

-2.60

2.19

-2.80

2.25

-2.20

2.85

-2.72

2.65

-3.22

2.15

-3.41

2.21

-2.85

2.20

Interpretation
1. As sig value is more than 5% (.946) H0 is accepted
2. From multiple comparison we can see that all values are more than 5%. H0 is accepted
3. So there is no sig diff in rating between any age group compared to other age groups

As per rating given to price


One way Annova
H0 : There is no sig D/B rating given to price in terms of age group
H1 : There is sig D/B rating given to price in terms of age group
ANOVA
rating given to price
Sum of Squares

Df

Mean Square

Between Groups

12.900

3.225

Within Groups

47.100

25

1.884

Total

60.000

29

Sig.

1.712

179

Multiple Comparisons
Dependent Variable: rating given to price
Tukey HSD
95% Confidence Interval

Mean
(I) Age of respondent

(J) Age of respondent

Difference (I-J)

Std. Error

15-25

25-35

-.500

.792

35-45

-1.200

.831

Sig.

.969
.606

Lower Bound

Upper Bound

-2.83

1.83

-3.64

1.24

25-35

35-45

45-55

>55

45-55

-1.500

.741

>55

-1.800

.831

15-25

.500

.792

35-45

-.700

.831

45-55

-1.000

.741

>55

-1.300

.831

15-25

1.200

.831

25-35

.700

.831

45-55

-.300

.782

>55

-.600

.868

15-25

1.500

.741

25-35

1.000

.741

35-45

.300

.782

>55

-.300

.782

15-25

1.800

.831

25-35

1.300

.831

35-45

.600

.868

45-55

.300

.782

.284
.225
.969
.915
.664
.533
.606
.915
.995
.957
.284
.664
.995
.995
.225
.533
.957
.995

-3.68

.68

-4.24

.64

-1.83

2.83

-3.14

1.74

-3.18

1.18

-3.74

1.14

-1.24

3.64

-1.74

3.14

-2.60

2.00

-3.15

1.95

-.68

3.68

-1.18

3.18

-2.00

2.60

-2.60

2.00

-.64

4.24

-1.14

3.74

-1.95

3.15

-2.00

2.60

Interpretation
1. As sig value is more than 5% (0.179) H0 is accepted
2. From multiple comparison we can see that all values are more than 5% and so there is no
sig diff in rating between any age group compared to other age groups
3. Kruskal wallis test (Alternative to One Way Annova)

Chi-square test
H0 : there is no sig D/B values of rating given to quality
H1 : there is sig D/B values of rating given to quality
Interpretation

Test Statistics

As Sig value is more than 5% H0 is accepted

rating given to
quality
1.667a

Chi-Square
Df

Asymp
. Sig.

.797

a. 0 cells (0.0%) have expected


frequencies less than 5. The
minimum expected cell
frequency is 6.0.

Two way chi-square

rating given to quality * gender of the respondent Crosstabulation


Count
gender of the respondent
male
rating given to quality

Total

Female

Total

strongly disagree

disagree

neutral

agree

strongly agree

16

14

30

The above table shows the responses of rating classified by gender


H0: there is no association between gender and rating given to quality
H1: there is association between gender and rating given to quality
Chi-Square Tests
Asymp. Sig. (2Value

Pearson ChiSquare

df

2.989a

sided)

.560

a. 10 cells (100.0%) have expected count less than 5. The minimum expected
count is 1.87.

Interpretation
As sig value is more than 5% (.56) H0 is accepted

RELIABILITY TEST:
Reliability Statistics
Cronbach's Alpha

interpretation

N of Items

.171

10

1. THERE IS LESS RELIABILITY


2. IT SHOULD BE LIES BETWEEN 0.6 0.9

KMO and Bartlett's Test


KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.
Bartlett's Test of Sphericity

Approx. Chi-Square
df
Sig.

.489
44.295
45

.502

Important point remember


1. KMO > 0.6
&
2. SIG < 0.05
EITHER OF THE ABOVE CONDITIONS HAVE TO BE SATISFIED FOR DOING
FACTOR ANALYSIS

Interpretation
the given data is not good for Factor analysis

Rotated Component Matrixa


Component
1

facilities available

.105

.038

.077

.812

Location

.045

.418

-.424

-.124

lighting inside

.197

-.056

.518

-.531

parking availability

-.101

.770

-.091

.315

responsiveness of employees

-.054

-.056

.751

.057

timeliness of service

-.339

.657

-.121

-.304

.400

-.097

-.527

.096

-.075

-.036

.251

.006

-.144

-.310

.565

.295

.078

assurance given by employees


Empathy
solution to the problem

.754
.711

reliability of service

.365

Extraction Method: Principal Component Analysis.


Rotation Method: Varimax with Kaiser Normalization.
a.

Rotation converged in 13 iterations.

For sample size 50 the factor loading (highest for each statement) should be more than 0.7 and
for 250 it should be minimum 0.4 (for sample sizes between 50 and 250, approximation has to be
used for finding factor loading ) in this study there are 5 statements are having less .7 and so they
are removed from questionnaire
1. Location
2. Lighting inside
3. Timeliness
4. Assurance
5. reliability

first component has empathy and solution to


problem. This component can be called as
EMPATHY.
Second component has parking and timeliness of
service (reliability of service is not considered as
the value is less .7) this component can be called as
SERVICE.
Third component has responsiveness (location and
Assurance are removed as the value is less than
0.7) this can be called as RESPONSIVENESS
Fourth component has facilities available (lighting
is removed as value is less than 0.7) this
component can be called as FACILITIES.
(the value mentioned above is called as factor loading)
In short the factor analysis has helped in
1.reducing the no.of statements by using factor loading

2.Combining statements which are similar into groups


called as components and naming them
3.Finding out relative importance of components by
using % of variance (total variance table)

KMO and Bartlett's Test

.572

Kaiser-Meyer-Olkin Measure of Sampling Adequacy.


Bartlett's Test of Sphericity

Approx. Chi-Square

899.915

df

351

.000

Sig.

Interpretation
1. KMO value is less than .6( 0.572)
2. sig is less than .05 (.000)
3. So data is reasonably good for factor analysis
Rotated Component Matrixa
Component
1
IMPORTANCE FOR
CREATING NEW CONTACTS
IMPORTANCE FOR NEED
FOR INFORMATION

-.263

.160

.288

.155

-.157

.503

.412

.070

.286

-.018

.121

.336

-.322

.571

IMPORTANCE FOR
INTERACTION WITH

.021

-.022

.295

.435

.404

.121

.014

.108

.060

-.070

.063

.099

.195

.798

.123

-.325

.642

.256

.104

.281

-.082

.109

-.368

.556

.284

.185

.091

.029

.164

-.230

.208

.833

-.089

.031

.127

.246

.089

-.042

.847

-.067

-.124

.055

-.054

.238

.038

.640

.271

.036

.187

.090

.301

.032

.735

.377

.024

-.253

.033

.288

-.110

.172

.731

-.063

.208

.353

.557

.156

-.163

.373

-.106

.141

.039

.332

.540

-.151

.373

.073

.098

.211

.805

-.180

.132

.045

.111

.146

-.099

.879

-.007

.140

.039

.135

.007

.015

.777

-.003

.055

.192

-.277

.051

-.065

.246

.509

.316

.205

.369

-.286

.205

-.111

.834

.051

-.052

.052

-.069

FRIENDS AND RELATIVES


IMPORTANCE FOR
ADVISING PEOPLE ON
HEALTH
IMPORTANCE FOR PHOTO
SHARING
IMPORTANCE FOR VIDEO
SHARING
IMPORTANCE FOR
ADVERTISING OF
PRODUCTS
IMPORTANCE FOR SELLING
PRODUCTS
IMPORTANCE FOR
PUBLISHING NEWS
IMPORTANCE FOR
PROMOTING BUSINESS
IMPORTANCE FOR FINDING
PROPERTY
IMPORTANCE FOR
RECEIVING ASSISTANCE
WHEN IN TROUBLE
IMPORTANCE FOR
CONNECTING WITH OLD
FRIENDS
IMPORTANCE FOR
GETTING JOB
RECOMMENDATIONS
IMPORTANCE FOR
GETTING NEW EMPLOYEES
IMPORTANCE FOR
GETTING STUDY
MATERIAL
IMPORTANCE FOR
HELPING IN BLOGGING
IMPORTANCE FOR
HELPING IN CHATTING

IMPORTANCE FOR MAKING

.301

.109

.833

-.032

.042

-.146

.045

.356

-.228

-.045

-.228

.208

.595

.186

.338

.031

.085

.051

.049

.799

-.015

.808

.002

.213

.174

.149

.046

-.053

.801

-.058

.241

.046

.020

-.020

.034

.824

.071

.041

.146

.152

.281

.063

.859

.203

.094

.102

.042

.169

.094

.344

-.096

.273

-.115

.539

.359

.003

.282

.126

.311

.285

.679

.101

.040

GROUPS
IMPORTANCE REGARDING
COMPLEXITY OF USE
IMPORTANCE OF LAYOUT
OF THE SITE
IMPORTANCE OF
PERFORMANCE OF SITE
IMPORTANCE OF COLOUR
SCHEMES OF SITE
IMPORTANCE OF VISUAL
ASPECTS OF SITE
IMPORTANCE OF
CUSTOMISATION OF SITE
IMPORTANCE FOR EASE OF
COMMUNICATION
IMPORTANCE GIVEN FOR
EASE OF CONNECTION

Extraction Method: Principal Component Analysis.


Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 13 iterations.

First component has


IMPORTANCE OF PERFORMANCE OF SITE
IMPORTANCE OF COLOUR SCHEMES OF SITE
IMPORTANCE OF VISUAL ASPECTS OF SITE
IMPORTANCE OF CUSTOMISATION OF SITE
This component can be called as SITE FEATURES.

Second Component has


IMPORTANCE FOR GETTING JOB
GETTING NEW EMPLOYEES
IMPORTANCE FOR GETTING STUDY MATERIAL
This component can be called as JOB ANALYSIS.

Third Component has


IMPORTANCE FOR HELPING IN CHATTING
IMPORTANCE FOR MAKING GROUPS
This component can be called as SOCIAL NEEDS

Fourth Component has


IMPORTANCE FOR ADVERTISING OF PRODUCTS
IMPORTANCE FOR SELLING PRODUCTS
IMPORTANCE FOR PROMOTING BUSINESS
This component can be called as Marketing strategy

Fifth Component has


IMPORTANCE FOR FINDING PROPERTY
IMPORTANCE GIVEN FOR EASE OF CONNECTION
This component can be called as Connection

Sixth Component has


IMPORTANCE OF LAYOUT OF THE SITE
This component can be called as Plant Layout

Seventh Component has


IMPORTANCE FOR ADVISING PEOPLE ON HEALTH
This component can be called as Health measures

Cluster Analysis
Final Cluster Centers
Cluster
1

site_features

3.68

3.95

Job_Ananlysis

3.08

3.69

Social_Needs

4.12

4.11

Marketing_Strategy

3.13

3.60

Connection

3.26

3.73

plant_layout

2.79

4.12

health_measures

2.34

3.73

Number of Cases in each


Cluster
Cluster

29.000

33.000

Valid

62.000

Missing

12.000

Interpretation
1. Based on hierarchical clustering we get 2 clusters

2. Based on final cluster we can see that for except social needs all other
components gives higher value (more important) in the second cluster. For
social needs both components are giving roughly same value.
3. As seen from the no.of cases second cluster is more important and so the
statements mentioned above are all areas of important for the study and
companies in social networking have to focus on all these parameters.

Dependent (related) Samples:


1. t-test
Paired Samples Correlations
N
Pair 1

Correlation

sales before advertisement

20

& sales After advertisement

Sig.

.836

.049

Interpretation
As seen from the above table the sig is .836 (more than 0.05) and so there is no correlation
between sales before advertisement and sales after advertisement.

Paired Samples Test


Paired Differences
95% Confidence
Std.
Mean

Deviation

Std. Error Interval of the Difference


Mean

Lower

Upper

Sig. (2t

df

tailed)

Pair sales before


1

advertisement sales After


advertisement

Hypothesis

149899.650

289197.513 64666.530

285248.252 14551.048 2.318

19

.032

H0 : there is no significant difference between sales before advertisement and sales


after advertisement
H1 : there is significant difference between sales before advertisement and sales
after advertisement

Interpretation
As sig value is 0.032 (less than 0.05), H0 is rejected and H1 is accepted

2. Non Parametric test (Wilcoxon signed rank test)

CORRELATION (BIVARIATE)
Correlations

sales of the product (in

Pearson Correlation

lakhs)

Sig. (2-tailed)
N
Pearson Correlation

sales of the

price of the

advertisement

product (in

product (in

expenditure(in

lakhs)

rupees)

lakhs)

.943**

.969**

.000

.000

20

20

20

**

.950**

.943

price of the product (in

Sig. (2-tailed)

rupees)

.000

advertisement

Pearson Correlation

expenditure(in lakhs)

Sig. (2-tailed)

.000

20

20

20

**

**

.969

.950

.000

.000

20

20

20

**. Correlation is significant at the 0.01 level (2-tailed).

REGRESSION
Model Summaryb
Change Statistics

Model

.971a

Adjusted R

Std. Error of

R Square

Square

Square

the Estimate

Change

Change

.943

.937

24.548

.943

Sig. F

141.864

df1

df2
2

Change
17

a. Predictors: (Constant), advertisement expenditure(in lakhs), price of the product (in rupees)
b.

Dependent Variable: sales of the product (in lakhs)

Interpretation
1. R-Square value shows the influence of independent variables on dependent variable
2. The normal range is 0.6 to 0.95.
3. Higher values indicated stronger influence

ANOVAa
Model
1

Sum of Squares
Regression
Residual
Total

df

Mean Square

170972.856

85486.428

10244.094

17

602.594

181216.950

19

F
141.864

Sig.

.000b

a. Dependent Variable: sales of the product (in lakhs)


b. Predictors: (Constant), advertisement expenditure(in lakhs), price of the product (in rupees)

.000

Interpretation
As seen from the Anova table, sig value is 0.000 (less than 0.05) which shows that there is
influence of independent variable on dependent variable

COEFFICIENT
Coefficientsa
Standardized
Unstandardized Coefficients
Model
1

B
(Constant)
price of the product (in
rupees)
advertisement
expenditure(in lakhs)

Std. Error

-684.770

168.318

1.312

1.043

11.289

2.785

Coefficients
Beta

Sig.

-4.068

.001

.232

1.258

.225

.748

4.054

.001

a. Dependent Variable: sales of the product (in lakhs)

Interpretation
1. Y (sales) = -684.77 + 1.312 (Price) + 11.289 (Advertisement)- Regression Equation
2. The constant shows influence of other independent variables which are not considered
for the study. If we take more variable then constant value will come down and
forecasting will come to more accurate.(Ideally constant has to be closer to zero)

Anda mungkin juga menyukai