Sie sind auf Seite 1von 13

CHAPTER IV

DATA ANALYSIS
I.

DEMOGRAPHIC
With 249 valid results, the researcher decided to input 215 into SPSS software.

Demographic results were also investigated.


Survey questionnaires were spread out not only by offline approach but also via the
internet for convenience, 249 questionnaires were collected. However, there were 34 invalid
surveys. The reason for those invalid results could be because the respondents did not belong to
the target group nor simply they did not finish all the given questions.

There were 117 females and 98 males among the 215 respondents. The gender difference
in this chart does no tell much about the comprehensiveness of the study.

Age differentiation is also belongs to Demographic like gender and help the researcher to
have an overview of the sample size. There were 16 respondents in group under 18, 115
respondents in group 18-25 and 70 respondents represent for group 25-40, 14 respondents in
group over 40. Second group represent for the largest part as 53.5%, third group dominates
32.6%, next is the first group with 7.4% the least respondents group accounts for 6.5%.

The respondents profession data is also collected. Students accounts for most part of the
profession percentage as 54.5%.As the data collected in age part, there was a large number of
respondents in this age.
Second place belongs to office worker group with45.5% .Free employee took only 0.5
%.The numbers reflect that a lot of respondent in this case is students.

As show in the chart above, people with income below 2 mil tend to use Grab taxi the most
with the result is 40%. Its fixed with the data in the Profession part-most user is students. The
second place is people who have income from 5 to 10 mil (26%). People who have income above
10 mil took 17.7% and from 2 to 5 mil took 16.3%.

II.

DESCRIPTIVE ANALYSIS
Table 1
Descriptive Statistics of Brand Awareness

N
I

have

seen

Descriptive Statistics
Minimum Maximum Mean

Std. Deviation

many

advertisements

for

GrabTaxi on Facebook, 215

3.35

1.026

215

2.88

.991

differentiated and easy 215

3.51

.836

3.48

.961

215

3.30

.930

215

3.07

.944

email and mobile phone


message.(Baw1)
Whenever think about
taxi service, I always
think

of

(Baw2)
GrabTaxis

GrabTaxi.
logo

is

to remember.(Baw3)
The specially color help
me

quickly

(Baw4)
I can

realize. 215
recognize

GrabTaxi

among

competing taxi brands.


(Baw5)
I am

familiar

GrabTaxi.(Baw6)

with

Valid N (listwise)

215

Six items of Attitudes towards the Brand variable have fairly high mean values in
comparison with the maximum value of 5. It implies that many respondents concent to the
measurable statements of this variable. In detail, the item GrabTaxis logo is differentiated and
easy to remember (Baw3) has the highest number of responsents agreement with the value of
mean (3.51), followed by Baw4 at 3.48.
Besides, items Baw1, Baw2, Baw5, and Baw6 are respectively ranked as the third to the
sixth position in order of respondents assent. Based on those statistics, it may be concluded that
respondents generally agree that GrabTaxi is popular and easy to perceive.

III.

RELIABILITY TEST
1. Brand Awareness
Reliability Statistic of Brand Awareness
Reliability Statistics
Cronbach's
Alpha
,867

N of Items
6

From the table above, the Cronbachs Alpha value is 0.867 (>0.6). It concludes that the
internal variables of Brand Awareness are closely related to each other.

Item Total Statistic of Brand Awareness


Item-Total Statistics
Cronbach's

baw1
baw2
baw3
baw4
baw5
baw6

Scale Mean if

Scale Variance if

Corrected Item-

Alpha if Item

Item Deleted
16,25
16,72
16,09
16,12
16,30
16,53

Item Deleted
13,149
14,064
14,907
14,309
13,614
13,456

Total Correlation
,715
,602
,606
,590
,735
,747

Deleted
,835
,856
,855
,858
,832
,830

The above table shows that all Corrected Item- Total Correlation are higher than 0.3, no
more item is deleted. The Cronbachs Alpha if Item Deleted column also demonstrates that it is
unnecessary to eliminate any more items from Brand Awareness.

IV. FACTOR ANALYSIS


1. Factor Analysis for Independent Variables
Firstly, factor analysis was run for all independent variables to check the construct validity of
the research model established in the literature review.
In this research, in order to get the final results, two rounds of factor analysis were required
to run because in the first round, there were some variables that needed deleting.
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

Bartlett's Test of Sphericity

.885

Approx. Chi-Square

1929.714

df

171

Sig.

.000

As can be seen from the table, the KMO value was 0.885 which was higher 0.5 and the
significant value of Bartlett's Test of Sphericity was 0.000 which was lower than 0.05. These
values reached the standard levels and were considered to be accepted.

Rotated Component Matrixa


Component
1

baw1

.812

baw6

.787

baw5

.781

baw2

.752

baw4

.667

baw3

.617

pq4

.816

pq3

.752

pq2

.747

pq5

.725

pq1

.680

bl4

.809

bl5

.703

bl2

.659

bl3

.567

bl1

.520

bas3

.828

bas2

.823

bas1

.798

Extraction Method: Principal Component Analysis.


Rotation Method: Varimax with Kaiser Normalization.
a. Rotation converged in 6 iterations.

After the factor analysis,19 variables were allocated into four components. Each component
included items that had high loadings. This result also indicates that there are 19 items grouped
into 4 independent variables.

2. Factor Analysis for Dependent Variable


After defining the construct of independent variables in the research model, factor analysis
was run again with the aim of testing the validity of the structure of dependent variables.
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

Bartlett's Test of Sphericity

.656

Approx. Chi-Square

110.105

df

Sig.

.000

As can be seen from the table, the KMO value was 0.822 which was higher 0.5 and the
significant value of Bartlett's Test of Sphericity was 0.000 which was lower than 0.05. These
values reached the standard levels and were considered to be accepted.
Component Matrixa

Component
1
pi2

.823

pi3

.794

pi1

.742

Extraction
Principal

Method:
Component

Analysis.
a.

extracted.

components

After running factor analysis for the four dependent variables, there was only one
component extracted. As a result, Varimax rotation could be applied.
As can be seen from the Component Matrix table, all of the three variables were grouped
into one component, and this component could be named as Purchase Intention because they had
exactly the same variables.
The factor Purchase Intention could be summarized as below:
No.
1
2
3

Statement
Purchase Intention
I will use GrabTaxi service in the near future
I would use GrabTaxi rather than any other service
I have intention to use GrabTaxi service for the future

Code
Pi
Pi1
Pi2
Pi3

In conclusion, after running factor analysis, the research model stated in the literature review
did not change. It still had four independent variables which were Brand Awareness, Brand
Association, Perceived Quality and Brand Loyalty and one dependent variable which was
Purchase Intention.

V.

CORRELATION & MULTIPLE REGRESSION

1. Correlation Testing
After allocating all variables into suitable factor, factor scores were calculated in order to
become new input for correlation testing and multiple regression analysis. In this study, factor
scores were calculated by taking the averages of all variables in one factor. Each average value
was then representative for each factor in the research model and was entered into SPSS formula
to estimate the results of correlation and regression.
Cooper and Schindler (2006, p.707) defined that correlation is the relationship by which
two or more variables change together. In this study, correlation testing was used to check
whether there was a probability of a relationship between the four independent factors and
Purchase Intention.
To be more specific, Pearson Correlation was applied. The value ranging from -1 and +1

can become an indicator of the strength of correlation between variables. The higher the Pearson
Correlation value is, the stronger the strength of the relationship can be.

Correlations
BAW
BAS
Pearson Correlation
1
.524**
BAW Sig. (2-tailed)
.000
N
215
215
Pearson Correlation
.524**
1
BAS
Sig. (2-tailed)
.000
N
215
215
Pearson Correlation
.254**
.339**
PQ
Sig. (2-tailed)
.000
.000
N
215
215
**
Pearson Correlation
.403
.483**
BL
Sig. (2-tailed)
.000
.000
N
215
215
Pearson Correlation
.477**
.490**
PI
Sig. (2-tailed)
.000
.000
N
215
215
**. Correlation is significant at the 0.01 level (2-tailed).

PQ
.254**
.000
215
.339**
.000
215
1
215
.543**
.000
215
.437**
.000
215

BL
.403**
.000
215
.483**
.000
215
.543**
.000
215
1
215
.587**
.000
215

PI
.477**
.000
215
.490**
.000
215
.437**
.000
215
.587**
.000
215
1
215

As can be seen from the Correlation table above, the Pearson Correlation value shows
that there could exist the relationship between Purchase Intention with each of the independent
factors. Based on the above result those factors BAW, BAS, PQ, BL had coefficients 0.477,
0.490, 0.437, 0.587 were indicators to predict that the strength of relationship of each pair of
dependent and independent variable.
Besides, all of the significant value was 0.000 which was lower than 0.05 also proved that
these values had statistical significance.
2.

Multiple Regressions

Finally, the multiple regression analysis was run. Based on this test, the relationships between
Purchase Intention and Brand Awareness, Brand Association, Perceived Quality and Brand
Loyalty could be investigated. Besides, multiple regression analysis could help identify and
measure the influence of each independent variable on the dependent one.

Model

Model Summary
R Square
Adjusted

Square
1
.668
.447
.436
a. Predictors: (Constant), BL, BAW, PQ, BAS
b. Dependent Variable: PI
a

R Std. Error of the


Estimate
.43270

With the first table, Model Summary, R Square was the most significant result. This value
indicates the degree to which the independent variables could explain the dependent variable. In
this study, R Square was 0.447 which is a quite low result. In other words, the research model
with four factors can interpret 44.70 % of the variability of Purchase Intention.
As for the 55.30 % which was not explained by the model in this study, it could be
elucidated by other factors not including in the model. Such factors can be Perceived Benefits
(Convenience, Distribution, Price), Demographic or perhaps other factors not mentioned in this
study.
ANOVAa
df
Mean Square
4
7.930
210
.187
214

Model

Sum of Squares
Regression
31.720
1
Residual
39.319
Total
71.039
a. Dependent Variable: PI
b. Predictors: (Constant), BL, BAW, PQ, BAS

F
42.353

Sig.
.000b

The next important value needs considering was the Significant level of the regression in the
ANOVA table. As can be concluded from the table above, the research model had statistical
significance due to low significant level valu
e (0.000<0.05).

Model

Coefficients
Unstandardized Coefficients
Standardized
B

Std. Error

Coefficients
Beta

Sig.

(Constant)
.964
BAW
.170
1
BAS
.110
PQ
.125
BL
.354
a. Dependent Variable: PI

.197
.048
.044
.055
.069

.217
.162
.140
.345

4.882
3.528
2.507
2.273
5.142

.000
.001
.013
.024
.000

Lastly, the Coefficients table gives the final results of the research as well as answers the
research questions stated at the beginning. It confirms which factors in the research model and
the four hypotheses had influence on Purchase Intention of GrabTaxi customer.
All factors were still included to analyze because they had low significant level (less than
0.05), which implied that they could reach statistical significant of this study. In other words,
these factors were proved to have influence on Purchase Intention.
After identifying which factors having impacts on the dependent variables, in order to
measure the level of impact, the researcher investigated the Standardized Coefficients Beta.
The higher the level of Standardized Coefficients is, the higher the level of influence that each
independent factor has on Purchase Intention.
As a result, the size of the influence level can be identified by looking at the
Unstandardized Coefficients. Furthermore, the sign before each Unstandardized Coefficient can
imply the direction of the relationship. If the sign of each factor is positive, it suggests that when
increase that factor, Purchase Intention also increases.
Based on this rule, Brand Loyalty can be considered as the factor that has the most
influence on Purchase Quality because it had the highest Unstandardized Coefficient which was
0.354. Due to its positive sign, Brand Loyalty and Purchase Intention had the same dimension in
their relationship.
If other factors were zero and the Brand Loyalty increased by one unit, Purchase
Intention also increased by 0.354 unit. Next to Brand Loyalty, the second influent factor was
Brand Awareness with B equal to 0.170 and the third place was Perceived Quality with B equal
to 0.125 and the last one was Brand Association with 0.110 of B.

Based on the analysis above, the regression equation can be written as:
PI = 0.964 + 0.354BL + 0.170BAW + 0.125PQ + 0.110BAS
The coefficient standing before each factor represents the size of the influence that it has
on the dependent variables. Besides, the sign before each coefficient implies the direction of
relationship between each pair of independent and dependent variables.
Considering four hypotheses established before, from the multiple regression analysis,
these hypotheses can be confirmed as below.
Hypotheses
H

There is a positive impact of brand awareness towards

1
H

customers purchasing intensions


There is a positive impact of brand association towards

2
H

customers purchasing intensions


There is a positive impact of perceived quality towards

3
H

customers purchasing intensions


There is a positive impact of brand loyalty towards

customers purchasing intensions

Confirmati
on
Support
Support
Support
Support

Das könnte Ihnen auch gefallen