• Tidak ada hasil yang ditemukan

Hypothesis Testing Hs

Dalam dokumen PDF utpedia.utp.edu.my (Halaman 110-115)

CHAPTER I CHAPTER I INTRODUCTION

CHAPTER 3 METHODOLOGY

S. D. Cronbach's a Items Factor

4.1.6 Hypothesis Testing Hs

H5: The availability of resources and skills set influence upon the usability of OARS.

10% of the respondents claimed to have no Internet connection available; 30%

claimed to have no or little knowledge of online reservation systems; the remaining 60% reported to have no credit or debit card available.

The relationship between available resources and the usage of online reservation systems is observed to haver= 0.32 with p<O.Ol. This shows that the relationship is mildly significant; the hypothesis is accepted.

4.2 Phase 1: Users' Flexibility in Terms of Compromising on SQAs

This study is to address the 2"d research question.

RQ2: To what extend flexible users can compromise with service quality attributes of Online Airline Reservation Systems?

4.2.1 Assumptions in ANOV A to Test H6

Before performing ANOV A, a very basic assumption of ANOV A was checked, i.e.

absence of outliers. Box-plot of the sample distribution was examined since it is a useful standard in data interpretation, reveals data symmetry, skewness and the presence of outliers. Moreover, it also facilitates in comparing more than one population without knowing anything about the underlying statistical distributions of those populations.

4.2.1.1 Box-and-Whisker Plot (H6)

In case of satisfaction level with existing SBTs, respondents who reported that they 'can compromise' on SQAs of the airline have a median at 3 (black line) as shown in Figure 4.1.

Satisfaction Level With existmg Self Bookmg Tools (SBTs)

I -H 1ghly Sa11sfiod, 5-H 1ghly Dtssallsfied 6

5

5 ....

Cll eo

"

"

a

4

-5 3: 3

]

j

2

Cll i 1

0

N • ~ n ~

C1n Comprom1se May Com pmmiso Cannot Com prom1se Users' fleJUblllly 1n compromISing on sernce q u.aluy annbul es

Figure 4.1: Box Plot showing Satisfaction Level with Existing SBTs

This represents neutral satisfaction level and at the same time indicates 50% of the data is greater than this value. Users' with any lesser satisfaction with existing SBTs are represented everything above median black line, while the users with higher satisfaction are represented everything below median black line. As shown by the top 'whisker', this group has greatest values but no outliers. Hence the data is normally distributed.

In case of satisfaction level with existing SBTs, respondents who reported that they 'may compromise' on SQAs of the airline have a median at 2 (black line). This represents high satisfaction level and at the same time indicates 50% of the data is greater than this value. Users' with any lesser satisfaction with existing SBTs are represented everything above median black line, while the users with higher satisfaction are represented everything below median black line. As shown by the top 'whisker', this group has greatest values and an outlier. The majority of the data is normally distributed.

In case of satisfaction level with existing SBTs, respondents who reported that they 'cannot compromise' on SQAs of the airline have a median at 2 (black line). This represents high satisfaction level and at the same time indicates 50% of the data is greater than this value. Users' with any lesser satisfaction with existing SBTs are represented everything above median black line, while the users with higher satisfaction are represented everything below median black line. As shown by the top 'whisker', this group has greatest values but no outliers. Hence the data is normally distributed.

4.2.1.2 Means Plot (H6)

The means plot as shown in Figure 4.2 shows that there is apparently an enormous difference between the satisfaction level of the three respondents groups, which appears not be the actual case. Therefore as a follow-up, the same results will be analyzed in a different chart to see the difference between the groups.

Satisfaction leve 1 V.'ith existing Self Booking Tools (SBTs)

I - lilghly Satisfrcd, ~ Highly Drssattstied

32r---~

§ 2.4

~ §

:1] 2.2

2.0 .1---:-1

Can Compromrse May Compromrse Cannot Compromise

Users' flexibrlrty m compromismg on servrce quality attributes

Figure 4.2: Means Plot on Satisfaction Level with Existing SBTs

4.2.1.3 T-Test (H6)

In this case the three groups are significantly different usmg a t-test (t=36.760, df=169, p=O.OOO) as shown in Table 4.3. 95% Confidence Interval (CI) is probability that the interval contains the true mean.

Table 4.3: T-Test on User's Flexibility with SBTs TestValue=O

- - - - · -··-· - - - -

t df Sig. (2- Mean 95% Confidence Interval tailed) Difference of the Difference

- - - . - -

How flexible

36.760 169 .000 2.07 Lower Upper

you are 1.96 2.18

4.2.1.4 Error Bars (H6)

The same results are now reproduced in the error bars, with 95% confidence intervals to have an idea of the variation in sample distribution. CJ of the groups is closely related to the results of the analysis of variance for these groups. The confidence interval for each graph below shows a linear pattern of the sample distribution which otherwise appeared to be showing huge variations in the simple means plot.

In error bars we intend to see if the mean of one group is included in the confidence interval of the other two groups - if so then there is likely no difference among the groups. Moreover, it is not relevant whether the error bars 'overlap' but whether the mean of one group 'overlaps' with the error bars of the other. The confidence intervals can overlap by as much as 25 percent of their total length and still show a significant difference between the means for each group.

Error Bar: Satisfaction level with existing Self Booking Tools (SBTs) I- Highly Satisfied, 5- Highly Dissatisfied

~ 3.5

"'

"'

~

3.0

1

~ 2.5

"

.Q

s

2.0

"' • I

I I

1.5

'0 Can Compromise

"

~·a Cannot Compromise

"

May Compromise

Users' flexibility in compromising on service quality attributes

Figure 4.3: Error Bar on Satisfaction Level with Existing SBTs

In Figure 4.3, 95% CI tells us that the satisfaction level of existing SBTs for the users who "can compromise" on SQAs of the airline is probably between 2.7 and 3.35, with group mean of3. Likewise, for users who "may compromise" it is probably between 2.4 and 2.83, with group mean of 2.6, and for users who "cannot compromise" it is probably between 1.8 and 2.3, with group mean of2.07.

The group means of users' who 'may compromise' shares a degree of confidence interval overlap with users who 'can compromise', thus the two groups may not necessarily be different from one another. Moreover, the group mean of users' who 'cannot compromise' does not share any degree of confidence interval overlap with either of the two groups, therefore, this particular group appears to be significantly different from the rest of the sample population. However, post-hoc tests can confirm this.

Dalam dokumen PDF utpedia.utp.edu.my (Halaman 110-115)