The role of perceived authenticity as the determinant to revisit heritage tourism destination in Penang - Theory and Practice in Hospitality and Tourism Research - page 295

Furthermore, Mohamad et al. (2013) add that
according to a Bucket Theory of Marketing,
“bringing in a new tourist and increasing the rev-
enue from existing customers are said to “fill the
bucket”. They add that tourists stop purchasing
or a decreased arrival for a period of time is por-
trayed as “a hole in the bucket”. They suggest an
understanding of the post-behavioral intention of
the tourist will help to stop money from leaking.
So far, however, insufficient attention and lim-
ited empirical investigation has been given to the
relationship between intention to revisit a heritage
destination, and the crucial factors of perceived
authenticity.

Confirmatory factor analysis (CFA) was con-
ducted prior to Structural Equation Modeling as
proposed by (Gerbing & Anderson, 1988). Findings
from the assessment of fitness for the measurement
model shows that all fitness indexes required in evalu-
ating the fitness of a model are achieved. It was indi-
cated that the result of (CFA) was excellent. Most
of the fitness indexes have achieved the required
level for measurement model. The measurement
model fit was great with p-value

= .000, GFI = 0.927,

CFI = 0.953, RMSEA = .080 and NORMED-

CHISQ 2.656 indicating a strong model fit.

The result of Cronbach alpha, average variance
extracted (AVE) and composite reliability (CR) for
perceived authenticity is (

=

α = 0.829, AVE = 0.648,

CR = 0.901), for satisfaction ( α = 0.880,

3 RESEARCH METHOD AV E = 0.726, CR = 0.888), and for revisit intention

( 0.864).

To validate the assessments of reliability, con-
vergent validity and discriminant validity, a tool
developed by Gaskin (2012) was utilized, namely
the validity master test. Gaskin (2012) proposed
several measures which are suitable to gauge valid-
ity and reliability. The results from the test show
(Table 1) that there were no issues concerning reli-
ability, convergent validity and discriminant valid-
ity. Therefore, the final measurement model can
proceed to the structural model, as the model has
achieved the requirements for a structural model.
Table 2 shows that the result of the struc-
tural model fit was great with p -value

α = 0.863, AVE = 0.614, CR =

The data was collected through a personally admin-
istered questionnaire using 10-point Likert scale,
to tourist traveled heritage destination in Penang.
Convenience sampling was use to obtain partici-
pants. Both international and domestic tourists
have an equal chance of being chosen as respond-
ents; however a screening interview was conducted
to domestic respondents and they were asked a few
questions to determine their origin. This procedure
was compulsory in order to differentiate tourists
from local people and to avoid biases.
Well-established and popular heritage destina-
tions in Penang were chosen as the appropriate
location for data collection, these include, Town
Hall, The Queen Victoria Memorial Clock Tower,
Kapitan Keling Mosque, For Cornwallis, Penang
State Museum, War Museum and Cheong Fatt Tze
Museum.

= .000,

GFI = 0.915, CFI = 0.944, RMSEA = .080 and

NORMEDCHISQ = 2.941 indicating a strong

predictive validity.

4.1 Hypothesis testing
The purpose of this study is to study the effect of
perceived authenticity on revisit intention (H 1 ), the
mediation effect of satisfaction between perceived
authenticities and revisit intention (H 2 ). H 1 was
supported with a significant level of ( p

4 RESULTS

A total of 255 useable questionnaires were collected
from the respondents. Analysis of the data revealed
that majority of the respondents was male (52.9%)
and female (47.1%). The age group were fairly
evenly distributed with the exception of “21-30'”
group consisting of 51 percent, and the group of
“31-39” consisting of 40 percent. 54.1 percent of
the respondents were married and 25.1 percent
were single and the rest were divorced (20.8%). The
respondents travel to heritage tourism destination
had a high level of education 85.1 percent from
tertiary education background, with a moderately
even split of those were manager or executive, 53.7
percent. Eighty percent of the respondents were
from international tourist while 20 percent were
from domestic tourist. Majority of the respond-
ents (80.8%) indicated that this was their first trip
to their heritage tourism destinations.

< .000). H 1

is supported as the unstandardized beta = 0.408,

standard beta 6.254.

As for the mediation effect (H 2 ), the direct
effect of PA on RIN is significant, as

= .065, t -value =

β = .065 and

p .000. As SATIS entered the model as mediator,

the direct effects of PA on RIN is reduced from
.065 to .064 due to the moderation effect of SATIS .
However, it is still significant since the probability

<

Table 1. Validity test.

CR AVE MSV ASV PA SATIS RIN

PA 0.921 0.854 0.217 0.187 0.924

SATIS 0.888 0.726 0.356 0.257 0.397 0.852

RIN 0.864 0.614 0.356 0.287 0.466 0.597 0.784