Page 131 - Special Topic Session (STS) - Volume 1
P. 131
STS419 Zuraeda Ibrahim et al.
used for the e-survey and 1000 copies of questionnaire were distributed to all
companies listed on the Main and ACE markets, identified through the Bursa
Malaysia website in year 2015. The questionnaire is divided into five main
sections. Section A is to identify characteristics and influential factors for
acceptance of Islamic banking products, Section B is to indicate the level of
awareness of respondents of Investment Accounts; Section C requests the
respondents to indicate the extent to which they agree with the concept of
Investment Account and their behaviour regarding investment. The
respondents are required to specify factors that would influence their
acceptance of Investment Accounts based on a 5-point scale, with 5 being
‘Strongly Agree’ and 1 being ‘Strongly Disagree’. Section D requests the
respondents to provide a demographic profile of their organisation; and
finally, Section E requests demographic details of the respondents. Despite
vigorous approaches done on the data collection part; this study could only
managed to get responses from 141 respondents. However, only 44 responses
have been analysed further because questionnaires with missing values to
Likert scale items were removed from the analysis, since structural equation
modeling (SEM) analysis is very sensitive to missing values for Likert scale
items (Hair et al., 2014). The statistical program that was used was Smart-PLS
2.0 M3 (Ringle et al, 2004). This statistical program validates the psychometric
properties of the measurement model and estimates the parameters of the
structural model. Before considering the results from the structural model
(parameters estimate), the quality of the measurement model was first
reviewed. The measurement model (refer Figure 1) was assessed by examining
convergent and discriminant validity, to determine the validity and reliability
of the measurement items, where all the constructs were reflective type
constructs.
3. Results
Table 1 below presents a summary of the results of the convergent validity
for the measurement model. Three indicators were not above the
recommended threshold of the loading, which is above 0.60, and were also
non-significant loading (t < 1.96), whereas all items meet the recommended
threshold loadings. In addition, the AVE for all constructs are more than the
recommended value, which is 0.50, and vary from 0.566 to 0.634, indicating
that the latent variables explain more than half of their indicator’s variance.
Moreover, the CR values for all constructs ranged from 0.714 to 0.947; thereby
these values also exceed the recommended threshold of 0.70. These results
indicate that the measurement model has demonstrated an adequate
reliability of the grouped items. However, in terms of internal consistency
reliability (i.e. Cronbach’s alpha) test, the ICR values ranged between 0.537 and
0.934, which indicates that the internal consistency for the model was at the
120 | I S I W S C 2 0 1 9