International Journal of Trade, Economics and Finance, Vol. 5, No. 5, October 2014
Reengineering Tax Service Quality Using a Second Order
Confirmatory Factor Analysis for Self-Employed
Taxpayers
Siti Normala bt Sheikh Obid and Bojuwon Mustapha
the self-employed taxpayers. (However, most significant
critics of the Serviqual model (SM) developed by V.A.
Zeithamal et al. [6] was criticized based on the number of
variable/dimension with their substantial stability when used
in different context).
Thus it was suggested that the items of Serviqual model is
a second order factors with multidimensional first order
variables [7], [8]. The advantage of employing a second order
CFA is numerous [9]. Given that its results are more
parsimonious with the use of smaller parameter to test the
hypotheses. Hence, a higher order factor represents the
pattern of the association that exists among the first order
[10]. Also, compound dimension structures are simply
interpretable with the use of a second order measurement
model [11]. The psychometric properties includes reliability,
convergent, and discriminant validity on the usage of online
tax system for tax administration effectiveness in Nigeria.
Abstract—Reengineering tax service quality using second
order confirmatory factor analysis, namely, responsiveness,
reliability, in formativeness, assurance and usability is viewed
as an enabler of tax administration effectiveness in collecting
government revenues. This paper examined the psychometric
properties of tax service quality items for tax administration
effectiveness on the self-employed taxpayers. A total of 181
received and usable data was collected and analyzed. The
findings shows the goodness fit indices are adequate with the
model fit indices showing the chi-square value of 215.334, DF
=146, p-value = 0.000, normed chi-square CMINDF) =1.475,
CFI =0.967 and RMSEA = 0.051. In other words, the findings
show that the factors help the tax administrator for providing
effective tax system.
Index Terms—Tax service quality, second-order CFA factor,
Bayesian and self-employed taxpayers.
I. INTRODUCTION
Due to the success of the pilot study in the United Stated,
most of the developed and developing nations have adopted
the use of online tax system for tax administration
effectiveness. This includes United Kingdom, Canada,
Taiwan, Australia, Malaysia, Singapore, Kenya, Thailand
and South Africa among other countries [1]. Also the
implementation was as a result of the benefit time saving,
convenience and cost reduction for both tax compliance and
tax administration [2].
However, tax service quality is the overall evaluation of the
quality of service being provided by the tax authority with the
use of information technology [3]. Parasuraman [4] defines
service quality as the scope to which information is conveyed
in line with the taxpayer’s expectation. Tax service quality
can also be interpreted as the use of resources by the tax
administration sector to achieve the b e s t output of the tax
system. In addition, [5] added that the general factors
contributing to the tax service quality level that involved the
availabilities of service and infrastructure facilities.
The five-factor solutions of tax service quality are more
towards evaluating the factors toward effective tax system.
Hence, this paper examines the psychometric properties of tax
service quality items for tax administration effectiveness on
II. LITERATURE
Initially, tax service quality is considered as a limited
paradigm that is challenging to understand [12]. The
development of the SM was based on the integration of
theoretical and empirical study result that considered tax
service quality as a multidimensional construct. This
construct used in the above study consists of five dimensional
variables, namely, tangibles, reliability, responsiveness,
assurance and empathy. Since then the application of SM in
the research had gained momentum to measure the service
quality in the area of education and legal profession [13],
[14]. Without any geographical challenges, the use of online
tax system has been viewed as an advantage in the filing of
tax return through the emergence and availability of
information technology [15]-[17].
The adoption of online tax system by taxpayers is
increasing because it is convenience, cost effective and time
saving [18]-[20]. Accordingly, the use of online tax system as
a new information technology is to achieve an effective tax
admiration system. Chang et al. [21] found that the effective
use of online tax system are related to the tax service quality
provided by the tax authority which also have effect on the
cost reduction from the taxpayers and tax authority.
Despite the benefits of online tax system the taxpayers still
faces some challenges with the service provided. The use of
Serviqual items was explored and concluded to be
problematic by [22]. The use of Serviqual model for the tax
service quality in this paper are composed of a five-factor
solution construct to include responsiveness, reliability,
Manuscript received April 25, 2014; revised July 15, 2014.
Siti Normala Bt Sheikh Obid is with the Department of Accounting
faculty Economic and Management Science, International Islamic University
Malaysia (e-mail: drcnormala@yahoo.com).
Bojuwon Mustapha is with the Faculty of Economic and Management
Science, International Islamic University Malaysia, P.O Box 10, 50728
Kuala Lumpur, Malaysia (e-mail: bojuwon@live.com).
DOI: 10.7763/IJTEF.2014.V5.410
429
International Journal of Trade, Economics and Finance, Vol. 5, No. 5, October 2014
are developed: (1) tax service quality is influenced by
five-factor solution. (2) Each of the items of tax service
quality has a non-zero loading on the hypothesed factor. The
remaining sections of this paper are literature review of tax
service quality, method, result, conclusion recommendation
and direction for future study. The next section discussed the
methodology of this paper.
informativeness, assurance and usability from the perspective
of the self-employed taxpayers crucial outcome of the
analysis to evaluate the usage of online tax system to the
measure of tax administration system.
The items used in this paper are scale to the tax service
quality in the context on online tax system with the tax
administration sector in Nigeria. The items were measured
with the expectation of the self-employed taxpayers to the tax
authority to evaluate the quality of the service provided. The
evaluation of the items was adapt and modified to the tax
service quality context [23]. Santos [3] described service
quality as perception of users in the evaluation of a system
based on their experience in it uses for a particular period of
time. Recently, Pantouvakis et al. [23] proposed the scale of
five key variables such as tangible, reliability, assurance,
responsiveness and empathy in the context of online book
trade. Furthermore, the empirically tested Serviqual model by
Parasuraman [4] and Zeithamal et al. [24] examined the
multidimensional items scale with fundamental service
quality aspect of five dimension which is applied and used in
this paper.
In addition, Ulrike Bauernfeind et al. [25] examines the
extent of tax service quality measurement using structural
equation modeling to find the existing relationship between
efficiency, information quality, security and usability. The
result indicates that responsiveness is the most influencing
factors in the process of online tax system. Based on this
study there are still areas of loopholes in the tax service
quality system where the conceptual model needs to be
improved. Santos [3] conducted an exploratory study to
understand the determinant of tax service quality in United
Kingdom school of business which consist of six to ten
members focus group. The finding shows that reliability has
the highest influence on the determinant of the tax service
quality.
The hypothesed model of tax service quality to the items in
the administered items are explained with the five factors i.e.
responsiveness, reliability, informativeness, assurance and
usability. According to Hair et al. [26] the CFA is a direct
strategy that researcher postulates in a particular model that
poised a set of relationships among variables with the
application of structural equation modeling (SEM) to
evaluate the adequacy of the conceptual model. Furthermore,
for the support of either the model fit the data nor the data fit
the model, was based on the initial model through the
application of the model comparisons with the Bayesian
analysis in the finding section of this paper.
Furthermore, Santos [3] investigates the variables that may
serve as a determent on tax service quality. The discussion
above is in line with the study by Parasuraman [4] who stated
reliability as a dimension which has the highest influence on
the tax service quality factor. Stiglingh [27] proposed a
theoretical framework for the tax service quality to include
the tax practitioners in the tax authority setting, using the
critical techniques of quantitative approach. Also Zeithamal
[24] conceptualize the construct by measuring the service
quality delivered through the website. These variables are
most applicable to the present study because an increase in
the quality of the variable will provide a significant output for
the tax administration system. On these notes, the hypotheses
III. METHODOLOGY
The components factor structure of a 19 items of quality
related to the measurement of tax service quality for tax
administration
effectiveness
was
examined.
The
measurement of tax service quality are composed of five
subscale
variables;
responsiveness,
reliability,
informativeness, assurance and usability. The items
comprising subscales were adapted from existing model of
Serviqual by Parasuraman [4].
The full set of the items are given a caption for in table 1 of
the exploratory analysis to identify the items that loaded
under each of the component factors. A seven point
likert-scale ranging from 1- strongly disagree to 7- strongly
agree was used to achieve the reliability scores on the scale
representing tax service quality. Data were collected from
self-employed taxpayers that registered with the federal
Inland Revenue service board.
Tax service quality is imprecise because it cannot be
directly measured. It is only reflected in a theory, which is
measured by other latent variables such as responsiveness,
reliability, informativeness, assurance and usability in the
context of this study. However, for the researcher to measure
this variable there is a need to undertake some statistical
analysis process. Therefore, the studies quantify tax service
quality by applying a statistical analysis on the items. Thus,
the tax service quality was measured via the use of the latent
variables with the applicability of administered items stating
the operational meaning of the variables. The sampling
technique is explained in the following section. The items
and the data were screened in order to test the second order
confirmatory factor analysis techniques in addressing the
model hypothesized.
A. Sampling
A random sampling technique was employed on the
registered taxpayers with the Federal Inland Revenue Service
Board. In general a total number of 500 self-employed
taxpayers were administered, said to be a representative from
the total population of 1,200 [28]. From 500 questionnaires,
200 were received and the usable one is 181. This account for
59% of the respondent response rate which is considered
adequate for the purpose of this paper [29].
During the procedure of data screening for outliers,19
dataset were removed as a result of Mahalanobis distance
values more than the χ2 value (χ2=40.00; n=12, p<0.001)
leaving a final 181 dataset to be analyzed. Statistical validity
tests and analysis were then conducted such as reliability test
and validity tests using second order CFA for construct
validity and discriminant validity for multicolinearity
treatment, composite reliability, and average variance
extracted, testing the fit for the hypothesized CFA model and
430
International Journal of Trade, Economics and Finance, Vol. 5, No. 5, October 2014
A. Assessing Validity and Reliability
In assessing the reliability and validity, Hair et al. [26]
suggested that the assessment is based on the dependability
between the dimensions of the variables used in the analysis.
Then the data were run using internal consistency test based
on a complete scale used with the reliability value, namely,
the Cronbach’s alpha and the overall reliability of individual
variables values. The result shows that the overall value of the
reliability values is higher than the threshold value of 0.70,
suggested by Hair et al. [26]. The details of the coefficient
value are shown in Table I. The result indicate that the
Cronbach’s alpha coefficient value of all the items
administered were acceptable and reliable for the analysis
with the overall value of 0.937.
The individual reliability scale value of the items based on
component are also analyzed which shows that
responsiveness, reliability, in formativeness, assurance and
usability, where the reliability scale value of Cronbach’ alpha
shows 0.854, 0.858, 0.859, 0.854 and 0.877, respectively.
We then employed construct validation to test on the
effectiveness on the application of maximum likelihood
(ML), for the second order of CFA. (The CFA is used to
hypotheses the model based on the underpinning theory and
model which will be detailed in the next sub section). Table I
also exhibits the mean and the standard deviation values of
the items which indicate that quality tax service will lead to
tax administration effectiveness, and hence increase in
revenue generation.
the revised model.
B. Instrumentation
The items focused on the five latent variables of tax
service quality that revolved around the tax administration
effectiveness based on the service provided to incorporate
and to internalize them as discussed in the literature review.
Every variable consists of several items. Where
responsiveness have4 items, reliability with 3 items, and
informativeness with 4 items, assurance 4 items and usability
with 4 items, respectively. Hence, these data were keyed in
using SPSS version 20.0 for component identification and
second order confirmatory factor analysis for the hypothesis
testing using AMOS 20.0 version.
IV. DATA ANALYSIS
An exploratory factor analysis (EFA) and second order
confirmatory factor analysis (CFA) was used to examine the
factor structure of the 19 items through the principle
component analysis if the EFA scales. All the 19 items were
subjected to EFA which generated five factor structure
components and the CFA was used to assess the fit indices of
the data to the model whether tax service quality is a factor
solution or multidimensional solution. We employed
multivariate analysis where the result suggested that the items
are not homogeneous (e.g., mean correlation among the a
priori scales or average item-total correlation).
TABLE I: THE INTERNAL CONSISTENCY OF THE CONSTRUCT EXTRACTED FOR TAX SERVICE QUALITY
Observed
Cronbach
Variable
Mean
SD
Alpha
Std loading
SMC=R2
AVE
CR
res3
re4
5.36
5.24
1.125
1.058
.575
.571
.727
.783
.819
.873
.529
.612
.671
.763
.775
.601
.762
.779
.805
.764
.581
.606
.647
.953
.909
.857
.734
.735
.541
res1
5.19
1.255
ree2
rel1
rel2
rel3
5.18
5.41
5.39
5.03
1.236
1.059
1.020
1.159
inf4
inf2
5.33
5.45
1.100
1.132
inf3
inf1
5.31
5.40
1.186
1.153
ass3
ass4
5.56
5.26
1.097
1.275
ass5
5.52
1.009
ass1
usb1
5.61
5.10
1.068
1.493
.587
.344
.815
.665
usb2
5.31
1.217
.870
.758
usb3
usb4
5.39
1.223
1.408
.835
.709
.697
.503
5.10
0.854
.758
.757
0.858
0.859
0.854
0.877
.60
.86
.68
.86
.60
.86
.63
.87
.65
.88
.584
Note: Res= responsiveness, Rel= reliability, Inf =Informativeness, Ass= Assurance and Usb= Usability
same construct, the composite reliability (CR) is applied.
The finding on AVE for responsiveness, reliability, in
formativeness, assurance and usability is 0.60, 0.68, 0.60, 0.63
and 0.65, respectively. While CR for responsiveness,
reliability, in formativeness, assurance and usability is 0.86,
0.86, 0.86, 0.87 and 0.88, respectively. The result shows that
AVE is above the threshold of 0.50 and CR is above 0.70.
B. Discriminant Validity of the Construct
The average variance (AVE) was carried out in order to
test the discriminant validity of the five components. The
result suggested that the higher the correlation square
supports the discriminant validity [30]. To test whether the
component items are interrelated with each other within the
431
International Journal of Trade, Economics and Finance, Vol. 5, No. 5, October 2014
759, res4 <-- res = 756, rel1 <-- rel = 819, rel2 <--rel =874,
rel3 <--rel =.775, ass1 <--ass =.587, ass2 <-- ass = .735, ass3
<-- ass =.953, ass4<--ass =.857, inf1 <-- inf =.763, inf2 <--ass
=.779. inf3 <-- inf = .805, inf4 <--inf = .763, usb1 <-- usb
=.835, usb2 <--usb = .899, usb3 <--usb =.795 and usb4 <-usb=.659 respectively. The next section explained the
discriminant validity of the model based on the data.
This indicates that the selected component have a good
measure with each of the items loading on a specific tax
service quality variables. In other words, the factor’s loading
of each of the variables have a significant loading that
indicate a very good and comprehensive model for the tax
service quality as suggested by [26] and [31].
V. SECOND ORDER CONFIRMATORY FACTOR ANALYSIS
(CFA)
The application of second order confirmatory factor
analysis is to test the construct validity of the administered
survey items. This indicate how each of the construct
component are been explained by the latent variables [26].
Thus, the high correlation of the items with the same
construct component is relatively high which shows the
evidence of construct validity of the items.
The regression weights based on the factor analysis are
correlated and the Square Multiple Correlation (SMC) of the
items specified by the construct contributed to the construct
validity of the model. Fig. 1 below show the conceptual
model with the five component variables and the 19 items
used for getting the perception of the respondent on the tax
service quality for tax administration effectiveness
A. Initial Model of Tax Service Quality
From the initial model of this paper the CFA results on the
hypothesized model shows that there is an inter-correlation
and significant regression weight and there is no offending
estimate issue on the loading of the items. An offending
estimate occur when there is inter-correlation value and
regression weight less than 0 and greater than 1 on the
inter-correlation of the items.
In addition, CFA result as exemplified as in Table II. It
was observed that the factor loading of first component,
responsiveness has a loading value range from 0.73 to 0.78.
The second component reliability has a loading value ranging
from 0.82 to 0.87. The third, in formativeness and the fourth,
assurance component shows a loading value of ranging from
0.76 to 0.81 and 0.59 to 0.95. The fifth component, usability
has a loading value ranging from 0.71 to .87, respectively.
(This regression weight estimates of the observed latent
variables are higher than the threshold 0f 0.50, as has been
suggested [32].
Thus, the regression weight based on each of the
component variable are responsiveness with value ranging
from 0.50 to.066; reliability with value range from 0.60 to
.076; in formativeness with the regression weight value
range from0.58 to 0.61; assurance with value ranging from
0.35 to 0.91;and finally, usability regression weight value
range between0.51 to 0.76. The above regression weight
value are acceptable due to higher threshold then suggested
by [31], which is 0.20. Thus, the high correlation of the items
with the same construct component is relatively high which
shows the evidence of construct validity of the items.
The items that best explained component is the one with
the highest loading on the same component which is res <-->
ifs .801. More so, the correlation coefficient among the five
variables are acceptable with group number and the default
model of res1 <-- res= .727, res2 <--res = 783, res3<-- res =
Fig. 1. Five-factor variable of second order CFA model for tax service
quality.
B. Comparison of Bayesian and Confirmatory Factor
Analysis
The maximum likelihood estimation (ML) of Likert scale
items have significant effects on the use of non-normally
distributed data with categories of variables less than 4
responses and a sample size less than 200, i.e. small sample
size [33]. This paper is found to be under this type of issue,
where the total number of cases is 181 which is less than 200
minimum cases, as suggested by Byrne (2009).
TABLE II: BAYESIAN CONFIRMATORY FACTOR ANALYSIS (CFA)
Construct items
ML
Bayesian
res2
res3
res4
inf2
<--<--<--<---
res
res
res
ifs
1.017
1.278
1.259
1.049
1.004
1.265
1.247
1.051
ass2
<--- ass
1.453
1.449
inf3
usb2
<--- ifs
<--- usb
1.137
.874
1.136
0.870
usb3
usb4
<--- usb
<--- usb
.840
.824
0.835
0.822
ass3
<--- ass
1.491
1.480
rel2
rel3
<--- rel
<--- rel
1.023
1.034
1.015
1.026
inf4
<--- ifs
1.048
1.049
ass4
<--- ass
1.417
1.414
Thus, Arbuckle [34] suggested to use Bayesian estimation
(BE) for the re-affirming the earlier CFA conducted in the
432
International Journal of Trade, Economics and Finance, Vol. 5, No. 5, October 2014
indicates that the five factor solution of tax service quality
model fit the data and there were no significant difference
between data and the model taking the measure consideration
of modification indices and the rule of thumbs in the
confirmatory factors analysis techniques.
There is also evidence of convergent, divergent and
discriminant validity for the five factor solution of tax service
quality as the value of the item ranges from .71 to .95. Based
on the above explanation, the goodness of fit index findings
recommends that the model did not generate any offending
estimate of the covariance matrix.
Basically the five factor solution of tax service quality fit
the measure of tax administration effectiveness using second
order confirmatory factor analysis for the model. For the
items of tax service quality that has a non-zero loading the
result indicates all the items are statistically significant with
value higher than 0.5 as shown in table 2.0. The loading are
significant with responsiveness .902, reliability .717,
informativeness .891, assurance 741 and usability .727. The
significant loading show that when tax service quality goes
up by one, standard deviation each of the variable will goes
up by the significant loading indicated above.
Furthermore, the RMSEA shows significant divergences
amongst the direct covariance and the indirect matrices by
supporting the degree of good fitness of the data. One of the
main aim of this paper is to examine and validate the value for
tax service quality TSQ framework as suggested by
Parasuraman (2005) where the TSQ quality is an important
variable in measuring the effectiveness of tax administration
system.
It is believe that the findings are significant and relevant to
tax authority and practitioners such as tax management to
embrace the quality tax service FIRS) it will led to increase in
revenue generation from self-employed taxpayers. In addition
the techniques used to evaluate the tax service quality is
proven to be psychometrically sound against the five factor
solution,
namely,
responsiveness,
reliability,
in-formativeness, assurance and usability. The findings show
that the tax service quality model is an important determinant
for measuring tax administration effectiveness that provide an
empowering relevant to the tax administration system
especially at the federal Inland Revenue Service Board
Nigeria.
The results could not be generalized because the studies
only focus onto a specific group of taxpayer, self-employed
taxpayers. Hence, for future research the study should take
into consideration of other types of taxpayers such as
corporate taxpayers and individual taxpayer’s. To
comprehend the present result, future study could apply other
techniques of analysis such as MPLUS and LISREL.
data analysis. (Bayesian analysis was further analyzed for the
appraisal of unstandardized loading regression weights).
Hence, the finding shows that little differences observed
between the loading value between ML estimation of CFA
and the Bayesian estimation analysis. This application justify
that the CFA with the application of ML estimation is
acceptable with the value of the re-specified model by the
comparative analysis, detailed as in Table II.
VI. DISCUSSION AND CONCLUSION
The initial specification of the model as indicated in Fig. 2
shows five-factor variables of Second order CFA model of
tax service quality. The figure shows the conceptual model
with the five component variables and the 19 items used to
attain the perception of the respondent on the tax service
quality for tax administration effectiveness. The CFA results
on the hypothesized model shows that there is an
inter-correlation and significant regression weight and there
is no offending estimate issue on the loading of the items.
Based on the same model the standard multiple
correlations for the responsiveness in this case are
endogenous. The first variable is estimated since tax service
quality cannot be directly measure. The result shows that
responsiveness explains 0.86 percent of variance while 0.14
percent of the variance is unexplained. The standard
regression weights for the itemized variable which are also
called the factor loading indicates that when item is 1 (res 1)
(res2), (res3) and (res 4), the standard deviation of
responsiveness goes up to .73, .78, .76 and .76, respectively.
The second result is the reliability constructs which
explain 61 percent variance and 39 percent of the variance is
unexplained. The standard regression weights for the
itemized variable indicate that when the item 1 (rel. 1) (rel. 2)
and (rel. 3) goes up by one, the standard deviation of
reliability will go up by .81, .87 and .78, respectively. Also
for the third variable, informativeness construct explain 79
percent variance and 21 percent is unexplained. The standard
regression weights for the itemized variable indicates that
when item 1 (inf 1) (inf 2), (inf 3) and (inf 4) goes up by one
then the standard deviation of reliability go up by .76, .78, .80
and .76, respectively.
The fourth variable, assurance, 64 percent of variance
was explained and 36 percent is unexplained. The standard
regression weights for the itemized variable indicates that
when item 1 (ass 1) (ass 2), (ass 3) and (ass 4) goes up by
one, the standard deviation of reliability goes up by .59, .74,
.95 and .86, respectively. Finally 0.61 percent of the
construct was explained by usability and 39 percent is
unexplained The standardized regression weights for the
itemized variable indicates that when item 1 (usb 1) (usb 2),
(usb 3) and (usb 4), the standard deviation of reliability will
go up by .82, .87, .83 and .71, respectively.
The result of the second order confirmatory factor analysis
support the adequacy of the hypothesed model with
chi-square value of X2 =215.334, DF =146, the p-value =
0.000, Normed chi-square CMINDF= 1.475, CFI =0.967,
and RMSEA = 0.051. However, all the fit indices items meet
up the threshold requirement since the value are higher than
the suggested threshold value. To conclude, the result
REFERENCES
[1]
[2]
[3]
433
J. Shaw, J. Slemrod, and J. Whiting, ―Administration and compliance,‖
in Dimens. Tax Des. Mirrlees Rev., J. Mirrlees, S. Adam, T. Besley, R.
Blundell, S. Bond. R. Chote, M. Gammie, P. Johnson, G. Myles J.
Poterba, Eds. Oxford Univ. Press, 2010.
C. E. McLure Jr and G. R. Zodrow, ―Administrative advantages of the
individual tax prepayment approach to the direct taxation of
consumption,‖ Heidelberg Congress on Taxing Consumption, 1990,
pp. 335–389.
J. Santos, ―E-service quality: a model of virtual service quality
dimensions,‖ Manag. Serv. Qual., vol. 13, no. 3, pp. 233–246, 2003.
International Journal of Trade, Economics and Finance, Vol. 5, No. 5, October 2014
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23] A. Pantouvakis, C. Chlomoudis, and A. Dimas, ―Testing the
SERVQUAL scale in the passenger port industry: a confirmatory
study,‖ Marit. Policy Manag., vol. 35, no. 5, pp. 449–467, 2008.
[24] V. A. Zeithaml and A. Malhotra, E-S-QUAL: A Multiple-Item Scale for
Assessing Electronic Service Quality, 2005.
[25] Ulrike Bauernfeind, T. Mayr, and A. Zins, ―A conceptual model for
quality dimensions for b2c recommender systems, ‖ ECIS Proceedings,
2006, pp. 1–12.
[26] J. F. Hair, B. C. William, B. J. Babin, and R. E. Anderson, Multivariate
Data Analysis, Seventh ed. pp. 1–758, 2010.
[27] M. Stiglingh, ―Authority setting for South Africa,‖ Int. Bus. Econ. Res.
J., vol. 12, no. 3, pp. 265–278, 2013.
[28] R. V Krejcie and D. W. Morgan, ―Determining sample size fpr reseach
activities,‖ NEA Res. Bull., vol. 38, pp. 607–610, 1970.
[29] R. J. Fox, M. R. Crask, and J. Kim, ―Mail survey response rate a
meta-analysis of selected techniques for inducing response,‖ Public
Opin. Q., vol. 52, no. 4, pp. 467–491, 1988.
[30] C. Fornell and D. F. Larcker, ―Structural equation models with
unobservable variables and measurement error: Algebra and statistics,‖
J. Mark. Res., pp. 382–388, 1981.
[31] R. B. Kline, Principle and Practice of Structural Equation Modeling,
Second ed. New York, 2005, pp. 1–385.
[32] J. F. Hair, M. Sarstedt, C. M. Ringle, and J. A. Mena, ―An assessment
of the use of partial least squares structural equation modeling in
marketing research,‖ J. Acad. Mark. Sci., vol. 40, no. 3, pp. 414–433,
2012.
[33] B. M. Byrne, Structural Equation Modeling with AMOS: Basic
Concepts, Applications, and Programming, CRC Press, 2009.
[34] J. L. Arbuckle, IBM SPSS Amos 19 User’s Giude, 2010, pp. 1–654.
A. Parasuraman, ―E-S-QUAL: A multiple-item scale for assessing
electronic service quality,‖ J. Serv. Res., vol. 7, no. 3, pp. 213–233,
Feb. 2005.
C.-W. Chen, ―Impact of quality antecedents on taxpayer satisfaction
with online tax-filing systems — An empirical study,‖ Inf. Manag., vol.
47, no. 5, pp. 308–315, 2010.
V. A. Zeithaml, A. Parasuraman, and A. Malhotra, ―Service quality
delivery through web sites: a critical review of extant knowledge,‖ J.
Acad. Mark. Sci., vol. 30, no. 4, pp. 362–375, 2002.
L. Liu, C. Li, and D. Zhu, ―A new approach to testing nomological
validity and its application to a second-order measurement model of
trust,‖ J. Assoc. Inf. Syst., vol. 13, no. 12, 2012.
Y.-S. Wang and Y.-W. Liao, ―Assessing eGovernment systems
success: A validation of the DeLone and McLean model of information
systems success,‖ Gov. Inf. Q., vol. 25, no. 4, pp. 717–733, 2008.
A. A. Adeyemi, D. Mustafa, and S. L. Oladipo, ―A second-order factor
gender-measurement invariance analysis of financial exclusion in
Ilorin, Nigeria,‖ Int. J. Trade, Econ. Financ., vol. 4, no. 6, pp. 398–402,
2013.
F. F. Chen, K. H. Sousa, and S. G. West, ―Teacher’s corner: Testing
measurement invariance of second-order factor models,‖ Struct. Equ.
Model., vol. 12, no. 3, pp. 471–492, 2005.
D. Sedera and G. Gable, ―A factor and structtural equation analysis of
the enterprise systems success measurement model,‖ Twenty-Fifth
International Conference on Information System, 2004, pp. 449–464.
A. Parasuraman and D. Grewal, ―The impact of technology on the
quality-value-loyalty chain: a research agenda,‖ J. Acad. Mark. Sci.,
vol. 28, no. 1, pp. 168–174, 2000.
P. Chugan and S. Pingle. (2012). Global Recession to global recovery:
Epoch strategies for marketing, family business and entrepreneurship.
[Online].
Available:
http://www.academia.edu/6875307/Global_Recession_to_Global_Rec
overy_Epoch_Strategies_for_Marketing_Family_Business_and_Entre
preneurship
A. Prakash, R. P. Mohanty, and S. P. Kallurkar, ―Service quality
modelling for life insurance business using neural networks,‖ Int. J.
Product. Qual. Manag., vol. 7, no. 3, pp. 263–286, 2011.
M. P. Gupta and D. Jana, ―E-government evaluation: A framework and
case study,‖ Gov. Inf. Q., vol. 20, no. 4, pp. 365–387, 2003.
L. Forlano, ―The emergence of digital government: international
perspectives,‖ Digit. Gov. Princ. best Pract., pp. 35–50, 2004.
W. F. Fox and M. N. Murray, ―The sales tax and electronic commerce:
So what’s new?‖ Natl. Tax J., pp. 573–592, 1997.
P. L. Poirier Jr, Electronic Tax Administration Advisory Committee
Annual Report to Congress (June 2011), DIANE Publishing, 2011.
M. Khosrowpour, Cases on Electronic Commerce Technologies and
Applications, IGI Global, 2006.
I. Chang, Y.-C. Li, W.-F. Hung, H.-G. Hwang et al., ―An empirical
study on the impact of quality antecedents on tax payers’ acceptance of
Internet tax-filing systems,‖ Gov. Inf. Q., vol. 22, no. 3, pp. 389–410,
2005.
P. Asubonteng, K. J. McCleary, and J. E. Swan, ―SERVQUAL
revisited: a critical review of service quality,‖ J. Serv. Mark., vol. 10,
no. 6, pp. 62–81, 1996.
A. Parasuraman, V. A. Zeithaml, and A. Malhotra, ―ES-QUAL a
multiple-item scale for assessing electronic service quality,‖ J. Serv.
Res., vol. 7, no. 3, pp. 213–233, 2005.
Siti Normala Sheikh Obid is an associate professor at
the Department of Accounting, Kulliyah Economic
and Management Sciences, International Islamic
University Malaysia. She obtained her MBA degree in
decision support system in USA, PhD degree in
economic (taxation) in UK. Her research interest
includes taxation, Islamic finance and banking, micro
finance, auditing, accounting, zakat and corporate
governance.
M. Bojuwon is currently pursuing his PhD degree in
accounting at International Islamic University
Malaysia. He obtained master of science degree in
International Accounting from University Utara
Malaysia and bachelor degree in accounting from
Kogi State University Anyigba Nigeria, a
Respectively. His current research focuses on the
perception of self-employed taxpayers on the usage on
online tax system. Other areas of his interest include
taxation, accounting information system and application of statistical
software (AMOS, EQS, LISREL MPLUS and PLS) in running quantitative
research.
434