The core framework’s Perception-based Attribute reduced-form structure allowed us to exploit the optimal properties of ordinary least squares in further data analysis. A regression with CSIBS as dependent variable and

USE, EOU, REL, SEC,

RES, and IMP as independent variables suggests that, given linear modeling assumptions, the consumer-satisfaction function implied by the present choice of core subset is statistically meaningful. In particular, we found that R²= 0.783, F = 109.867, df = ( 6, 175), p < 0.001, ^Attribute and that the regression coefficients for all six core attributes are statistically significant (see Table 2).

Properties of the consumer-satisfaction function, as hypothesized under H1–H6, are therefore supported by the data.

According to a standard result in statistics, the constant term in a linear regression equation captures autonomous effects on the dependent variable. In the present exercise, this can be interpreted as representing the empirical influence of service-quality variables not included in the core subset, in the sense of partitioning a universal set of quality attributes into a ^{USE EOU} subset containing {USE, ^{USE 1.380} EOU, REL, SEC, RES, ^EOU

IMP} and a complement ^REL

SEC

subset of other attributes.

RES

(Excluded variables hav- _IMP ing nothing to do with service quality would be Table 3. Estimated marginal factored into the regres- ^{rates of substitution.} sion equation’s error term.) The fact that the constant term was found to be statistically nonsignificant (in Table 2, the coefficient b_CONST = 0.277, t = 1.323, p = 0.187) indicates that these “left-out” service-quality attributes do not, even in the aggregate, significantly affect consumer satisfaction with Internet banking. This result, together with the fact that we found all six regressors to be statistically significant, suggests that the approach employed to construct the core framework is justified, both empirically and analytically.

Since hypotheses H1–H6 are supported by the

Usefulness (USE)
Ease of Use (EOU)
Reliability (REL)
Security (SEC)
Responsiveness (RES)
Continuous Improvement (IMP)
Consumer Satisfaction toward Internet
Banking Services (CSIBS)

Items Under
Likert Scaling
8
12
6
6
6
5
4

Table 1. Data reliability test.

Cronbach data, our core framework

can be applied to identify ^0.825 and evaluate strategies in ^0.907 Internet-banking man-

0.796

_0.897 agement. First consider _0.876 the fundamental strate- 0.815 gic-managerial problem 0.806 of market development.

Given the positive impact of perceived use-

fulness, it would be possible to enhance consumer satisfaction and demand by increasing the variety of banking and financial services offered over the Internet. From a micro-level analysis of the survey data, we discovered that respondents consider ease of navigation particularly impor- ^{t Sig.} tant when judging ease of _0.187 use. This finding suggests _0.028 that the upgrading of e- 0.014 banking Web sites should 0.003 be planned with this ^0.001 function in mind. Given ^0.026 the problems posed by ^0.000 counterfeit bank Web sites and the resulting disincentive effects on Table 2. Regression results. demand [ 11], our findings with regard to security and reliability supply a compelling reason to allocate more resources to combat fraudulent banking over the Internet.

The positive impact of reliability highlights the importance of satisfying

^{REL SEC RES IMP} customer expectations
^{1.460 1.770 1.160 2.900} with regard to error-free
1.058 1.283 0.841 2. 101 accounting and service
^{1.212 0.795 1.986} implementation. It is par-
0.655 1.638
ticularly important for
2.500
operations and IT man-
agers not to forget this
desideratum in the midst of popular efforts by Internet
banks to compete by offering rewards and discounts. It
has been suggested [ 9] that as customers become more
accustomed to online transactions, their concern over
security would ease [ 9]. If Internet banks consistently
demonstrated competence and a commitment to
enhancing information safety and privacy protection,
both the demand side and supply side of the market
would benefit.

Since responsiveness enhances consumer satisfaction when transacting online, initiatives by Internet banks to increase promptness and attentiveness in e-commu-nications should improve demand. Finally, the positive impact of continuous improvement suggests that in the

OLS

Coefficients b_k

Constant 0.277 1.323 USE 0.100 2.222 EOU 0.138 2.479 REL 0.146 3.042 SEC 0.177 3.407 RES 0.116 2.253 IMP 0.290 4.365