论文部分内容阅读
摘 要：本文是“電子政务服务使用信任、有效感知性、态度与持续意愿”的延伸研究。已有研究表明，继续使用电子政务服务的意愿独立于外部环境，感知有用性发挥着充分的中介作用，直接催化用户的继续使用意愿[75]。该篇文章将继续探讨在社交媒体不可或缺的情况下，“如果政务扩展（增强）到社交媒体应用的电子服务连接，它能作为用户感知有用性和继续使用意愿之间的中介吗？”研究框架主要采用技术接受模型（TAM）和期望确认理论模型（ECT/ECM）来验证研究假设和中介效应。结果显示，大多数用户对加强电子服务与社交媒体应用的连通性表达了积极的看法。在社交媒体的介入下，感知有用性不再直接与持续行为相关，而是被社交媒体的中介作用所替代。
关键词：电子政务，感知有用性，SEM，社交媒体。
Abstract：This article is an extended study of “Trust， Perceived Useful， Attitude and Continuance Intention to Use E-Government Service： An Empirical Study in Taiwan”. Previous research has shown that the intention to continue using e-government services is independent from the external environment， and the perceived usefulness plays a full mediating role that directly catalyzes users' continuance use intention [75]. This article will continue to explore under the social media indispensability， “If the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？” The research framework is mainly adapted from technology acceptance model （TAM） and expectation confirmation theory model （ECT/ECM） in validating the research hypotheses and mediating effect. The results showed that most users express positive views on enhancing e-service connectivity to social media application. With the intervention of social media， perceived usefulness is not directly related to continuous behavior， but is replaced by the mediating role of social media.
Keywords： E-government， Perceived usefulness， SEM， Social media.
1.Introduction
We have noticed from Taiwan’s TDOAS report that as early as 2009， the initial access rate of e-information and e-transactions services on government portals had reached medium levels of 50.8% and 30.3%. However， over the past eight years， the ratio of these two services continued to fall to 35.4% and 27.7% in 2017. The most striking of these is that most users indicate that they prefer to choose social media as a source of relevant information. In view of the above， previous research has shown that the intention to continue using internal e-government services is independent from the external environment （regardless of the flourishing development of generic social network application）. Obviously， the perceived usefulness plays a full mediating role that directly catalyzes users' continuance use intention. From the other end of the spectrum， since perceived usefulness of an e-service is positively associated with the attitude towards users’ continuance intention; people may inclined to give high evaluations and comments on the deployment of new technologies or innovative applications， if the user finds that service is as useful as perceived then will forms a positive attitude towards reuse behavior. Bhattacherjee et al. also point out that sometimes users' perceptions of usefulness will directly drive their continued intent， because they find that the service meets their needs， then it is possible to bypass the validation process and continue to use it directly [58]. In practice， social media is extensively used in online commercial activities and political campaigning since 2000， whilst governments still struggle to understand how best to use social media strategy to implement on operational activities and the digital transformation. Compared with the real world of e-commerce， online services and message responses provided by government authorities are sometimes insufficient and lack of convenient alternatives. Therefore the question arises， under the social media indispensability， if the government extends （enhances） e-service connectivity to social media application， and can it serve as a mediator between users’ perceived usefulness and continued use intention？ We will discuss in detail in this study
2.Theoretical Framework and Hypotheses
2.1Social Media Roles in the E-Government Scenario
E-government initiative is an innovative approach emerged in the late 1990′s to running public service more efficiency in government agencies. Which can be considered the second revolution of public administration after New Public Management （NPM） [56] [64] as part of an effort to make the public service more "businesslike" by using private sector management models. Observed from the “United Nations E-Government Survey 2008-2018”， it has significantly changed the way services are provided and the basic relationship between government and citizens. The Waseda University's “2018 International Digital Government Rankings Survey” states that countries in all regions of the world are continuing their efforts to enhance digital governance and promote digital innovation [74]. At the same time， most of the governments have increased their excellent achievements in citizen-centric approach and demand-pull online services. When we look at government online services or e-governance provisioned at this moment， it actually includes the implementation of some new concepts， especially in some practice of digital inclusion covering such electronic democracy， decision transparency and accountability [10] [38] [48].
The 2018 UN e-government survey clearly shows that digitalization and social media are the focus of government network development at this stage both have changed the nature of cooperation in service delivery [56]. The Waseda’s report also shows that there are some new trends in the use of ICTs in government activities that will have potential impacts in the coming years， such as in “Expanding the scale of smart city and e-local government” boundary [74]. Such e-services delivery at local governments would be more innovative and lower thresholds for reforming bureaucracies than in higher-level public sectors， its primarily derives from the flexibility of adopting new technologies and applications in those online activities. In practice， both surveys emphasize that policy makers and executors should strive to explore a more appropriate service mechanisms in current social atmosphere， especially through the expansion of social media application and connections. As a modern country， the ability to maintain honesty and trust with citizens is essential to building a good relationship between the two. Just like a business， social media provides a perfect way to stay transparent and clear with their customers. It’s not only a good way to share memes and keep up with the trends， but also a very powerful solution for government organizations to interact with the public and disseminate information or services.
2.2 Social Media Roles in the TAM Scenario
The technology acceptance model （TAM） introduced by Davis， Bagozzi and Warshaw in 1989， it derived from the theory of reasoned action （TRA） [13] [17] to predict the acceptability of a specific system and to identify the modification factors in assessment [20]. TAM replaces TRA’s attitude construct with two different beliefs named， perceived usefulness and perceived ease of use， both are used as significant beliefs to influence his or hers intention to adopt a newly developed technology [8] [12] [51] [52]. Perceived usefulness （PU） is the degree to which a user thinks a technology would enhance performance or productivity in workplace; it effective reflects the function of information system （IS） performance and measure the overall productivity， effectiveness and efficiency. Perceived ease of use （PEOU） as "the degree to which a person believes that using a particular system would be free from effort"， which is to measure the extent of self-efficacy， including ease of learning， understanding， operationand control [26] [27] [37] [43].
The TAM model demonstrates the direct linkage between perceived usefulness and behavioral intention to use， which is stronger than another attitude in perceived ease of use. It represents that perceived usefulness act as a major determinant that will influence actual system use [12]. The subsequent studies also confirmed that perceived usefulness also substantially affects attitudes during ante and post acceptance stages of IS use; while perceived ease of use is non-significant associated with attitude in the post acceptance stage [12] [34] [58]. However， if two new IS offer the same functionality， most users will find useful one that is easier to use [15].
If we switch to the contemporary United States， “Social Media Use in 2018” by Pew Research shows that 88% of people 18-29 years old have at least one social media account [71]， and over 60% of 13 to 17-year-olds have at least one profile on social media， with many spending more than two hours per day on social networking sites [72]. From the perspective of our research context， those scenarios experienced in Taiwan， Japan or U.S. are all seem to be able mapping to the TAM's theoretical framework; the prevailing of social media trend has indeed influence the evaluation of PU and PEOU toward attitude and behavioral intention to re-act [54] [61]. Based on the above discussion， we replace attitude in TAM's model （Figure 2） by focus into social media application to be consistent with the research context （Figure 1）. Therefore， the following research hypotheses are proposed： H1： Perceived usefulness in e-government service is positively associated with social media application toward users’ intention to continue use.
H2： Social media application in e-government service is positively associated with users’ intention to continue use.
2.3The Continuance Intention in ECT and ECM
The expectation confirmation theory （ECT） is a theory that derived from cognitive dissonance theory. It seeks to identify what determines the individual consumer’s satisfaction and whether they will repurchase the product or service [39] [57] [66] [68]， which being widely used in behavior research of consumer marketing and information system adoption [6] [33] [45] [55]. ECT model assumes that pre-expectations and perceived performance will lead to post-purchase satisfaction. In other words， when one’s disconfirmation is positive which will posit to increase post-satisfaction， and if negative， the result will be reversed. The following research by Oliver （1980） also confirmed the hypothesis test of the ECT model [57]. Compared with TAM's initial behavior and planned behavior theory （TPB） research， the framework of ECT can be regarded as a dynamic process model of post hoc research [49]. However， Bhattacherjee argued that ECT mainly ignores potential changes in users' expectations after using new products or systems， and the impact of these changes on subsequent cognitive processes.
Bhattacherjee’s expectation confirmation model （ECM） is mainly derived from the continuation of the three models ECT， TAM and TPB [1] [12] [21]. ECM focuses on three major variables in determining the intention of continuous use， including expectations， satisfaction and confirmation， while the pre-consumption expectation in ECT is replaced by ECM’s post-consumption expectations. In ECM， the ex-post expectation is represented by perceived usefulness， since perceived usefulness is a cognitive belief salient to IS use.Although the theoretical extension of ECM was mainly drawn upon ECT， ECM’s framework is distinct from ECT since the initial adoption behavior does not represent an intention that would automatically lead to continued use. Bhattacherjee asserts that perceived usefulness just as the cognitive belief in IS acceptance contexts which may also be related [58] [59]. Compared to TAM's model， ECM is believed to contribute a more meaningful dimension in exploring attitudes toward users’ perceived usefulness， satisfaction and continued use intention [12] [43] [44] [58] [60] [65]. Teo， Srivastava， and Jiang conducted an empirical study in 2008， combining online trust with IS's success model （DeLone/McLean） and the framework of ECM to explore the factors that contribute to the success of e-government services. The research indicates that the intent of continuing to use e-government websites is similar to users revisiting specific business portals. Especially in the post-adoption phase， users' intentions to continue using government websites often follow repetitive behaviors and are affected by feedback mechanisms. Teo， Srivastava， and Jiang’s finding can be regarded as consistent with the arguments of the ECM model. [50] [53] [62]. Extrapolating from these arguments， here we propose the third research hypothesis：
H3： Perceived usefulness in e-government service is positively associated with users’ intention to continue use.
Figure 1. The extended frameworkFigure 2. The original framework
（Social media intervention） *** p<0.001， ** p<0.01， * p<0.05
3.Method
The research framework was mainly adapted from technology acceptance model （TAM） and expectation confirmation theory model （ECT/ECM）， in validating research hypotheses and mediating effect. The research framework posits two exogenous variables of perceived usefulness （the performance of e-government service）， and social media application （the government extends （enhances） e-service connectivity）. The third variable is the intention to continue using e-government services as an endogenous variable. The original questionnaire consisted of 20 questions attributed to three constructs， which was adapted from the existing literature of validated scales （Table 1）. The corresponding answers are coded as a five-point Likert scale from “strongly disagree” to “strong agree”. The scale first invites five graduate students majoring in business and social sciences to conduct a pre-test， to check that each item in the questionnaire is clear and easy to understand without any ambiguity， to ensure the reliability and validity in the measurement model. After removing the low-reliability items from the original question， the total Cronbach's Alpha value exceeds the standard 0.7 [42]， from the perspective of statistical analysis， this scale is suitable for entering the next stage of the formal questionnaire survey.
The sample allocation was conducted by stratified proportional sampling method based on the 2018 demographic data in Ministry of Interior Taiwan [14]. The respondents mainly invited 400 subjects from the PCCU campus， based on their different attributes and their experience of accessing electronic services on public portals. The survey is conducted in a timely manner and careful explained in advance to ensure the validity of the scale and achieve an effective response rate of 100%. In Taiwan， the age of e-government adopters is mainly distributed between 20 and 39， accounting for 38.5% of total users. Therefore， the design of the sample distribution not only conforms to the outline of the TDOAS survey， but also avoids potential bias to provide a lively and effective scale for our research [69]. Under the social media indispensability， the assessment tool will adopt structural equation model （SEM） technique to identify “If the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？” The first advantage of using the SEM multiple regression analysis is both measurement and structural models can be incorporated into simultaneous analysis [16] [74]. The measurement model includes factors loading that are expected to converge on their proposed constructs; while structural model involves examining the proposed hypotheses among theoretical framework and providing an overall model fit. The second advantage is SEM is being capable of impute relationships between unobserved constructs from observable variables [41] [63] [73]. According to the above procedure of research method， it should be able to effectively achieve the study goals of this article.
4.The Measurement Model
The assessment of normality shows that all variables are conforms to the suggested standard values， skewness < 2， and kurtosis < 7 [23] [31] [36]. Multivariate kurtosis c.r. value is 114.85 also in line with Bollen and Stine （1993） suggested threshold [7] [22] [24]. The preliminary offending test illustrates all error variances were ranged from 0.04 to 0.38 of significance， positive support; standard errors of variances are range from 0.01 to 0.04 of significance， and factor loadings without exceeding or very close to value 1.0. Further， as part of justifying in measurement model， the study will proceed with confirmatory factor analysis （CFA） to verify convergent validity and discriminant validity of the scale [9] [18] [19] [47].
For convergent validity， Table 2 shows the standardized factor loadings to the corresponding variable are highly significant that range from 0.59 to 0.95. CR on perceived usefulness is 0.889， social media application is 0.935 and intention is 0.858; AVE on perceived usefulness is 0.674， social media application is 0.744， and intention is 0.670. All the observed variables have reach a highly degree of reliability and convergent validity [9] [18] [28] [29] [30]. For discriminant validity， the study adopts bootstrap estimation method of confidence interval （CI）， which includes： bias-corrected percentile， percentile， and point estimation method （Table 3）. If the interval contains a value of 1 at the 95% confidence level of the correlation coefficient， the null hypothesis is rejected， which means that there is discriminative validity between the two constructs， and vice versa [16] [11]. Table 4 shows each of confidence interval estimation doesn’t contain value 1; it indicates the discriminant validity was satisfied for every variable of the proposed model. 5.The Structural Model
After proposing the modification process of the model， Table 4 shows three fit indices including the absolute fit indices， incremental fit indices， and parsimonious fit indices， all meet the recommended thresholds [2] [3] [32]. The path coefficients and hypotheses test in SEM model （Figure 3） indicates that perceived usefulness is both positively associated with social media application and continuance use intention toward e-government services. H1 and H3 are supported. However， the direct link between social media application and users’ intention is also found to be significantly associated. Therefore， the proposed hypotheses H1， H2 and H3 among theoretical framework are all supported.
6.The Mediation Effect
In social sciences research， a mediation model is to identify the process that underlies an observed relationship between an independent variable and a dependent variable via third mediator variable; especially when they appear haven’t a definite connection [5] [25] [73]. In our context， we will continue to discuss under the social media indispensability， if the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？
Since we have acknowledged that three proposed hypotheses H1， H2 and H3 are all supported， obtained regression weights γ=0.77， β=0.32， andγ=0.18 in sequence. Whereas， we also observed the direct path between “ICU←PU” in Table 5， if out of indirect effect （γ=0.25） from the total effect （γ=0.43）， leaving only （γ=0.18） in direct effect. Obviously， this may probably cause by third mediator variable （social media application）， from practical of view， the strengths of three loading factors are inconsistent， and it’s need further to clarify the causal relationship between the variables [46].
Table 6 shows the indirect effect and direct effect of perceived usefulness on users’ intention， using product of coefficients， bootstrapping CI of bias-corrected and percentile and MacKinnon of PRODCLIN2 methods[35] [73]. In bootstrapping CI at 95% level， if the upper and lower limit of the estimated value does not including zero， which represent exists the mediation effect. The indirect effects in （PU→BI）， product of coefficients Z = 3 > |1.96|， bootstrap CI and MacKinnon PRODCLIN2 doesn’t contain zero; test results indicate that indirect effect is significant exists. However， the direct effects of product of coefficients Z = 1.47 < |1.96| threshold， both bootstrap CI and MacKinnon PRODCLIN2 contain zero， representing the direct effects doesn’t exist [40]. From technical point of view， with the intervention of social media， perceived usefulness is not directly related to continuous behavior， but replaced by the mediating role of social media. 7.Discussions and Implications
This article is an extended study of “Trust， Perceived Useful， Attitude and Continuance Intention to Use E-Government Service： An Empirical Study in Taiwan”. Previous studies has verified that the intention to continue using e-government services is independent from the external environment; while perceived usefulness plays a full mediating role that direct catalyzes users' continuance use intention. However， accompany with the prevailing of social media trend， it has indeed influence the commons perception toward service quality and effectiveness on government portals [56]. Along the same vein， the study conducted an in-depth analysis continues to explore under the social media indispensability， if the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？ Through the path coefficients and hypotheses test in SEM model （Figure 3）， we find that perceived usefulness is both positively associated with social media application and continuance use intention toward e-government services. H1 and H3 are supported. The direct link between social media application and users’ intention （H2） is also found to be significantly associated. However， we found that the strength of three paths loading factor are not consistent， which may cause by the mediation effect of social media application， and need proceed to the next stage of clarifying the causal relationship between variables. The final results showed that most users take a positive attitude to this issue， but if intervened through the use of social media， the perceived usefulness does not seem to be directly related to continuous use behavior， but is replaced by a full mediation role connecting social media.
From a practical point of view， the "United Nations e-government survey" and "The Waseda-IAC's international e-government ranking survey" both clearly support the findings of this article. They emphasized that today's governments need to continuously evaluate the innovative trends in ICT use， which will have a potential impact on e-government service models in the coming years [74]. This mainly includes social media， the Internet of Things， and big data analytics， which can provide more attractive and innovative mechanisms for government online services. For example， Waseda's annual survey retains ten key indicators and 35 sub-indicators to accurately and precisely assess the latest developments in e-government in the ICT sector in all target countries. We can easily identify the top ten countries in 2017 and 2018， such as Denmark， Singapore， United Kingdom， United States， Japan， and South Korea [74]. Their government websites all have basic common elements that allow users to perform cluster-type communicated with each other and access via links on social media for relevant information and services. In practice， although our research has found interventions in the connection of e-service and social media application; we still need to emphasize the performance of perceived usefulness is still expected to be the most salient factor influencing user’s ex-post attitude. People may be inclined to give high evaluations and comments on the deployment of new technologies or innovative applications， if the user finds that service is as useful as perceived then will forms a positive attitude towards reusing. But sometimes users' different perceptions of usefulness will directly drive their continued intent， because they find that the service meets their needs （or the only choice）， then it is possible to bypass the validation process and continue to use it directly [58]. The success of e-government initiatives is contingent upon citizen’s satisfaction， and continued use. Compared with the e-commerce environment， governments must recognize the common fact that most of the public sector online services and feedback mechanisms are insufficient or lack convenient alternatives. Therefore， as a modern government， the most important duty is to constantly assess the impact of its innovative use of ICT to different modes of e-government services， in order to meet the continuing needs of the users.
8.Conclusions and limitations
This paper highlights the importance of continuous use behavioral research on a specific information system， especially in the context of e-government framework. From the perspective of empirical research， regardless of the differences in theoretical models， hypothesis verification process or research fields， the results can be compared with previous studies for further exploration. Since e-government initiative emerged late 1990′s， it advocate "everything is service" to running public service in a more efficiency way， yet policy makers should pay more attention to users continued intentions under the premise of sustainable development.Here， we can make conclude that our research findings is consistent with Bhattacherjee’s assertion as stated in the literature review， ECM’s framework is distinct from ECT since the initial adoption behavior does not represent an intention that would automatically lead to continued use behavior. However， if it intervenes through the deployment of new technologies or innovative applications during the verification process， it will directly inspire users' intentions to continue using beyond cognitive beliefs. The research questionnaire is mainly targeted at ethnic groups between 20 and 25 years of age in order to meet most attributes of the TDOAS survey and to minimize personal cognitive biases to provide a lively and effective analysis condition [69]. However， the study measurement items may not be comprehensive enough in exploring user's cognitive beliefs and behaviors intention. Secondly， on the premise of future research， it is recommended to moderately extend the theoretical framework to other related constructs， such as online security and ICT applications issues. Finally， we do believe this empirical study has made a substantial contribution in current stage.
References
[1]Ajzen， I.， （1991）， “The theory of planned behavior，” Organizational Behavior and Human Decision Processes， Vol.50， No.2， pp.179-211.
[2]Anderson， J. C. and D. W. Gerbing ， （1988）， “Structural equation modeling in practice： A review and recommended two-step approach，” Psychological Bulletin， Vol.103， No. 3， pp.411-423.
[3]Bagozzi， R. P. and Y. Yi， （1988）， “On the evaluation of structural equation models，” Journal of the academy of marketing science， Vol.16， No.1， pp.74-94.
[4]Bagozzi， R. P.， Davis， F. D. and P. R. Warshaw， （1992）， “Development and test of a theory of technological learning and usage，” Human Relations， Vol.45， No.7， pp.660-686.
[5]Baron， R. and D. Kenny， （1908）， “The moderator-mediator distinction in social psychological research： Conceptual， strategic， statistical considerations，” Personality Social Psychology， Vol.51， pp.1173-1182.
[6]Belanger， F. and L. Carter， （2008）， “Trust and Risk in E-Government Adoption，” Journal of Strategic Information Systems， Vol.17， pp.165-176.
[7]Bollen， K. A. and R. A. Stine， （1993）， Bootstrapping goodness-of-fit measures in SEM：Testing structural equation models， K. A. Bollen and J. S. Long， Eds.： Sage.
[8]Burnkrant， R. E. and J. P. Jr. Thomas， （1988）， “The Structure and Antecedents of the Normative and Attitudinal Components of Fishbein's TRA，” Journal of Experimental Social Psychology， Vol.24， pp.66-87.
[9]Campbell， D. T. and D. W. Fiske， （2004）， “Convergent and discriminant validation by the multitrait–multimethod matrix，” Psychological Bulletin， Vol.56， pp.81-105， 1959.
[10]Chu， P.， Hsiao， N.， Lee， F. and C. Chen， （2003）， “Exploring success factors for Taiwan's government e-tendering system： behavior perspectives from end users，” Government Information Quarterly， Vol.21， No.2， pp.219-234. [11]Cohen， J.， Cohen， P.， West， S. G. and L. S. Aiken， （2003）， Applied multiple regressions： Correlation analysis for the behavior sciences， 3rd ed.： Erlbaum.
[12]Davis， F. D.， （1989）， “Perceived usefulness， perceived ease of use and user acceptance of information technology，” MIS Quarterly， Vol.13， No. 3， pp.319-340.
[13]Davis， F. D.， Bagozzi， R. P. and P. R. Warshaw， （1989）， “User acceptance of computer technology： a comparison of two theoretical models，” Management Science， Vol.35， No.8， pp.982-1003.
[14]MOI. （2018）， Ministry of Interior， Statistical Yearbook-The Department of Statistics，： MOI Taiwan.
[15]Dillon， A. and M. G. Morris， （1996）， “User acceptance of information technology： Theories and models，” Annual Review of Information Science and Technology， Vol.31， pp.3-32.
[16]Efron， B.， （1987）， “Better Bootstrap Confidence Intervals，” Journal of the American Statistical Asso.， Vol.82， No. 397， pp.171-185.
[17]Fishbein， M. and I. Ajzen， （1975）， Belief， Attitude， Intention and Behavior： An Introduction to Theory and Research， Reading， Massachusetts： Addison-Wesley.
[18]Fornell， C. and D. Larcker， （1981）， “Evaluating structural equation models with unobservable variables and measurement error，” Journal of marketing research， Vol.18， pp.39-50.
[19]Gao， S.， Mokhtarian， P. and R. Johnston， （2008）， “Non-normality of Data in SEM Transportation Research Record，” Journal of the Transportation Research Board， Vol.2082， pp.116-124.
[20]Gefen， D.， Karahanna， E. and D. Straub， （2003）， *Trust and TAM in online shopping： an integrated model，” MIS Quarterly， Vol.27， No.1， pp.51-90.
[21]Joey， F. G.， （2004）， “The theory of planned behavior and Internet purchasing，” Internet Research， Vol.14， No.3， pp.198-212.
[22]Hair Jr.， J. F.， Anderson， R. E.， Tatham， R. L. and W. C. Black， （1998）， Multivariate data analysis， 5th ed.： Prentice-Hall.
[23]Hallow， L. L.（1985）， Behavior of Some Elliptical Theory Estimators with Non-normality Data in a Covariance Structures Framework： A Monte Carlo study， Ph.D. Dissertation： UCLA.
[24]Hair Jr.， J. F.， Black， W. C.， Babin， B. J.， Anderson， R. E. and R. L. Tatham， （2006）， Multivariate data analysis， 6th ed.： Prentice-Hall.
[25]Hayes， A. F. （2009）， “Beyond Baron and Kenny： Statistical mediation analysis in the new millennium，” Communication Monographs， Vol.76， pp.408-420.
[26]Hsu， C.， Lu， H. and H. Hsu， （2007）， “Adoption of the mobile Internet： An empirical study of multimedia message service （MMS），” Omega， Vol.35， pp.715-726. [27]Hung， S.， Chang， C. and T. Yu， （2006）， “Determinants of user acceptance of the e-gov service： The case of online tax filing and payment system，” Government Information Quarterly， Vol.23， pp.97-122.
[28]Jackson， D. L.， Gillaspy Jr. J. A. and R. Purc-Stephenson， （2009）， “Reporting Practices in CFA： An Overview and Some Recommendations，” Psychological Methods， Vol.14， No.1， pp.6-23.
[29]J？reskog， K. G.， （1967）， “A general approach to confirmatory maximum likelihood factor analysis，” ETS Research Bulletin Series， Vol.2， pp.183-202.
[30]J？reskog， K. G. and D. S？rbom， （2000）， Analysis of Linear Structural Relationships by Maximum Likelihood， Instrumental Variables， and Least Squares Methods LISREL VI， 7th ed.： Scientific Software.
[31]Kenny， D. A. and C. M. Judd， （1984）， “Estimating the nonlinear and interactive effects of latent variables，” Psychological Bulletin， Vol.96， No.1， pp.20-210.
[32]Kenny， D. A. and D. B. McCoach， （2003）， “Effect of the number of variables on measures of fit in SEM，” Structural Equation Modeling： A Multidisciplinary Journal， Vol.10， No.3， pp.333-351.
[33]Kotler， P.， Kartajaya， H. and I. Setiawan， （2010）， Marketing 3.0： From Products to Customers to Human Spirit： Wiley & Sons.
[34]Nicolas，C. L.， Castillo， F. JM. and H. Bouwmanb， （2008）， “An assessment of advanced mobile services acceptance： Contributions from TAM and diffusion theory，” Information & Management， Vol.45， pp.359-364.
[35]MacKinnon， D. P. （2008）， Introduction to Statistical Mediation Analysis： Erlbaum.
[36]Mardia， K.V.， （1970）， “Measures of multivariate skewness and kurtosis with applications，” Biometrika， vol.57， pp.519-530.
[37]Medlin， B. D. （2001）， The factors that may influence a faculty member's decision to adopt electronic technologies in instruction， Ph.D. dissertation： Virginia Tech.
[38]O’Reilly， T. （2005）， What Is Web2.0： Design Patterns and Business Models for Next Generation of Software： O’Reilly Media.
[39]Peng， Y. and S. Tao， （2009）， “A study of information searchers on cosmetic blogs： Lifestyle and purchase behavior，” The Journal of Advertising and Public Relations， Vol.32， pp.91-123.
[40]Preacher， K. J. and A. F. Hayes， （2008）， “Asymptotic and resampling strategies for assessing and comparing indirect effects in multi-mediator models，” Behavior Research Methods， Vol.40， No.3， pp.879-891.
[41]Reisinger， Y. and T. Lindsay， （1999）， “Structural equation modeling， SEM with Lisrel︰application in tourism，” Tourism Management， Vol.20， pp.71-88. [42]Cortina， J. M.， （1993）， “What is Coefficient Alpha？ An Examination of Theory and Applications，” Journal of Applied Psychology， Vol.78， No.1， pp.98-107.
[43]Sahli， A. B. and L. Patrick， （2014）， “Sing the Decomposed Theory of Planned Behavior （DTPB） to Explain the Intention to Book Tourism Products Online，” Int. Journal of Online Marketing， Vol.4， No.1， pp.1-10.
[44]Schaupp， L.C.， Carter， L. and M. E. McBride， （2010）， “E-file adoption： A study of U.S. taxpayers’ intentions，” Computers in Human Behavior， Vol.26， pp.636-644.
[45]Schierz， P. G. Schilke， O. and B. W. Wirtz， （2010）， “Understanding consumer acceptance of mobile payment services： An empirical analysis，” E-Commerce Research and Applications， Vol.9， pp.209-216.
[46]Shrout， P. E. and N. Bolger， （2002）， “Mediation in experimental and nonexperi-mental studies： New procedures and recommendations，” Psychological Methods， Vol.7， No.4， pp.422-445.
[47]Stevens， J. P. （1996）， Applied multivariate statistics for the social sciences， 3rd ed.： Lawrence Erlbaum Associates.
[48]Sugiyama， K. and T. Andree， （2010）， The Dentsu Way： McGraw-Hill.
[49]Sumita， U. and R. sogai， （2009）， Development of e-marketing contract structure based on consumer-generated contents and its optimal strategy， social system and management discussion series： U. of Tsukuba.
[50]Tan， M. and T. Teo， （2000）， “Factors Influencing the Adoption of Internet Banking，” Journal of the Association for Information Systems， Vol.1， No.5， pp.1-42.
[51]Taylor， S. and P. Todd， （1995a）， “Decomposition and crossover effects in the TPB： a study of consumer adoption intentions，” International Journal of Research in Marketing， Vol.12， pp.137-155.
[52]Taylor， S. and P. Todd， （1995b）， “Understanding information technology usage： A test of competing models，” Information Systems Research， Vol.6， No.2， pp.144-176.
[53]Teo， T. and S. Pok， （2003）， “Adoption of WAP-enabled mobile phones among Internet users，” Omega， Vol.31， pp.483-498.
[54]Vijayasarathy， L.， （2004）， “Predicting consumer intentions to use online shopping： the case for an augmented technology acceptance model，” Information & Management， Vol.41， pp.747-762.
[55]Wei， P. S. and L.H. Peng， （2013）， “An examination of celebrity endorsements and online customer review influence female consumers shopping behavior，” Computers in Human Behavior， Vol.29， pp.193-201.
[56]UN， （2016）， E-Government Survey： UN Department of Economic and Social A？airs （DESA）. [57]Oliver， R. L.， （1980）， “A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions，” Journal of Marketing Research， Vol.17， No. 4， pp.460-469.
[58]Bhattacherjee， A.， （2001）， “Understanding Information Systems Continuance： An Expectation Confirmation Model，” MIS Quarterly， Vol.25， No.3， pp.351-370.
[59]Stone， R. W. and L. Baker-Eveleth， （2013）， “Students’ expectation， confirmation， and continuance intention to use electronic textbooks，” Computers in Human Behavior， Vol.29， No. 3， pp.984-990.
[60]Limayem， M. and M. K. C. Cheung， （2008）， “Understanding information systems continuance： The case of Internet-based learning technologies，” Information & Management， Vol.45， No.4， pp.227-232.
[61]Gefen， D.， Karahanna， E. and D. W. Straub， （2003）， “Trust and TAM in online shopping： An integrated model，” MIS Quarterly， Vol.27， No.1， pp.5-90.
[62]Teo， T.， Shirish Srivastava， C. and L. Jiang， （2008）， “Trust and Electronic Government Success： An Empirical Study，” Journal of Management Information Systems， Vol.25， No.3， pp.99-132.
[63]Alzahrani， A. I. （2011）， Web-based e-Government Services Acceptance for G2C-A Structural Equation Modeling Approach， Ph.D. dissertation： De Montfort University.
[64]Jeong， C. H. （2007）， Fundamental of Development Administration， Selangor： Scholar Press.
[65]Karahanna， E.， Straub， D. and N. Chervany， （1999）， “Info. Tech. Adoption Across Time： A Cross Sectional Comparison of Pre and Post Adoption Beliefs，） MIS Quarterly， vol.23， no2， pp.183-21.
[66]Festlnger， L. （1957）， A Theory of Cognitive Dissonance： Stanford University Press.
[67]George， J. F. （2004）， “The TPB and Internet purchasing，” Internet Research， Vol.14， No.3， pp.198-212.
[68]Festinger， L.， （1962）， “Cognitive dissonance，” Scientific American， Vol.207， No.4， pp.93-107.
[69]TDOAS. （2017）， Taiwan digital opportunity annual survey： National Development Council.
[70]Bertot， J. C.， Jaeger， P. T. and J. M. Grimes， （2010）， “Using ICT to create a culture of transparency： E-gov. and social media as openness and anti-corruption tool for societies，” Gov. Inf. Quarterly， Vol.27， pp.264-271.
[71]Smith， A. and M. Andeson， （2018）， Social Media Use in 2018： Pew Research Center.
[72]Hajirnis， A.， （2015）， “Social media networking： Parent guidance required，” The Brown University Child and Adolescent Behavior Letter， Vol.31， No. 12， pp.1-7.
[73]MacKinnon， D. P. （2008）， Introduction to Statistical Mediation Analysis， Mahwah： Erlbaum.
[74]Waseda-IAC Survey. （2018）， The 2018 Waseda-IAC International Digital Government ranking survey： The Waseda University.
[75]Tsui， H. D.， （2019）， “Trust， Perceived Useful， Attitude and Continuance Intention to Use E-Government Service： An Empirical Study in Taiwan，” IEICE Transacts on Information and Systems， Vol.E102-D， No.12， pp.2524-2534.
关键词：电子政务，感知有用性，SEM，社交媒体。
Abstract：This article is an extended study of “Trust， Perceived Useful， Attitude and Continuance Intention to Use E-Government Service： An Empirical Study in Taiwan”. Previous research has shown that the intention to continue using e-government services is independent from the external environment， and the perceived usefulness plays a full mediating role that directly catalyzes users' continuance use intention [75]. This article will continue to explore under the social media indispensability， “If the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？” The research framework is mainly adapted from technology acceptance model （TAM） and expectation confirmation theory model （ECT/ECM） in validating the research hypotheses and mediating effect. The results showed that most users express positive views on enhancing e-service connectivity to social media application. With the intervention of social media， perceived usefulness is not directly related to continuous behavior， but is replaced by the mediating role of social media.
Keywords： E-government， Perceived usefulness， SEM， Social media.
1.Introduction
We have noticed from Taiwan’s TDOAS report that as early as 2009， the initial access rate of e-information and e-transactions services on government portals had reached medium levels of 50.8% and 30.3%. However， over the past eight years， the ratio of these two services continued to fall to 35.4% and 27.7% in 2017. The most striking of these is that most users indicate that they prefer to choose social media as a source of relevant information. In view of the above， previous research has shown that the intention to continue using internal e-government services is independent from the external environment （regardless of the flourishing development of generic social network application）. Obviously， the perceived usefulness plays a full mediating role that directly catalyzes users' continuance use intention. From the other end of the spectrum， since perceived usefulness of an e-service is positively associated with the attitude towards users’ continuance intention; people may inclined to give high evaluations and comments on the deployment of new technologies or innovative applications， if the user finds that service is as useful as perceived then will forms a positive attitude towards reuse behavior. Bhattacherjee et al. also point out that sometimes users' perceptions of usefulness will directly drive their continued intent， because they find that the service meets their needs， then it is possible to bypass the validation process and continue to use it directly [58]. In practice， social media is extensively used in online commercial activities and political campaigning since 2000， whilst governments still struggle to understand how best to use social media strategy to implement on operational activities and the digital transformation. Compared with the real world of e-commerce， online services and message responses provided by government authorities are sometimes insufficient and lack of convenient alternatives. Therefore the question arises， under the social media indispensability， if the government extends （enhances） e-service connectivity to social media application， and can it serve as a mediator between users’ perceived usefulness and continued use intention？ We will discuss in detail in this study
2.Theoretical Framework and Hypotheses
2.1Social Media Roles in the E-Government Scenario
E-government initiative is an innovative approach emerged in the late 1990′s to running public service more efficiency in government agencies. Which can be considered the second revolution of public administration after New Public Management （NPM） [56] [64] as part of an effort to make the public service more "businesslike" by using private sector management models. Observed from the “United Nations E-Government Survey 2008-2018”， it has significantly changed the way services are provided and the basic relationship between government and citizens. The Waseda University's “2018 International Digital Government Rankings Survey” states that countries in all regions of the world are continuing their efforts to enhance digital governance and promote digital innovation [74]. At the same time， most of the governments have increased their excellent achievements in citizen-centric approach and demand-pull online services. When we look at government online services or e-governance provisioned at this moment， it actually includes the implementation of some new concepts， especially in some practice of digital inclusion covering such electronic democracy， decision transparency and accountability [10] [38] [48].
The 2018 UN e-government survey clearly shows that digitalization and social media are the focus of government network development at this stage both have changed the nature of cooperation in service delivery [56]. The Waseda’s report also shows that there are some new trends in the use of ICTs in government activities that will have potential impacts in the coming years， such as in “Expanding the scale of smart city and e-local government” boundary [74]. Such e-services delivery at local governments would be more innovative and lower thresholds for reforming bureaucracies than in higher-level public sectors， its primarily derives from the flexibility of adopting new technologies and applications in those online activities. In practice， both surveys emphasize that policy makers and executors should strive to explore a more appropriate service mechanisms in current social atmosphere， especially through the expansion of social media application and connections. As a modern country， the ability to maintain honesty and trust with citizens is essential to building a good relationship between the two. Just like a business， social media provides a perfect way to stay transparent and clear with their customers. It’s not only a good way to share memes and keep up with the trends， but also a very powerful solution for government organizations to interact with the public and disseminate information or services.
2.2 Social Media Roles in the TAM Scenario
The technology acceptance model （TAM） introduced by Davis， Bagozzi and Warshaw in 1989， it derived from the theory of reasoned action （TRA） [13] [17] to predict the acceptability of a specific system and to identify the modification factors in assessment [20]. TAM replaces TRA’s attitude construct with two different beliefs named， perceived usefulness and perceived ease of use， both are used as significant beliefs to influence his or hers intention to adopt a newly developed technology [8] [12] [51] [52]. Perceived usefulness （PU） is the degree to which a user thinks a technology would enhance performance or productivity in workplace; it effective reflects the function of information system （IS） performance and measure the overall productivity， effectiveness and efficiency. Perceived ease of use （PEOU） as "the degree to which a person believes that using a particular system would be free from effort"， which is to measure the extent of self-efficacy， including ease of learning， understanding， operationand control [26] [27] [37] [43].
The TAM model demonstrates the direct linkage between perceived usefulness and behavioral intention to use， which is stronger than another attitude in perceived ease of use. It represents that perceived usefulness act as a major determinant that will influence actual system use [12]. The subsequent studies also confirmed that perceived usefulness also substantially affects attitudes during ante and post acceptance stages of IS use; while perceived ease of use is non-significant associated with attitude in the post acceptance stage [12] [34] [58]. However， if two new IS offer the same functionality， most users will find useful one that is easier to use [15].
If we switch to the contemporary United States， “Social Media Use in 2018” by Pew Research shows that 88% of people 18-29 years old have at least one social media account [71]， and over 60% of 13 to 17-year-olds have at least one profile on social media， with many spending more than two hours per day on social networking sites [72]. From the perspective of our research context， those scenarios experienced in Taiwan， Japan or U.S. are all seem to be able mapping to the TAM's theoretical framework; the prevailing of social media trend has indeed influence the evaluation of PU and PEOU toward attitude and behavioral intention to re-act [54] [61]. Based on the above discussion， we replace attitude in TAM's model （Figure 2） by focus into social media application to be consistent with the research context （Figure 1）. Therefore， the following research hypotheses are proposed： H1： Perceived usefulness in e-government service is positively associated with social media application toward users’ intention to continue use.
H2： Social media application in e-government service is positively associated with users’ intention to continue use.
2.3The Continuance Intention in ECT and ECM
The expectation confirmation theory （ECT） is a theory that derived from cognitive dissonance theory. It seeks to identify what determines the individual consumer’s satisfaction and whether they will repurchase the product or service [39] [57] [66] [68]， which being widely used in behavior research of consumer marketing and information system adoption [6] [33] [45] [55]. ECT model assumes that pre-expectations and perceived performance will lead to post-purchase satisfaction. In other words， when one’s disconfirmation is positive which will posit to increase post-satisfaction， and if negative， the result will be reversed. The following research by Oliver （1980） also confirmed the hypothesis test of the ECT model [57]. Compared with TAM's initial behavior and planned behavior theory （TPB） research， the framework of ECT can be regarded as a dynamic process model of post hoc research [49]. However， Bhattacherjee argued that ECT mainly ignores potential changes in users' expectations after using new products or systems， and the impact of these changes on subsequent cognitive processes.
Bhattacherjee’s expectation confirmation model （ECM） is mainly derived from the continuation of the three models ECT， TAM and TPB [1] [12] [21]. ECM focuses on three major variables in determining the intention of continuous use， including expectations， satisfaction and confirmation， while the pre-consumption expectation in ECT is replaced by ECM’s post-consumption expectations. In ECM， the ex-post expectation is represented by perceived usefulness， since perceived usefulness is a cognitive belief salient to IS use.Although the theoretical extension of ECM was mainly drawn upon ECT， ECM’s framework is distinct from ECT since the initial adoption behavior does not represent an intention that would automatically lead to continued use. Bhattacherjee asserts that perceived usefulness just as the cognitive belief in IS acceptance contexts which may also be related [58] [59]. Compared to TAM's model， ECM is believed to contribute a more meaningful dimension in exploring attitudes toward users’ perceived usefulness， satisfaction and continued use intention [12] [43] [44] [58] [60] [65]. Teo， Srivastava， and Jiang conducted an empirical study in 2008， combining online trust with IS's success model （DeLone/McLean） and the framework of ECM to explore the factors that contribute to the success of e-government services. The research indicates that the intent of continuing to use e-government websites is similar to users revisiting specific business portals. Especially in the post-adoption phase， users' intentions to continue using government websites often follow repetitive behaviors and are affected by feedback mechanisms. Teo， Srivastava， and Jiang’s finding can be regarded as consistent with the arguments of the ECM model. [50] [53] [62]. Extrapolating from these arguments， here we propose the third research hypothesis：
H3： Perceived usefulness in e-government service is positively associated with users’ intention to continue use.
Figure 1. The extended frameworkFigure 2. The original framework
（Social media intervention） *** p<0.001， ** p<0.01， * p<0.05
3.Method
The research framework was mainly adapted from technology acceptance model （TAM） and expectation confirmation theory model （ECT/ECM）， in validating research hypotheses and mediating effect. The research framework posits two exogenous variables of perceived usefulness （the performance of e-government service）， and social media application （the government extends （enhances） e-service connectivity）. The third variable is the intention to continue using e-government services as an endogenous variable. The original questionnaire consisted of 20 questions attributed to three constructs， which was adapted from the existing literature of validated scales （Table 1）. The corresponding answers are coded as a five-point Likert scale from “strongly disagree” to “strong agree”. The scale first invites five graduate students majoring in business and social sciences to conduct a pre-test， to check that each item in the questionnaire is clear and easy to understand without any ambiguity， to ensure the reliability and validity in the measurement model. After removing the low-reliability items from the original question， the total Cronbach's Alpha value exceeds the standard 0.7 [42]， from the perspective of statistical analysis， this scale is suitable for entering the next stage of the formal questionnaire survey.
The sample allocation was conducted by stratified proportional sampling method based on the 2018 demographic data in Ministry of Interior Taiwan [14]. The respondents mainly invited 400 subjects from the PCCU campus， based on their different attributes and their experience of accessing electronic services on public portals. The survey is conducted in a timely manner and careful explained in advance to ensure the validity of the scale and achieve an effective response rate of 100%. In Taiwan， the age of e-government adopters is mainly distributed between 20 and 39， accounting for 38.5% of total users. Therefore， the design of the sample distribution not only conforms to the outline of the TDOAS survey， but also avoids potential bias to provide a lively and effective scale for our research [69]. Under the social media indispensability， the assessment tool will adopt structural equation model （SEM） technique to identify “If the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？” The first advantage of using the SEM multiple regression analysis is both measurement and structural models can be incorporated into simultaneous analysis [16] [74]. The measurement model includes factors loading that are expected to converge on their proposed constructs; while structural model involves examining the proposed hypotheses among theoretical framework and providing an overall model fit. The second advantage is SEM is being capable of impute relationships between unobserved constructs from observable variables [41] [63] [73]. According to the above procedure of research method， it should be able to effectively achieve the study goals of this article.
4.The Measurement Model
The assessment of normality shows that all variables are conforms to the suggested standard values， skewness < 2， and kurtosis < 7 [23] [31] [36]. Multivariate kurtosis c.r. value is 114.85 also in line with Bollen and Stine （1993） suggested threshold [7] [22] [24]. The preliminary offending test illustrates all error variances were ranged from 0.04 to 0.38 of significance， positive support; standard errors of variances are range from 0.01 to 0.04 of significance， and factor loadings without exceeding or very close to value 1.0. Further， as part of justifying in measurement model， the study will proceed with confirmatory factor analysis （CFA） to verify convergent validity and discriminant validity of the scale [9] [18] [19] [47].
For convergent validity， Table 2 shows the standardized factor loadings to the corresponding variable are highly significant that range from 0.59 to 0.95. CR on perceived usefulness is 0.889， social media application is 0.935 and intention is 0.858; AVE on perceived usefulness is 0.674， social media application is 0.744， and intention is 0.670. All the observed variables have reach a highly degree of reliability and convergent validity [9] [18] [28] [29] [30]. For discriminant validity， the study adopts bootstrap estimation method of confidence interval （CI）， which includes： bias-corrected percentile， percentile， and point estimation method （Table 3）. If the interval contains a value of 1 at the 95% confidence level of the correlation coefficient， the null hypothesis is rejected， which means that there is discriminative validity between the two constructs， and vice versa [16] [11]. Table 4 shows each of confidence interval estimation doesn’t contain value 1; it indicates the discriminant validity was satisfied for every variable of the proposed model. 5.The Structural Model
After proposing the modification process of the model， Table 4 shows three fit indices including the absolute fit indices， incremental fit indices， and parsimonious fit indices， all meet the recommended thresholds [2] [3] [32]. The path coefficients and hypotheses test in SEM model （Figure 3） indicates that perceived usefulness is both positively associated with social media application and continuance use intention toward e-government services. H1 and H3 are supported. However， the direct link between social media application and users’ intention is also found to be significantly associated. Therefore， the proposed hypotheses H1， H2 and H3 among theoretical framework are all supported.
6.The Mediation Effect
In social sciences research， a mediation model is to identify the process that underlies an observed relationship between an independent variable and a dependent variable via third mediator variable; especially when they appear haven’t a definite connection [5] [25] [73]. In our context， we will continue to discuss under the social media indispensability， if the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？
Since we have acknowledged that three proposed hypotheses H1， H2 and H3 are all supported， obtained regression weights γ=0.77， β=0.32， andγ=0.18 in sequence. Whereas， we also observed the direct path between “ICU←PU” in Table 5， if out of indirect effect （γ=0.25） from the total effect （γ=0.43）， leaving only （γ=0.18） in direct effect. Obviously， this may probably cause by third mediator variable （social media application）， from practical of view， the strengths of three loading factors are inconsistent， and it’s need further to clarify the causal relationship between the variables [46].
Table 6 shows the indirect effect and direct effect of perceived usefulness on users’ intention， using product of coefficients， bootstrapping CI of bias-corrected and percentile and MacKinnon of PRODCLIN2 methods[35] [73]. In bootstrapping CI at 95% level， if the upper and lower limit of the estimated value does not including zero， which represent exists the mediation effect. The indirect effects in （PU→BI）， product of coefficients Z = 3 > |1.96|， bootstrap CI and MacKinnon PRODCLIN2 doesn’t contain zero; test results indicate that indirect effect is significant exists. However， the direct effects of product of coefficients Z = 1.47 < |1.96| threshold， both bootstrap CI and MacKinnon PRODCLIN2 contain zero， representing the direct effects doesn’t exist [40]. From technical point of view， with the intervention of social media， perceived usefulness is not directly related to continuous behavior， but replaced by the mediating role of social media. 7.Discussions and Implications
This article is an extended study of “Trust， Perceived Useful， Attitude and Continuance Intention to Use E-Government Service： An Empirical Study in Taiwan”. Previous studies has verified that the intention to continue using e-government services is independent from the external environment; while perceived usefulness plays a full mediating role that direct catalyzes users' continuance use intention. However， accompany with the prevailing of social media trend， it has indeed influence the commons perception toward service quality and effectiveness on government portals [56]. Along the same vein， the study conducted an in-depth analysis continues to explore under the social media indispensability， if the government extends （enhances） e-service connectivity to social media application， can it serve as a mediator between users’ perceived usefulness and continued use intention？ Through the path coefficients and hypotheses test in SEM model （Figure 3）， we find that perceived usefulness is both positively associated with social media application and continuance use intention toward e-government services. H1 and H3 are supported. The direct link between social media application and users’ intention （H2） is also found to be significantly associated. However， we found that the strength of three paths loading factor are not consistent， which may cause by the mediation effect of social media application， and need proceed to the next stage of clarifying the causal relationship between variables. The final results showed that most users take a positive attitude to this issue， but if intervened through the use of social media， the perceived usefulness does not seem to be directly related to continuous use behavior， but is replaced by a full mediation role connecting social media.
From a practical point of view， the "United Nations e-government survey" and "The Waseda-IAC's international e-government ranking survey" both clearly support the findings of this article. They emphasized that today's governments need to continuously evaluate the innovative trends in ICT use， which will have a potential impact on e-government service models in the coming years [74]. This mainly includes social media， the Internet of Things， and big data analytics， which can provide more attractive and innovative mechanisms for government online services. For example， Waseda's annual survey retains ten key indicators and 35 sub-indicators to accurately and precisely assess the latest developments in e-government in the ICT sector in all target countries. We can easily identify the top ten countries in 2017 and 2018， such as Denmark， Singapore， United Kingdom， United States， Japan， and South Korea [74]. Their government websites all have basic common elements that allow users to perform cluster-type communicated with each other and access via links on social media for relevant information and services. In practice， although our research has found interventions in the connection of e-service and social media application; we still need to emphasize the performance of perceived usefulness is still expected to be the most salient factor influencing user’s ex-post attitude. People may be inclined to give high evaluations and comments on the deployment of new technologies or innovative applications， if the user finds that service is as useful as perceived then will forms a positive attitude towards reusing. But sometimes users' different perceptions of usefulness will directly drive their continued intent， because they find that the service meets their needs （or the only choice）， then it is possible to bypass the validation process and continue to use it directly [58]. The success of e-government initiatives is contingent upon citizen’s satisfaction， and continued use. Compared with the e-commerce environment， governments must recognize the common fact that most of the public sector online services and feedback mechanisms are insufficient or lack convenient alternatives. Therefore， as a modern government， the most important duty is to constantly assess the impact of its innovative use of ICT to different modes of e-government services， in order to meet the continuing needs of the users.
8.Conclusions and limitations
This paper highlights the importance of continuous use behavioral research on a specific information system， especially in the context of e-government framework. From the perspective of empirical research， regardless of the differences in theoretical models， hypothesis verification process or research fields， the results can be compared with previous studies for further exploration. Since e-government initiative emerged late 1990′s， it advocate "everything is service" to running public service in a more efficiency way， yet policy makers should pay more attention to users continued intentions under the premise of sustainable development.Here， we can make conclude that our research findings is consistent with Bhattacherjee’s assertion as stated in the literature review， ECM’s framework is distinct from ECT since the initial adoption behavior does not represent an intention that would automatically lead to continued use behavior. However， if it intervenes through the deployment of new technologies or innovative applications during the verification process， it will directly inspire users' intentions to continue using beyond cognitive beliefs. The research questionnaire is mainly targeted at ethnic groups between 20 and 25 years of age in order to meet most attributes of the TDOAS survey and to minimize personal cognitive biases to provide a lively and effective analysis condition [69]. However， the study measurement items may not be comprehensive enough in exploring user's cognitive beliefs and behaviors intention. Secondly， on the premise of future research， it is recommended to moderately extend the theoretical framework to other related constructs， such as online security and ICT applications issues. Finally， we do believe this empirical study has made a substantial contribution in current stage.
References
[1]Ajzen， I.， （1991）， “The theory of planned behavior，” Organizational Behavior and Human Decision Processes， Vol.50， No.2， pp.179-211.
[2]Anderson， J. C. and D. W. Gerbing ， （1988）， “Structural equation modeling in practice： A review and recommended two-step approach，” Psychological Bulletin， Vol.103， No. 3， pp.411-423.
[3]Bagozzi， R. P. and Y. Yi， （1988）， “On the evaluation of structural equation models，” Journal of the academy of marketing science， Vol.16， No.1， pp.74-94.
[4]Bagozzi， R. P.， Davis， F. D. and P. R. Warshaw， （1992）， “Development and test of a theory of technological learning and usage，” Human Relations， Vol.45， No.7， pp.660-686.
[5]Baron， R. and D. Kenny， （1908）， “The moderator-mediator distinction in social psychological research： Conceptual， strategic， statistical considerations，” Personality Social Psychology， Vol.51， pp.1173-1182.
[6]Belanger， F. and L. Carter， （2008）， “Trust and Risk in E-Government Adoption，” Journal of Strategic Information Systems， Vol.17， pp.165-176.
[7]Bollen， K. A. and R. A. Stine， （1993）， Bootstrapping goodness-of-fit measures in SEM：Testing structural equation models， K. A. Bollen and J. S. Long， Eds.： Sage.
[8]Burnkrant， R. E. and J. P. Jr. Thomas， （1988）， “The Structure and Antecedents of the Normative and Attitudinal Components of Fishbein's TRA，” Journal of Experimental Social Psychology， Vol.24， pp.66-87.
[9]Campbell， D. T. and D. W. Fiske， （2004）， “Convergent and discriminant validation by the multitrait–multimethod matrix，” Psychological Bulletin， Vol.56， pp.81-105， 1959.
[10]Chu， P.， Hsiao， N.， Lee， F. and C. Chen， （2003）， “Exploring success factors for Taiwan's government e-tendering system： behavior perspectives from end users，” Government Information Quarterly， Vol.21， No.2， pp.219-234. [11]Cohen， J.， Cohen， P.， West， S. G. and L. S. Aiken， （2003）， Applied multiple regressions： Correlation analysis for the behavior sciences， 3rd ed.： Erlbaum.
[12]Davis， F. D.， （1989）， “Perceived usefulness， perceived ease of use and user acceptance of information technology，” MIS Quarterly， Vol.13， No. 3， pp.319-340.
[13]Davis， F. D.， Bagozzi， R. P. and P. R. Warshaw， （1989）， “User acceptance of computer technology： a comparison of two theoretical models，” Management Science， Vol.35， No.8， pp.982-1003.
[14]MOI. （2018）， Ministry of Interior， Statistical Yearbook-The Department of Statistics，： MOI Taiwan.
[15]Dillon， A. and M. G. Morris， （1996）， “User acceptance of information technology： Theories and models，” Annual Review of Information Science and Technology， Vol.31， pp.3-32.
[16]Efron， B.， （1987）， “Better Bootstrap Confidence Intervals，” Journal of the American Statistical Asso.， Vol.82， No. 397， pp.171-185.
[17]Fishbein， M. and I. Ajzen， （1975）， Belief， Attitude， Intention and Behavior： An Introduction to Theory and Research， Reading， Massachusetts： Addison-Wesley.
[18]Fornell， C. and D. Larcker， （1981）， “Evaluating structural equation models with unobservable variables and measurement error，” Journal of marketing research， Vol.18， pp.39-50.
[19]Gao， S.， Mokhtarian， P. and R. Johnston， （2008）， “Non-normality of Data in SEM Transportation Research Record，” Journal of the Transportation Research Board， Vol.2082， pp.116-124.
[20]Gefen， D.， Karahanna， E. and D. Straub， （2003）， *Trust and TAM in online shopping： an integrated model，” MIS Quarterly， Vol.27， No.1， pp.51-90.
[21]Joey， F. G.， （2004）， “The theory of planned behavior and Internet purchasing，” Internet Research， Vol.14， No.3， pp.198-212.
[22]Hair Jr.， J. F.， Anderson， R. E.， Tatham， R. L. and W. C. Black， （1998）， Multivariate data analysis， 5th ed.： Prentice-Hall.
[23]Hallow， L. L.（1985）， Behavior of Some Elliptical Theory Estimators with Non-normality Data in a Covariance Structures Framework： A Monte Carlo study， Ph.D. Dissertation： UCLA.
[24]Hair Jr.， J. F.， Black， W. C.， Babin， B. J.， Anderson， R. E. and R. L. Tatham， （2006）， Multivariate data analysis， 6th ed.： Prentice-Hall.
[25]Hayes， A. F. （2009）， “Beyond Baron and Kenny： Statistical mediation analysis in the new millennium，” Communication Monographs， Vol.76， pp.408-420.
[26]Hsu， C.， Lu， H. and H. Hsu， （2007）， “Adoption of the mobile Internet： An empirical study of multimedia message service （MMS），” Omega， Vol.35， pp.715-726. [27]Hung， S.， Chang， C. and T. Yu， （2006）， “Determinants of user acceptance of the e-gov service： The case of online tax filing and payment system，” Government Information Quarterly， Vol.23， pp.97-122.
[28]Jackson， D. L.， Gillaspy Jr. J. A. and R. Purc-Stephenson， （2009）， “Reporting Practices in CFA： An Overview and Some Recommendations，” Psychological Methods， Vol.14， No.1， pp.6-23.
[29]J？reskog， K. G.， （1967）， “A general approach to confirmatory maximum likelihood factor analysis，” ETS Research Bulletin Series， Vol.2， pp.183-202.
[30]J？reskog， K. G. and D. S？rbom， （2000）， Analysis of Linear Structural Relationships by Maximum Likelihood， Instrumental Variables， and Least Squares Methods LISREL VI， 7th ed.： Scientific Software.
[31]Kenny， D. A. and C. M. Judd， （1984）， “Estimating the nonlinear and interactive effects of latent variables，” Psychological Bulletin， Vol.96， No.1， pp.20-210.
[32]Kenny， D. A. and D. B. McCoach， （2003）， “Effect of the number of variables on measures of fit in SEM，” Structural Equation Modeling： A Multidisciplinary Journal， Vol.10， No.3， pp.333-351.
[33]Kotler， P.， Kartajaya， H. and I. Setiawan， （2010）， Marketing 3.0： From Products to Customers to Human Spirit： Wiley & Sons.
[34]Nicolas，C. L.， Castillo， F. JM. and H. Bouwmanb， （2008）， “An assessment of advanced mobile services acceptance： Contributions from TAM and diffusion theory，” Information & Management， Vol.45， pp.359-364.
[35]MacKinnon， D. P. （2008）， Introduction to Statistical Mediation Analysis： Erlbaum.
[36]Mardia， K.V.， （1970）， “Measures of multivariate skewness and kurtosis with applications，” Biometrika， vol.57， pp.519-530.
[37]Medlin， B. D. （2001）， The factors that may influence a faculty member's decision to adopt electronic technologies in instruction， Ph.D. dissertation： Virginia Tech.
[38]O’Reilly， T. （2005）， What Is Web2.0： Design Patterns and Business Models for Next Generation of Software： O’Reilly Media.
[39]Peng， Y. and S. Tao， （2009）， “A study of information searchers on cosmetic blogs： Lifestyle and purchase behavior，” The Journal of Advertising and Public Relations， Vol.32， pp.91-123.
[40]Preacher， K. J. and A. F. Hayes， （2008）， “Asymptotic and resampling strategies for assessing and comparing indirect effects in multi-mediator models，” Behavior Research Methods， Vol.40， No.3， pp.879-891.
[41]Reisinger， Y. and T. Lindsay， （1999）， “Structural equation modeling， SEM with Lisrel︰application in tourism，” Tourism Management， Vol.20， pp.71-88. [42]Cortina， J. M.， （1993）， “What is Coefficient Alpha？ An Examination of Theory and Applications，” Journal of Applied Psychology， Vol.78， No.1， pp.98-107.
[43]Sahli， A. B. and L. Patrick， （2014）， “Sing the Decomposed Theory of Planned Behavior （DTPB） to Explain the Intention to Book Tourism Products Online，” Int. Journal of Online Marketing， Vol.4， No.1， pp.1-10.
[44]Schaupp， L.C.， Carter， L. and M. E. McBride， （2010）， “E-file adoption： A study of U.S. taxpayers’ intentions，” Computers in Human Behavior， Vol.26， pp.636-644.
[45]Schierz， P. G. Schilke， O. and B. W. Wirtz， （2010）， “Understanding consumer acceptance of mobile payment services： An empirical analysis，” E-Commerce Research and Applications， Vol.9， pp.209-216.
[46]Shrout， P. E. and N. Bolger， （2002）， “Mediation in experimental and nonexperi-mental studies： New procedures and recommendations，” Psychological Methods， Vol.7， No.4， pp.422-445.
[47]Stevens， J. P. （1996）， Applied multivariate statistics for the social sciences， 3rd ed.： Lawrence Erlbaum Associates.
[48]Sugiyama， K. and T. Andree， （2010）， The Dentsu Way： McGraw-Hill.
[49]Sumita， U. and R. sogai， （2009）， Development of e-marketing contract structure based on consumer-generated contents and its optimal strategy， social system and management discussion series： U. of Tsukuba.
[50]Tan， M. and T. Teo， （2000）， “Factors Influencing the Adoption of Internet Banking，” Journal of the Association for Information Systems， Vol.1， No.5， pp.1-42.
[51]Taylor， S. and P. Todd， （1995a）， “Decomposition and crossover effects in the TPB： a study of consumer adoption intentions，” International Journal of Research in Marketing， Vol.12， pp.137-155.
[52]Taylor， S. and P. Todd， （1995b）， “Understanding information technology usage： A test of competing models，” Information Systems Research， Vol.6， No.2， pp.144-176.
[53]Teo， T. and S. Pok， （2003）， “Adoption of WAP-enabled mobile phones among Internet users，” Omega， Vol.31， pp.483-498.
[54]Vijayasarathy， L.， （2004）， “Predicting consumer intentions to use online shopping： the case for an augmented technology acceptance model，” Information & Management， Vol.41， pp.747-762.
[55]Wei， P. S. and L.H. Peng， （2013）， “An examination of celebrity endorsements and online customer review influence female consumers shopping behavior，” Computers in Human Behavior， Vol.29， pp.193-201.
[56]UN， （2016）， E-Government Survey： UN Department of Economic and Social A？airs （DESA）. [57]Oliver， R. L.， （1980）， “A Cognitive Model of the Antecedents and Consequences of Satisfaction Decisions，” Journal of Marketing Research， Vol.17， No. 4， pp.460-469.
[58]Bhattacherjee， A.， （2001）， “Understanding Information Systems Continuance： An Expectation Confirmation Model，” MIS Quarterly， Vol.25， No.3， pp.351-370.
[59]Stone， R. W. and L. Baker-Eveleth， （2013）， “Students’ expectation， confirmation， and continuance intention to use electronic textbooks，” Computers in Human Behavior， Vol.29， No. 3， pp.984-990.
[60]Limayem， M. and M. K. C. Cheung， （2008）， “Understanding information systems continuance： The case of Internet-based learning technologies，” Information & Management， Vol.45， No.4， pp.227-232.
[61]Gefen， D.， Karahanna， E. and D. W. Straub， （2003）， “Trust and TAM in online shopping： An integrated model，” MIS Quarterly， Vol.27， No.1， pp.5-90.
[62]Teo， T.， Shirish Srivastava， C. and L. Jiang， （2008）， “Trust and Electronic Government Success： An Empirical Study，” Journal of Management Information Systems， Vol.25， No.3， pp.99-132.
[63]Alzahrani， A. I. （2011）， Web-based e-Government Services Acceptance for G2C-A Structural Equation Modeling Approach， Ph.D. dissertation： De Montfort University.
[64]Jeong， C. H. （2007）， Fundamental of Development Administration， Selangor： Scholar Press.
[65]Karahanna， E.， Straub， D. and N. Chervany， （1999）， “Info. Tech. Adoption Across Time： A Cross Sectional Comparison of Pre and Post Adoption Beliefs，） MIS Quarterly， vol.23， no2， pp.183-21.
[66]Festlnger， L. （1957）， A Theory of Cognitive Dissonance： Stanford University Press.
[67]George， J. F. （2004）， “The TPB and Internet purchasing，” Internet Research， Vol.14， No.3， pp.198-212.
[68]Festinger， L.， （1962）， “Cognitive dissonance，” Scientific American， Vol.207， No.4， pp.93-107.
[69]TDOAS. （2017）， Taiwan digital opportunity annual survey： National Development Council.
[70]Bertot， J. C.， Jaeger， P. T. and J. M. Grimes， （2010）， “Using ICT to create a culture of transparency： E-gov. and social media as openness and anti-corruption tool for societies，” Gov. Inf. Quarterly， Vol.27， pp.264-271.
[71]Smith， A. and M. Andeson， （2018）， Social Media Use in 2018： Pew Research Center.
[72]Hajirnis， A.， （2015）， “Social media networking： Parent guidance required，” The Brown University Child and Adolescent Behavior Letter， Vol.31， No. 12， pp.1-7.
[73]MacKinnon， D. P. （2008）， Introduction to Statistical Mediation Analysis， Mahwah： Erlbaum.
[74]Waseda-IAC Survey. （2018）， The 2018 Waseda-IAC International Digital Government ranking survey： The Waseda University.
[75]Tsui， H. D.， （2019）， “Trust， Perceived Useful， Attitude and Continuance Intention to Use E-Government Service： An Empirical Study in Taiwan，” IEICE Transacts on Information and Systems， Vol.E102-D， No.12， pp.2524-2534.