1 Introduction

It is hard to define clearly what E-government system is, although its enabled services have been immersed in people’s lives for years. In general, E-government is defined as “utilizing the Internet and the world-wide-web for delivering government information and services to citizens” by UN and American Society for Public Administration (ASPA) (UN and ASPA 2002). It indicates that the essence of this system is to promote and deliver transparent, convenient, and effective public services to citizens by applying advanced web technologies.

In literature, researchers studied E-government phenomenon from three major perspectives: what is E-government, what is the evolvement of E-government service, and what do citizens expect from E-government (Gil-Garcia and Martinez-Moyano 2007). Some researchers regard E-government as the application of web technologies to promote government services efficiently and effectively (Brown and Brudney 2001), and the government is virtually organized through agencies who are structured and enabled by web technologies (J. E. Fountain 2001). Additional technology deployed apart from web technologies may include database, multimedia, and so on (Jaeger 2003). On the other hand, some researchers regard E-government as electronic means that enable interactions between the government and receivers of public services, such as citizens, private sectors, and other public administrations (Means and Schneider 2000). Researchers are prone to capture this phenomenon from different aspects and define it accordingly, thus no well-recognized definition is commonly agreed yet (Halchin 2004; Yildiz 2007).

Several reasons are accounted for this difficulty in defining E-government systems, especially from the perspective of citizens who are the primary service receivers. According to Yildiz (Yildiz 2007), first, unlike the objectives of E-government, which are relatively easy to define, there is a lack of works describing the involved stakeholders’ activities and the specific technology deployed in such systems. Therefore, the meaning of E-government that stems from its deployed social context, such as primary social groups and implemented strategies, is blurred and overlooked (Yildiz 2007). Second, for citizens with diverse characteristics and background, E-government may have different meanings and most of the citizens may not grasp its overall image clearly (Torres et al. 2005; Grant and Derek 2005). Third, it is not the advancing technology defining what E-government is, rather it is the service delivery process and key players’ behaviours making us understand the “evolving nature of the E-government concept better (Yildiz 2007, pp.10) (Jonas 2000; Hwang et al. 1999). Since the ultimate purpose of E-government systems is to deliver better services to citizens, understanding such systems with an emphasis on” heterogeneous citizens’ background, preference, and expectation towards public services is important and key to improve the service provision process and service quality.

1.1 E-Government Service Evolvement

There are also different ways to categorize E-government services. From the interactivity perspective, the online services can be divided into static information provision, one-way (government to customer) interaction, mutual interaction, and transactional service (Arduini et al. 2010). The services could be further summarized into two categories: information provision service and transactional service (Venkatesh et al. 2012). Compared with sheer information provision services, transactional services which involve online electronic transactions are more sophisticated and require additional procedures to complete. Co-created public service between citizens and public sectors are also emerging rapidly, such as citizen sourcing and do-it-yourself-government (Linders 2012).

On the other hand, according to Layne and Lee (Layne and Lee 2001), the services can be divided into following stages from an evolutionary perspective. In the initial stage, governmental website containing certain government information is classified and open to the public, thus only one-way communication is guaranteed. The second stage evolves to a two-way communication in which online transactions are provided. The third stage is the vertical integration of central and local public sectors by sharing information resources via database. Normally, modification and improvement on the service delivery process is required. The final stage is the horizontal integration in which different functional areas are integrated horizontally and delivered via one single portal (Gil-Garcia and Martinez-Moyano 2007).

Gil-Garcia et al. (Gil-Garcia and Martinez-Moyano 2007) also argued that there is a trend of E-government movement from national level to local sectors along with the system evolution, and the stages vary across administration levels. It is very possible that the E-government service is already well designed at national level, whilst still at an initial stage amongst the local administrative sectors (Gupta and Jana 2003). Furthermore, the target citizens, administrative processes, and technological sophistication will be different at each evolvement stage (Holden et al. 2003). The divergent demand and pressure from end-users, i.e. citizens and private sectors, may also influence the features of initiatives (Gil-Garcia and Martinez-Moyano 2007).

In reality, poor design of E-government systems that only aligns with common practices and ignores the needs of citizens often fail to entice citizens (Meuter et al. 2000; Rai and Sambamurthy 2006). Comprehending and designing use-centric public service are thus remained complicated and challenging along with the continuous service provision to citizens. Without identifying the involved stakeholders and exploring the process through which the services are delivered to end-users, as well as the social context and environment, one could not profoundly and thoroughly understand the E-government phenomenon. Hence, adopting new methodologies to investigate this phenomenon from a new perspective is critical for designing and providing the services with new insights.

In this work, we will review the theoretical foundations and methodologies deployed in E-government research in past decades, from which we could posit and justify a service system perspective and an agent-based simulation approach. We then propose a corresponding conceptual agent-based framework to identify and abstract the system components and their interactions as a holistic system. It will serve as the first step towards designing E-government services by adopting a service system perspective and micro-level simulation approaches complementing those traditional ones.

2 Theoretical Aspect

E-government is a relatively new research field closely relating to other research realms, such as information systems and public administration (Heeks and Bailur 2007). Most of the extant works examined different aspects of this phenomenon, for instance, E-government service adoption and diffusion process, and governance issues from either end-user sectors or public governor perspectives (Gil-Garcia and Martinez-Moyano 2007; Cordella and Iannacci 2010; Luna-Reyes and Gil-Garcia 2011; van Velsen et al. 2009; Venkatesh et al. 2012). There are few systematic studies investigating E-government system from a holistic view (Fields et al. 2007; Arduini et al. 2010) and especially from a citizen-centred service system perspective.

According to Heeks and Bailur’s study (Heeks and Bailur 2007), although they expected that various theories from other disciplines had been applied in E-government research, surprisingly that most of the theoretical frameworks used were stage model-based and category-based focusing on E-government’s evolvement stages and features. Theoretical frameworks from the governance literature, and models and schemas from the information system literature contribute most to the E-government research (Heeks and Bailur 2007). Theories built only from and for E-government studies are therefore expected. Within the “theories” applied, the dominant ones are from two major fields: information systems adoption/diffusion and information systems in organizations, which will be discussed respectively in the following.

2.1 Innovation Adoption

With respect to information systems deployed in E-government services, besides the technical development considering security issues (Kaliontzoglou et al. 2005), adoption and diffusion processes of E-government services have attracted most of the attention. There is a stream of work focusing on the innovation adoption and assimilation involved in E-government phenomena. Various factors are analysed to examine their influence on citizens’ intention to adopt the services, such as trust, governmental leadership, and different measurements of services provided (Bélanger and Carter 2008; Luk 2009; Lean et al. 2009). However, the information technology itself does not define what E-government system is (Yildiz 2007), though it may shape the way of how public services are delivered to the public and how involved stakeholders communicate with each other. In the following, we will review some classic innovation diffusion models applied in those studies.

2.1.1 Macro-Level Diffusion Model

By 1970s, several main innovation diffusion models have been proposed and established, and their variants were widely applied in different contexts afterwards (Meade and Islam 2006). Macro-level diffusion models, such as Bass’s diffusion model (M Bass 1969), were proposed to forecast the first-time adoption and to evaluate the market penetration with a focus on collective adoption behaviours. For instance, Bass’s model (M Bass 1969) captures both homogeneous population’s desire to innovate and to imitate others that drive the collective adoption behaviours. A large body of mathematical models have reflected a general “Bell shape” curve for period-by-period adoption and a “S-shape” curve for cumulated adopters (Teng et al. 2002).

On the other side, heterogeneous individuals are considered in Roger’s innovation diffusion model (Rogers 2003). Individuals are categorized as innovators, early adopters, early majorities, late majorities, and laggards of which the percentage is a normal distribution, thus creating the S-shape adoption curve (Meade and Islam 2006). Those who are better educated and of a higher social-economic status tend to have a lower adoption threshold, thus a higher adoption rate of innovations.

This strand of macro-level models is particular useful when the market penetration or market share of innovations is of primary interest (Schramm et al. 2010). However, public services are different from innovations in the sense that service adoption is an evolving process along with the influences from end-users directly or indirectly, who learn from their past experience or the environment whilst utilizing the services.

2.1.2 Micro-Level consumer’s Behaviour Model

Regarding micro-level behaviour models, there is a strand of well-defined social-psychological theories based on micro-level consumer’s adoption models, amongst which Technology Acceptance Model (TAM) (Davis 1989), Theory of Planned Behaviour (TPB) (Fishbein and Ajzen 1975), and Unified Theory of Acceptance and Use of Technology (UTAUT) (Venkatesh et al. 2003) are influential. This strand of psychological-social models treats intention as the predicator of behaviours (Zhang and Nuttall 2011), and has been empirically validated in E-government research (Schaupp et al. 2010; Lean et al. 2009; Gunasekaran and Ngai 2008; Bélanger and Carter 2008). TRA-based theories which study the intention and behaviour of users taking into account not only the service itself but also the social contexts are believed to be more suitable for E-government studies (Arduini et al. 2010). However, Shareef et al. (M. A. Shareef et al. 2011) argued that those classic innovation adoption models were not capable to reflect complex citizens’ adoption behaviours, and critical influential factors should be defined separately since they may differ along with the development of service levels. Also, this set of theory focused only on the causal relationship underlying the phenomenon and left the dynamic service delivery process unexplored. The rich and meaningful interactions among stakeholders were therefore either simplified or ignored.

Besides, there is also a strand of explanatory models such as consumer diffusion paradigm proposed (Gatignon and Robertson 1985). According to those works, the adoption of innovations depends on three major aspects: individual’s characteristics, perceived innovation properties, and social influences posed on individuals (Schramm et al. 2010). Regarding E-government studies, Tung and Rieck (Tung and Rieck 2005) proposed a theoretical framework based on the diffusion of innovation theory, network externalities, adoption barriers, and influence from social aspects to analyse the E-government adoption among business organizations in Singapore. Extensive frameworks based on adoption theories and models have been proposed (M. Shareef et al. 2010; Klievink and Janssen 2009; van Dijk et al. 2008; Schedler and Summermatter 2007) and we will not review each of them here. However, different from innovations, E-government services will be improved in terms of efficiency and quality continuously, and new services will gradually emerge and be delivered during the assimilation period (Arduini et al. 2010).

In summary, there is a missing co-evolved process between the adoption behaviours and the E-government service diffusion process in above models. We should bridge the citizen’s adaptive behaviours to the macro-level adoption patterns emerged from different social groups, and based on which evaluate supporting strategies to prompt the service usage. The comparison of above models is summarized in Table 5.1.

Table 5.1 Comparison of innovation diffusion models
Full size table

2.2 Governance

Another stream of frameworks adopted in E-government studies is from the governance literature, such as public administration and political science. They advanced the understanding of E-government phenomena from a government structure perspective and explored the relations between government, information technology, and institutions. However, it ignored the characteristics of heterogeneous primary users and their valuable feedback to the services, which is critical to the success of E-government systems, even from a governance perspective.

Governance should not be defined as a physical entity, i.e. government and governing individual citizens. Rather it is a process about authorities and citizens interacting with each other and guiding themselves (Holliday and Kwok 2004). Information and communication technologies are becoming very critical in changing the administration work of government to some extent. E-government as an emerging application of information technology can facilitate the governing process and invite a broader range of citizens participating in public affairs. With respect to E-government research, there have been extensive works devoted to study how the E-government system is inter-weaved with governance processes and how it can influence the inter-organizational management and external relationship with citizens (Holliday and Kwok 2004; Dias and Rafael 2007; Hossain et al. 2011). Theories such as institutional theory (Richard 1995) and enacted technology framework (J. Fountain 2007) were applied, although the impact still remains controversial.

Researchers also borrow concepts from organizational information technology and e-Business fields (Lee and Rao 2009). There is also a group of works focusing on the design issues of E-government services under either governmental settings or user contexts by different approaches, such as content analysis, process modelling, and output evaluation (Buchanan and McMenemy 2011; Dias and Rafael 2007; Vassilakis et al. 2007). But E-government is different from either of those fields due to the monopolistic nature of government (Srivastava and Teo 2010; Hossain et al. 2011), and the services are voluntary to use since traditional front desk services are still available as an alternative. In this sense, E-government deserves its own theories and methodologies.

2.3 Service System Perspective

E-government system is not only about the information technology deployed in public sectors to facilitate public administration, rather, it is a service provision process that involves both citizens as service receivers and public sectors as service providers. Undoubtedly, citizens with divergent background and different views of E-government play a key role in making E-government systems meaningful and successful (Akman et al. 2005). On one hand, citizens with divergent background will influence the services in both direct and indirect ways (Farrell and Saloner 1986). Pressures from citizens to improve the efficiency and usability of services have a direct impact on the service provided, whilst interactions between citizens and the government, together with its resulted adoption rate have an indirect impact in the sense that even the innovation provided by public sectors is affected by what is expected from the receiver side (Arduini et al. 2010). In other words, service receivers may transfer their unique experiences whilst dealing with the services to the service provider, and eventually enhance the service design at certain degree (Barras 1986).

On the other hand, regarding organizational concerns, innovation-relevant decisions reflect the vision and strategy of public sectors, and limited resources are subject to relocation among different social groups. The system should satisfy individual’s needs, rather than the opposite case. Therefore, besides the basic user-friendly requirements of E-government systems, there should be more considerations on citizens’ diverse behaviours of using different types of services, which will have substantial impacts on the service design. Furthermore, government should help different groups of citizens conquer the usage barriers, and policy design on supporting strategies should take citizen’s needs into account as well, which makes the E-government services more complicated and challenging to understand and design (Arduini et al. 2010).

Therefore, in order to understand E-government systems as holistic ones, a service system perspective, from which individual elements of the system can be defined, integrated, and analysed (Anthopoulos et al. 2007; Demirkan et al. 2008), such as citizens, organizational resources, social-cultural context, strategy, and initiatives (Wimmer 2002), is a suitable tool to investigate this phenomenon. Services are defined as “the application of competence and knowledge to create value between providers and receivers” (Goldstein et al. 2002; Spohrer et al. 2007). From this definition, a service system involves not only technology, people, and organization, but also the shared knowledge and social context; the dynamic process is more important than static entities (Checkland 1999; Vargo and Lusch 2004; Demirkan et al. 2008); and citizens participate in and influence the delivery process directly or indirectly. In this sense, services can also be viewed as “a series of interactions between service provider and clients that result in an observable output” (Spohrer et al. 2007).

3 Methodological Aspect

With respect to the E-government research field, Heeks and Bailur (Heeks and Bailur 2007) concluded that although most of the works do not have an explicit epistemological stance beneath the stated methodology, the prevailing but not explicitly stated epistemology stances are positivism, and somewhere in-between to social constructionist. Accordingly, both qualitative and quantitative methods have been applied in this field. In the following, we will briefly review the methodologies applied in E-government research with the associated underlying epistemological stance.

3.1 Quantitative Studies

Regarding E-government research, a positivism stance implies an objective or realism ontology. This stance presumes that pivotal factors do exist in influencing the E-government development, and are controlled by the underlying causal laws (Heeks and Bailur 2007). Orlikowski and Baroudi (Orlikowski and Baroudi 1991) identified features of positivism studies as “evidence of formal propositions, quantifiable measures of variables, hypotheses tested, and the drawing of inferences about a phenomenon from the sample to a stated population”. The researchers holding this stance seek to find measurable pivotal variables including technological, social, and psychological issues which might influence the system outcomes, and to figure out the corresponding causal relationships. Accordingly, an empiricist epistemology will be placed by which the data gathered during the research are considered independent from the researchers who are observing and experimenting to acquire knowledge of such underlying causal relations (Heeks and Bailur 2007). Under such epistemological stances, quantitative research methods, such as large-scale questionnaire and survey, are primarily adopted.

Some works from this stream focus on technical issues that might influence users’ (both citizens and business sectors) adoption behaviour of different E-government services, whilst some others focus on social and psychological factors. In Venkatesh et al.’s work (Venkatesh et al. 2012), they conducted a web-based survey and identified four key factors that affect citizen’s intention of using transactional E-government services, which are usability, computer resource requirement, technical support provision, and secure provision. Schaupp et al. (Schaupp et al. 2010) evaluated the influence of factors such as performance expectancy, effort expectancy, trust, risk, social influence, and supporting facilities on U.S. taxpayers’ intention of using E-file. Lean et al. (Lean et al. 2009) conducted an exploratory study examining the Malaysian’ intention of using E-government and concluded that trust, perceived usefulness, perceived advantage and image have positive effect whilst perceived intention has a negative effect. With respect to business sectors-oriented services, some other works investigate the influence of E-government services assimilation on business value creation processes in organizations (Hossain et al. 2011).

By holding this deterministic view and conducting quantitative studies, researchers can observe and identify key factors which influence the system outcome, and explore the underlying causal laws in different cases. However, this stream of empirical study ignores the dynamic process and interactions among stakeholders which also play a key role. In addition, the particular social context is ignored which is critical for E-government research.

3.2 Qualitative Studies

In contrast, a social constructivism stance implies a subjective ontology by which the meaning of objects (even physical objects) assigned by different stakeholders matters the most. Under this stance, the focus of E-government research is to understand the meaning of this phenomenon constructed by each individual when using E-government services. It is assumed that the subjective understanding and interests of researchers cannot be detached from the meaning construction process neither (Heeks and Bailur 2007). Qualitative research methods such as unstructured interview and documentation analysis are applied.

Kamal et al. (Kamal et al. 2011) deployed a qualitative multiple case study approach to examine the role of different stakeholders, their perception of technology integration solutions in UK local governments, and their involvement in the adoption process. The reason of why those aspects are vital to such technology integration projects was emphasized and discussed. Cordella and Iannacci (Cordella and Iannacci 2010) proposed an e-Government enactment framework to analyse the intricacies involved in the deployed technology which is viewed as the carrier for achieving e-Government objectives. The complex relations among technologies and the political logic were also examined to investigate how they shaped the e-Government initiatives.

Most of the qualitative studies have tried to elucidate the recursive relationship among information technology, organizational structure, and social context and how the information technology was designed and deployed to achieve long-term interests inscribed in E-government initiatives by conducting theory-based case studies (Luna-Reyes and Gil-Garcia 2011). However, the qualitative studies that rely on verbal interpretation are too flexible to provide more rigorous quantifiable results due to its inherited characteristics, and new approaches are necessary to complement the theoretical discussions in order to acquire a better understanding of the relationships, consequences, and dynamic processes.

3.3 Agent-Based Simulation

To bridge the research gap left by extant studies with traditional methodologies, a new approach which can follow a holistic perspective is necessary and an agent-based modelling (ABM) approach (Gilbert 2008) is a promising candidate. “Simulation is a third way of doing science”, claimed by Axelrod (Axelrod 1997, pp. 5). He argued that deduction aims to derive logical consequences from a set of premises, whilst induction explores empirical data searching for any pattern. Different from these two traditional ways of doing science, simulation explores the data that are generated by simulation models embedding a set of pre-defined rules with a set of premises. By modelling a system and inputting specific data, researchers could observe the resulted emergent and unexpected output data even with simple embedded rules.

Across broad applicable areas, Axelrod (Axelrod 1997, pp. 3–4) summarized the purposes of simulation as “prediction, performance, training, entertainment, education, proof and discovery”. Different from rational individuals which are usually assumed in deduction, adaptive behaviours are modelled and simulated, through which the contingent consequences raised from non-linear rules could be analysed (Axelrod 1997). Agent-based simulation (modelling) is one of the major simulation paradigms. It is characterized as a “bottom-up” simulation approach which can capture micro-level individuals’ decision making and interactions among individuals and against the environment to analyse the resulted macro-level phenomena. This bottom-up approach is particularly useful to elucidate the underlying dynamic processes (Axelrod 1997).

ABM has been used extensively in both organizational innovation studies and innovation diffusion studies, respectively (Garcia 2005), but not in the service diffusion and adoption process studies yet. It can explore the interactions among involved stakeholders, and investigate how the interactions lead to collective behaviours. Besides, it can also be used to evaluate various supporting policies based on microscopic dynamics, and open new ways of collaboration among stakeholders. It is therefore a justified and promising approach to study the E-government phenomenon and to evaluate supporting strategies along with the service diffusion process since the whole process is dynamic and heterogeneous stakeholders are involved. The position of this research in terms of methodology is thus illustrated in Table 5.2. Entry “Macro-level” and “Micro-level” in the result analysis column refers to the result analyses from macro-level and micro-level perspective, respectively.

Table 5.2 Position of this work in E-government literature: Methodology
Full size table

In summary, we aim to scrutinize the E-government phenomenon from a service system perspective by applying an agent-based approach. The micro-level characteristics of involved stakeholders, individuals’ adaptive behaviours, interactions among stakeholders, as well as the resulted macro-level phenomenon are under investigation. Advancing understanding and insight on the dynamic behaviours of this service system at both micro-level and macro-level are expected to inspire better designed systems and strategies. Next, we will present a conceptual model of the service system, through which involved stakeholders are identified, explained, and integrated, and individual’s abstract adaptive behaviours and interactions among stakeholders are articulated in order to examine and understand the system more profoundly and thoroughly.

4 Conceptual Framework

We introduce the conceptual agent-based framework to study E-government from a holistic perspective in this section. We first identify key stakeholders directly involved in the system, and define their interactive behaviours subsequently.

Although public sectors could gain benefits through promoting and developing E-government systems, citizens will gain much more from the services and influence the services directly and indirectly. E-government systems therefore should be viewed and examined as a series of interactions between public sectors and citizens, whilst the divergent services provided act as the intermediate point (Akman et al. 2005).

The common stakeholders of E-government systems are end-users, such as citizens and private sectors (or business organizations), service providers, and employees of the service provider (Gouscos et al. 2007). For the public sector, we will treat it as one single stakeholder without distinguishing among different departments engaged in E-government services, although the departmental coordination (or one-stop E-government) is also an important research direction of E-government systems (Wimmer 2002; Gouscos et al. 2002; Dias and Rafael 2007). Employees of the service provider (i.e. staff working for government) are not considered as well, since citizens are the most important parts and to certain extent the benefit gained by employees are similar to that gained by citizens (Gouscos et al. 2007). In the following, we will go into the details of each stakeholder, respectively.

4.1 Public Services

According to Gronroos’s work (Gronroos 1988, 2001), the public service provided (transactional service and informational service) could be divided into three kinds: core services, facilitating services, and supporting services. Core services are the major functional services provided by individual departments, such as online transactional service. Facilitating services are aided services that help citizens complete the core services, such as computer resources and personal e-certificate that are necessary to do online transactions. Supporting services are value-added services that are often optional, such as technical support (Venkatesh et al. 2012). Services are evaluated against different criteria. For instance, usability is defined for core services, but not for the supporting services (Venkatesh et al. 2012).

Different services including transactional services and information provision services will be provided on two service channels, traditional office counter, and E-government websites. No significant differences are defined for the service provision between these two channels, but the consumed time and effort on each channel differ. For instance, online tax filing requires more efforts but less time compared to tax filing at traditional counters. Services provided through different channels will be evaluated against three major criteria, including easier, faster, and better services (Gouscos et al. 2007). Easier services and faster services are those require little effort and time to carry out, and better services are those with supporting services during the process.

4.2 Citizen-Side Learning and Channel Selection

The ultimate purpose of E-government is to provide services to citizens, and further to encourage them to engage in public affairs (Gouscos et al. 2007). There are no universally applicable services and thus the corresponding target users vary (Venkatesh et al. 2012). There is also a trade-off among service properties. For instance, citizens expect a higher level of security and privacy measures for online transactions, but may give up due to the complexity and poor usability involved in routine procedures (Venkatesh et al. 2012); but for other services, citizens just expect a convenient and easy-to-operate process without any security concerns.

There still exist discrepancies in the usage amongst target citizens, who have different preferences towards provided services associated with their education level, gender, and economic status (Gouscos et al. 2007). The preferences have impact on how citizens from different social groups evaluate multiple service channels and further influence the E-government usage behaviours. Different from innovation or product adoption, which is most of the time a one-time behaviour, the ability of taking up service will evolve along with the service improvements during the service diffusion processes. Citizens will learn adaptively how to use E-government systems either from their past experience or from the external environment including other citizens and the public sector.

Therefore, we will not design citizens as traditional rational agents who pursue to maximize their utilities, rather they are rule-based heterogeneous agents who will learn and behave adaptively against the environment. We categorize citizens into groups by their gender, educational level, and economic status (workers, students, and house makers). Each group values time and effort differently, therefore their selection of a service channel to take up the services will vary. Additionally, we allow citizens to improve their ability of using E-government systems through learning within their local community. Here the community composed of different kinds of citizens is treated as the “environment” of the model (Gilbert 2008). Deguchi’s social learning dynamics analytical model (Deguchi 2004) will be revised for modelling the adaptive learning behaviours.

4.3 Government-Side Resource Allocation

Public sectors hold two major objectives. One objective is to provide different kinds of services through multiple service channels, and technical support for using E-government, such as FAQ, email, and hotline services. In order to minimize the digital gap accounting to attributes such as educational level, the resource allocation in terms of user support should be improved continuously to satisfy the divergent requirements from all groups. The other objective is to propose and enact strategies promoting E-government services and facilitating citizens in using the services, such as public propaganda and educational workshops.

The promotion of E-government is considered as a way to increase citizen’s awareness of E-government services and to spread correct information/knowledge of E-government; and organizing educational programmes aims to prompt citizens’ ability of using technologies, such as organizing regular IT workshops within local communities and setting E-government self-help machines next to traditional service counters. Since education is a key factor of improving the uptake of E-government services (Gouscos et al. 2007), such educational programmes are expected to help citizens conquer barriers of using E-government services. Yet, the effectiveness of supporting strategies varies across social groups, thus optimized allocation of limited resources to address the needs fairly is critical.

We assume that the public sector will apply simple heuristics to allocate educational resources in an optimized way, such that the demands raised from different social groups can be satisfied simultaneously. We only consider E-government services in the general sense in this framework, and do not further distinguish particular services, adopted IT strategies and implemented initiatives in each local context.

4.4 A Holistic View

The interactions among stakeholders are realized through citizens’ usage of public services provided through different channels along with the improved public support, and illustrated in Fig. 5.1. Basically, different public services are provided on both channels: traditional counter and E-government enabled by the government. The government will implement supporting strategies to entice more citizens to utilize E-government, such as learning programmes helping citizens mitigate the effort of using E-government, and public propaganda increasing citizens’ awareness. Citizens will influence the public sector through the channel selection behaviour. More specifically, citizens will evaluate the channels by taking into account the service attributes, their corresponding preferences and received support from the public sector, and then choose a particular channel to take up services. Citizens will also learn from their off-line community to improve their abilities of using technologies, which may reduce the barriers of using online services. The resulted adoption rate will influence the allocation of supporting resources among different citizen groups, which will further influence the citizens’ channel selection through supporting services, as a dynamic process.

Fig. 5.1
figure 1

Conceptual Framework

Full size image

4.5 Gaming and Simulation

For an agent-based simulation approach, the validation process is one of the critical components. There are several ways to validate the models, in terms of guaranteeing consistency, against empirical data collected from the real world or stylized facts (Fagiolo et al. 2007). This stream of methods is very useful in reproducing specific social phenomena, yet our objective is to examine the link between underlying mechanisms and the resulted adoption behaviours. Ohori and Takahashi (Ohori and Takahashi 2012) introduced an analytical method for the validation of agent-based simulation, named “scenario analysis”. This analytical method differs from previous methods by which average results of simulation runs are presented. Rather it describes the result of each run without modification. This method has the advantage of taking divergent situations, design policies, and possible changes into account when interpreting the simulation results, which can facilitate the decision-making process.

Gaming is another useful tool to triangulate the simulation results when conducting scenario analyses. When designing agent-based models to examine the impact of learning mechanisms on citizens’ collective behaviours, we can unfold existing gaming protocols to integrate the learning mechanism. Public goods game has been widely applied in studying the social dilemma between individual interests and collective benefits, and especially suited to investigate such behaviours within groups (Camerer and Fehr 2004). In accordance with simulation models, we can design a game with a set of scenarios to be played by human players, and triangulate the results with those from simulation runs. We can also design the learning mechanism as a passive-learning process in the sense that it is embedded in the gaming session and calculated automatically whilst the players are informed by the learning results only. Alternatively, we can also allow the players to learn from their past experience or from the environment by limiting available information to the players. Based on the results and knowledge derived from gaming sessions as empirical evidence, we can further improve the agent-based model design. We can also introduce the participation of machinery agents to the games, which may induce unexpected behaviours of human players, and leverage the “wisdom of crowds” to explore collaboration mechanisms.

5 Concluding Remarks

In this work, we first reviewed the theoretical and methodological aspects of E-government phenomena. Based on the discussion, we argued that this phenomenon can be studied as an integrated service system considering the heterogeneous stakeholders and their interactive mechanisms from a bottom-up perspective, and thus new approaches should be adopted. This work serves as the first step towards understanding and designing E-government services by adopting a service system perspective.

We subsequently proposed an agent-based framework of E-government to identify the characteristics of heterogeneous stakeholders, their adaptive behaviours, and the interaction mechanisms among them from a service system perspective. Such integration provides a holistic view of the system which captures the dynamic service provision process and the macro-level phenomenon emerged from the micro-level interactions. It will complement studies carried by traditional approaches and advance the understanding of such systems from a bottom-up perspective. Future simulation works can be conducted based on this conceptual framework, through which the underlying dynamic process can be examined and analysed, and E-government supporting strategies can be proposed and evaluated.