1. Introduction
Both the private (see, e.g., Barrett et al., 2006) and public sectors (e.g., Cordella & Bonina, 2012) attach great importance to the introduction of information and communication technology (ICT). New technologies affect nearly all aspects of life, including how people interact with the authorities (Kumar et al., 2017). As underscored by Güler and Büyüközkan (2023, p. 1) the advent of digital technologies marks a transformative juncture for societies.
In the public sector, the use of information technology to enable and improve the efficiency of the provision of government services and information to citizens, employees, businesses and agencies is referred to as “e-governement” (Carter & Bélanger, 2005, p. 5). E-government and digital government are associated with positive effects for public organisations worldwide (Gil-Garcia & Flores-Zúñiga, 2020; Güler & Büyüközkan, 2023), such as cost reductions, improved services, time savings and increased effectiveness and efficiency (Alshehri & Drew, 2010). Furthermore, through the digital provision of services, interacting with government organisations is no longer limited to working hours or office locations (Kumar et al., 2017).
Despite the organisational benefits evident in e-government initiatives, the acceptance of such solutions by individuals remains variable, contingent upon diverse factors. The success of e-government initiatives hinges upon citizens’ willingness to adopt these services (Carter & Bélanger, 2005, p. 6). In Europe, the adoption of e-government initiatives, such as electronic services, is relatively low. For instance, the percentage of individuals (ages 16 to 74) within the EU that use the Internet for interaction with public authorities was only 57% in 2018. (Pérez-Morote et al., 2020, p. 1). An examination of the use of e-government services by the European population revealed a notable lack of widespread adoption, indicating a critical barrier to realising the potential benefits of e-government initiatives. Governments aim to increase citizen acceptance of their e-government initiatives to promote their use (Carter et al., 2016) and hence in the long run, profit from the positive effects associated with e-government. Consequently, there is a pressing need for comprehensive research into the determinants shaping citizens’ demand for and utilization of such services (Ma & Zheng, 2018).
Not all e-government initiatives, such the provision of a new electronic service through the introduction of an information system, are successful, and sometimes people are reluctant to use and adopt e-government solutions (Distel, 2020; Rey-Moreno et al., 2018). Different scholars have investigated the antecedents of e-government adoption and acceptance. Many draw on well-established models and frameworks from information systems literature, such as the technology acceptance model (TAM) (see, e.g., Wang, 2003), the unified theory of acceptance and use of technology (UTAUT) (see, e.g., Israel & Tiwari, 2011), its extension UTAUT 2 (see, e.g. Alharbi et al., 2017), and Roger’s diffusion of innovation (DOI). It is common for researchers to combine parts of these and other models into a new research model to fit their particular research focus (see, e.g., Carter & Bélanger, 2005; Sang & Lee, 2009). The models are well-suited to explain behavioural intentions for using a particular solution and the actual use of a technology (see, e.g., Davis, 1989; Venkatesh et al., 2003).
What remains unclear is which factors can lead to technology acceptance in cases where the solution in question has not yet been implemented. The models fail to explain how acceptance works in these cases as they contain components that are only relevant to existing information systems (e.g., usability aspects like “perceived ease of use”) and explain “usage” rather than “acceptance”. The question remains which factors lead people to accept e-government solutions as more of a conceptual idea than an actual technological solution. Further, the models fail to consider personal characteristics such as digital literacy or trust in government, which might be relevant for e-government acceptance. In the context of e-government, it is assumed that social, political, and cultural influence factors are important for technology acceptance (Al-Hujran et al., 2015).
This study aims to address the research gap by investigating the pre-implementation public acceptance of e-government applications and constructing an integrated model encompassing individual factors by examining a specific case of an e-government initiative. Specifically, this research investigates factors influencing voter acceptance of the direct-democratic ballot concerning the introduction of electronic proof of identity (e-ID) in Switzerland in March 2021. Accordingly, the research question of this article is:
What are the information system characteristics, individual factors (including socio-demographic aspects, competencies, and values), and trust- and political-orientation-related factors that influence public acceptance of e-ID?
Electronic identification (e-ID) technology plays a crucial role in ensuring secure authentication for digital services, facilitating online identification, authentication, and signing (Kemppainen et al., 2023; Melin et al., 2016). The area of e-ID is of particular importance in the context of government initiatives, as the provision of e-Government and e-services depends on identity management (Tsap et al., 2020b, p. 340).
The introduction of e-ID interacts with a country’s governmental and societal context (Mettler & Guenduez, 2019), while the direct democratic system in Switzerland permits an examination of public acceptance of specific policies, such as e-government initiatives. The subject of this study is a referendum held on creating a legal basis for introducing an e-ID solution, which the majority of Swiss voters rejected in March 2021. To investigate the factors influencing pre-implementation acceptance of an e-ID, this study utilises the VOX dataset, comprising a post-ballot survey of approximately 3,000 eligible voters (Swissvotes, 2021). Furthermore, to address shortcomings in existing technology acceptance models, we propose a revised public information system acceptance model – the government information system acceptance model (GISAM). The GISAM incorporates factors from various technology acceptance models and previous studies, tailored to explain e-government acceptance at a conceptual stage rather than an established developed e-government solution. This case is particularly suited to explain ex-ante acceptance of an e-ID system because investigating a ballot proposal in a direct-democratic context offers valuable insights into the attitudes and opinions of the general public towards a new technology. The VOX dataset offers rich data that not only allows to test different theory-based potential influence factors but also a set of potential control variables. By integrating these factors and combining them in a new model (GISAM), we pursue an integrative approach towards public (ex-ante) technology acceptance. The clear direct-democratic rejection of an e-government solution that is the subject of this study underpins the findings of studies suggesting that citizens may be strongly opposed to e-government solutions (Rey-Moreno et al., 2018).
Given governments’ vested interest in expanding e-government solutions, understanding the determinants of e-government acceptance is of paramount importance. By identifying the factors driving acceptance, governments can tailor initiatives to increase public acceptance, thereby maximising associated benefits such as public value and o operational efficiency.
This article is structured as follows: It commences with a theoretical framework, which consists of an explanation of the need for e-IDs, an overview of global developments regarding e-IDs, in general, and through a Swiss lens in particular, before explaining the current state of research regarding different dimensions of public acceptance of e-ID solutions. From these theoretical foundations, the proposed GISAM is derived and explained. Subsequently, the methodology and the findings are presented, followed by a discussion is and conclusions.
2. Electronic Identification with e-IDs
2.1 The need for e-IDs
One dimension of e-government is the digitisation of public service delivery (Cahlikova, 2021), which changes how governments interact with citizens (Lindgren et al., 2019). As new technologies (e.g., artificial intelligence or sensor technologies) push into e-government service delivery with an expectation that such developments will continue in the future (Lindgren et al., 2019), it is important to understand how such technologies are accepted. If not accepted, an e-ID solution is unlikely to be adopted by citizens (Crompton, 2010).
Another dimension of e-government is the electronic practice of democracy that comprises the digitisation of “existing democratic practices” (e.g., e-voting) and the development of “new ones that become a possibility in the electronic environment” (e.g., e-participation) (Cahlikova, 2021, p. 28). A necessary condition for e-government in general and e-voting in particular is the possibility of a secure, unique system (Kofler et al., 2003; Zwattendorfer & Slamanig, 2015) that allows for “identification, authentication and digital signing via the internet” (Melin et al., 2016, p. 2). While in traditional voting settings, the physical features of an individual can be compared with official documents (e.g., a passport or identity card) to verify someone’s identity, this cannot be done in a digital setting (Lentner & Parycek, 2016). Therefore, corresponding identification in the digital space is needed to ensure safe e-government services and prevent misuse, such as identity fraud. Identity management stands as a cornerstone in the development of e-government and the provision of electronic services, upon which e-government and electronic service provision heavily depend (Tsap et al., 2020b, p. 340). According to Gartner (2023) “Identity and Access Management (IAM) is a security and business discipline that includes multiple technologies and business processes to help the right people or machines to access the right assets at the right time for the right reasons, while keeping unauthorized access and fraud at bay”. In the government context, one means of ensuring a safe way of identity and access management for identification purposes is the introduction of electronic identity (e-ID) (Melin et al., 2016). Attempts and successful introductions of e-IDs have been discussed in literature and practice for a long time (see e.g., Kanwar et al., 2022; Tsap et al., 2019).
The human urge to produce and collect data can be traced back to the Upper Palaeolithic period (35,000–10,000 BCE), where the first attempts at markings to represent things took place (Beynon-Davies, 2021). Today, in the information age, governments collect significant amounts of citizen data, sometimes without their explicit knowledge (Stoycheff et al., 2019). This manifests itself, for example, in governments increasingly accessing citizens’ electronic personal data in connection with the provision of e-government services (Taylor & Lips, 2008) and might influence the foundations of citizen-government relationships (Lips, 2010). Of course, the aim of identifying citizens through technical means is not new to e-IDs. Indeed, it has existed for a long time through means of identification such as “[…] caller-line identification, pin numbers, […] smart cards […], biometric systems […], facial and number plate recognition systems […] and internet-based access to government websites supported by varying types and intensities of identification and authentication” (Taylor & Lips, 2008, p. 144). However, with an e-ID, they might reach new dimensions.
When introducing e-IDs, van Dijck & Jacobs (2020) argued that the legal and technical operability often dominates the design of an e-ID, while public concerns about personal data (e.g. privacy and security concerns, user empowerment and control over one’s own data) are neglected in the discussion. Particularly concerning biometric data, which is often the case in e-IDs, it can be argued that there is potential for violating people’s right to privacy (Barkane, 2022).
2.2 Global state of the introduction of e-IDs
Different implementation systems and application areas make it difficult to compare e-IDs among countries in detail. Nevertheless, it can be said that e-ID efforts are ongoing in most countries worldwide (Kanwar et al., 2022), even if their degrees of maturity vary significantly (Mettler & Guenduez, 2019). In the European Union, the approaches and implementation of e-IDs vary considerably despite plans to establish a European digital identity (Fitri, 2022). As Fitri (2022) explained, only five European countries (Belgium, Estonia, Finland, Norway and Sweden) have managed to establish e-IDs that are actively used by at least 40 per cent of the population. While most other European countries have managed to implement national digital identity schemes, their adoption faces privacy concerns (e.g., the UK) (Fitri, 2022). COVID-19 has served as a driver for establishing digital identities for retrieving government services and their interoperability (Aldane, 2022). In an international working group, Australia, Canada, Finland, Israel, New Zealand, Singapore, the Netherlands, and the UK have shared their experiences with digital identities, including the reaction to COVID-19, regarding requirements for “mutual recognition, and/or interoperability between [the] member countries” (Commonwealth of Australia (Digital Transformation Agency), 2022, p. 3).
From an institutional point of view, interactions with other actors (e.g. from the financial sector) influence the design of e-IDs and e-government (Cordella & Paletti, 2019; Eaton et al., 2018), and administrative, financial and technical constraints have to be taken into consideration (Bostan et al., 2017). Consequently, countries vary not only in their degree of e-ID adoption but also in cooperating with private organisations offering e-ID solutions, ranging from solely private (e.g., Sweden) or public-private partnerships (e.g. Austria) to public responsibility (e.g., Germany) (Guenduez et al., 2022). Van Dijck & Jacobs (2020, p. 896) argued that e-IDs based on decentralised attribute-based systems on a non-profit and non-state basis might be established due to “[p]ublic concerns such as privacy, security, user empowerment and control over one’s personal information” as they allow users to “control the design of a system’s architecture as well as data management” (p. 899).
From an individual point of view, Kalvet et al. (2018, p. 431) emphasised that the general acceptance of a centralised database of identity documents is high, while “acceptance of the inclusion of fingerprints, iris, etc. into database varies substantially across countries.” Other studies, such as the literature review by Tsap et al. (2019), argued that cultural and historical factors may influence e-ID acceptance.
2.3 Developments in Switzerland
In Switzerland, initial e-government initiatives date back to the mid-1990s (Chappelet, 2004). Today, different strategic documents at the federal, cantonal, and local government levels guide the introduction of e-government initiatives in Switzerland (Cahlikova, 2021). Despite the strategic importance attributed to e-government, owing to a lack of human resources and a legal basis, Swiss municipalities have failed to satisfy the public need for digital service delivery (Buess et al., 2019). As Cahlikova (2021, p. 73) explained, there is a further hurdle to the introduction of e-government in Switzerland: “The absence of a central organ that would send a strong signal of the government’s commitment to implementing e-Government constitutes an important obstacle to its general development.” Consequently, Switzerland currently ranks 30th out of 36 countries in Europe concerning overall e-government maturity (European Commission et al., 2021).
Switzerland introduced the so-called SwissID, an electronic identification option (Sury, 2020), in 2017 (Bühlmann, 2017). It is a service provided by the SwissSign Group, a joint venture between government-related businesses, financial companies, insurance companies and health insurance companies. This allows access to various services in the private and public sectors through a single account (Plattner, 2022). As Sury (2020) explained, a condition of the SwissID was to fulfil the planned requirements of the e-ID Act to be recognised as a state identification unit once the e-ID Act came into force. However, since this has not yet happened, the state does not recognise the SwissID (Sury, 2020). As with any form of e-participation (Cahlikova, 2015), introducing an e-ID solution acceptable to Swiss citizens, the government and the private sector has proven complex owing to the different requirements of these stakeholders (e.g., concerning policies and regulations, access to data and ICT resources, accountability and manageability of the solution (Sialm & Knittl, 2016)). In Switzerland, its citizens and parliament are the legislators, and voters must approve changes in federal law (Swiss Parliament, 2022). Accordingly, a referendum was held in March 2021 to decide on the planned e-ID Act – which is the basis for the analysis in this article – and this was rejected by a majority of Swiss voters (Swiss Federal Department of Justice and Police EJPD, 2021). The proposed e-ID Act is described in more detail in Chapter 4.1.
3. Literature Review: Public Acceptance of e-ID Solutions
E-government describes “the use of information and communication technologies (ICT) […] to improve the efficiency, effectiveness, transparency, accountability, and activities of public sector organizations” (Sang & Lee, 2009, p. 71) and has been developed since the early 1990s in accordance of the rise of the internet (Bannister & Connolly, 2012). Organisational change, such as changes in structures and operations of public organisations, are often associated with e-government initiatives, and in the more developed parts of the world, e-government is widely used (Twizeyimana & Andersson, 2019). However, in the context of e-government, individuals and their needs as stakeholders are of central importance (Pleger et al., 2021), while people may refuse to use or adopt e-government solutions (Rey-Moreno et al., 2018), sometimes because of their lack of attention to user needs (Crompton, 2010; Pleger et al., 2020; Singh et al., 2022) or lack of value and privacy concerns (Aichholzer & Strauß, 2010). It is assumed that e-government adoption is influenced “by the reciprocal interactions of personal, behavioural, and environmental factors” (Zhao et al., 2019, p. 1). Understanding the factors that lead people to accept e-government initiatives is therefore crucial for governments when introducing new e-government solutions efficiently and in accordance with people’s needs.
When assessing technology acceptance, scholars often draw on various ICT acceptance models to explain the acceptance of e-government. One of the foremost models is the technology acceptance model (TAM) (Hofmann et al., 2012). The TAM (Davis, 1985, 1989; Davis et al., 1989) is based on the theories of planned behaviour (TPB) and reasoned action (TRA) and aims to explain technology acceptance based on a user’s internal beliefs, attitudes, and intentions (Davis et al., 1989). Another often-used model, the unified theory of acceptance and use of technology (UTAUT), was developed by Venkatesh et al. (2003) by combining different technology acceptance models and approaches such as the theory of reasoned action, theory of planned behaviour, TAM, diffusion of innovation, motivational model, social cognitive theory, model of PC utilisation, and a combined theory of behaviour and TAM. However, these models focus on technology adoption in cases where a technological solution already exists and are therefore not well-suited to explaining the ex-ante acceptance of e-government solutions. If a solution does not yet exist, many components are still unknown because they are yet to be defined. The acceptance of such a solution is therefore more akin to the acceptance of a theoretical construct than an existing solution, which is why we assume that other factors are important for the acceptance process.
3.1 Dimensions of public acceptance of e-ID solutions
Previous research has identified several factors that might influence the acceptance of e-ID solutions. These factors can be aligned in different dimensions – individual socio-demographic characteristics, individual competencies and values, attitudes information system characteristics, trust, and political orientation. The following sections expand on these dimensions.
Dimension 1: Individual socio-demographic characteristics
The first dimension covers individual socio-demographic characteristics independent of the e-service topic, namely e-ID, and includes age, gender, and education. The UTAUT (Venkatesh et al., 2003) and previous studies (ELKheshin & Saleeb, 2020; Tiits et al., 2014) suggest that these factors will have a moderating influence on e-ID acceptance – but overall, the evidence is mixed (Kalvet et al., 2018). Additionally, Zhao et al. (2019) found that personal factors (age, gender, education, and income) paired with behavioural factors and personal factors paired with environmental factors influence e-government adoption.
Dimension 2: Individual competencies and values
Dimension 2 covers competencies and values unique to every individual that might influence their acceptance behaviour. Dahi and Ezziane (2015), Schepers and Wetzels (2007), and Xie et al. (2017) all suggested that an individual’s subjective norms play a role in e-ID acceptance. Subjective norms describe an individual’s perception or belief that others think they should (or should not) behave in a certain way (Dahi & Ezziane, 2015; Schepers & Wetzels, 2007). It is similar to the “social influence” factor investigated by Chan et al. (2010) and “perceived compatibility” by Carter and Bélanger (2005). However, research that includes subjective norms in the context of technology adoption has produced mixed results (Schepers & Wetzels, 2007).1
Digital literacy, as another aspect of interest, includes factors like “understanding”, “seeing reasons/purpose”, “knowing how to use”, and “comprehending” – demonstrating that an individual understands how a technical system works and can use it (Tsap et al., 2019, p. 181). An earlier study by Harbach et al. (2013, p. 10) claimed that “participants expressed a general reluctance to adopt new services or technologies on the Internet, due to a feeling of insecurity and negative reports in the news. They showed no interest or motivation to gain an understanding of a new mechanism.” These individuals, therefore, express a low degree of digital literacy, and their attitude towards an e-government solution might be different if they possessed a higher degree of digital literacy. In line with this assumption, the findings of Tsap et al. (2020a, p. 167) indicate that “in general users are knowledgeable and tech-savvy. They demonstrate knowledge of potential risks when it comes to security and privacy, capabilities and limitations of the existing system, principles of its functioning, etc.” Therefore, digital literacy is an essential factor when understanding user acceptance (Hofmann et al., 2012; Tsap et al., 2020a; Venkatesh et al., 2003).2
Dimension 3: Attitudes towards information system characteristics
The third dimension includes anticipated information system (IS) characteristics, a common dimension in established technology acceptance models (see e.g., Davis, 1985; Venkatesh et al., 2003). In 2021, the year the referendum was held, the SwissID had two million users (SwissSign AG, 2021) compared to 8.7 million inhabitants (Federal Statistical Office (FSO), 2022). Within these two million SwissID users were also people without voting rights (e.g., foreign citizens) that were in possession of the SwissID for practical reasons, such as public transportation, signing documents, authentication of e-mail communication. Of the 8.7 million inhabitants, 5.5 million were Swiss citizens with voting rights (Federal Statistical Office, 2022). It is unknown, how many of the SwissID holders in 2021 were Swiss citizens and therefore allowed to vote on the proposed eID Act. In an extreme scenario, where every SwissID holder was a Swiss citizen, the percentage of citizens in possession of the SwissID would be 36 per cent – the actual value is expected to be lower due to the wide range of application areas of the SwissID outside of services bound to Swiss nationals. Therefore, it can be argued that IS characteristics played a subordinate role in the decision of citizens to vote for or against the referendum. However, as the IS characteristics are dominant in many technology acceptance models, these factors are also included in this study.
Perceived risk is a construct that consists of “importance and probability of loss” (Dowling, 1986, p. 194). Various studies have shown that the risks citizens perceive may influence their intention to use an e-government service (Carter & Bélanger, 2005; Xie et al., 2017). The perceived risks associated with the introduction of e-ID include “document forgery”, “identity theft” (Tiits et al., 2014, p. 4) or “the use of personal information for purposes other than those initially stated”, “surveillance”, and “collection of personal data” (Kalvet et al., 2018, p. 431). Axelsson and Melin (2012) studied Swedish student attitudes towards e-ID, and through focus groups, they identified usability and security as critical factors in trusting e-ID and e-government in general.
Privacy concerns are “associated with risks, fears, threats to citizens’ right to be violated in relation to their digital identities” (Tsap et al., 2019). This concept is also linked to trust (Tsap et al., 2020a). Privacy is paramount regarding identity management and identification documents issued by governments (Kalvet et al., 2018). Privacy concerns in the context of an e-ID from an individual point of view can stem from fear of violation of a “citizen’s right to personal privacy in general and data privacy in particular” (Beynon-Davies, 2006, p. 17).
Performance expectancy is one of the factors contained in the UTAUT (Venkatesh et al., 2003), and it describes what and how citizens can expect to benefit from a technological solution (Tiits et al., 2014). In a recent survey study, (Tsap et al., 2020a) investigated public acceptance and preferences for e-ID from an Estonian perspective – a country where “more than 2/3 of citizens regularly use e-ID today” (p. 159). Their findings suggest that citizens favour IS characteristics such as convenience, security, speed, and co-existing multiple authentication methods. In another article (Tsap et al., 2020b, p. 341) identify three key features of e-ID acceptance: convenience, security, and speed. Each of these features contributes to the overall ‘ease of use’ factor.
Dimension 4: Trust and political orientation
The fourth dimension includes individual factors about trust and political views, and trust contains the dimensions “reliability” and “integrity” (Carter & Bélanger, 2005). Accordingly, trust in government in the context of e-government can be described as the perception of a government’s reliability and integrity when offering electronic services (Carter & Bélanger, 2005). Trust in government is an important aspect of e-government acceptance as it reduces individual risk perceptions, e.g., regarding data misuse (Dutton et al., 2005) and is an indicator of confidence in a government’s digital competencies (Distel, 2020; Hofmann et al., 2012; Sialm & Knittl, 2016). Trust in government has previously been identified as having a strong effect on e-government adoption (Carter & Bélanger, 2005; Distel, 2020; Tsap et al., 2019; Xie et al., 2017) and e-ID adoption (Tsap et al., 2019). A previous study by Tiits et al. (2014) in the context of ePassports found that the public trusted the government with identity codes and photos but was more sceptical about storing eye iris images and DNA, financial data, and location data. Citizens expect ePassports to safeguard their data from forgery, be accurate and reliable in identifying individuals, and offer better protection from identity theft. Consequently, the most significant risks associated with ePassports are misuse of personal data and surveillance (Tiits et al., 2014). A study by Kalvet et al. (2018) suggests that people trust the government to issue secure identity documents. However, people also perceive risks with government-issued identity documents, namely, “the use of personal information for purposes other than those initially stated (function creep) and covert surveillance or collection of personal data by government (Kalvet et al., 2018, p. 431).”
In the solution proposed by the Swiss government about which the referendum was held, issuing e-ID could have been undertaken by private or public partner organisations (Swiss Federal Department of Justice and Police EJPD, 2021). To launch successful e-ID solutions, Crompton (2010, p. 297) advises governments to cooperate with other partners as “a comprehensive, centralised stand-alone approach is almost certain to fail.” Guenduez et al. (2022) showed that trust in private businesses (paired with trust in the government) might influence the likelihood of the involvement of a private partner in public e-ID issuance. For this reason, trust in private businesses is also relevant to this study. For example, Tiits et al. (2014) ascertained that people were more reluctant to share their biometric data with private than public organisations.
Another important dimension is trust in the internet, which reflects trust in online services or technology in general (Tsap et al., 2019). Previous studies have identified trust in the internet as relevant in e-government adoption (Carter & Bélanger, 2005; Hofmann et al., 2012), while others could not confirm this relationship (Distel, 2020). It, therefore, seems worth investigating.
Like other individual factors, political orientation or political interest are largely ignored in traditional technology acceptance models such as the TAM, and their influence remains unclear (Brown et al., 2002). However, since political orientation can be related to other factors, such as trust (Chen et al., 2017), we consider it essential to include “political orientation” and “political interest” in this study. Furthermore, voting can be viewed as an expression of political behaviour that, in turn, can be influenced by a person’s political values (Leimgruber, 2011). Therefore, in the direct-democratic Swiss context, it can be assumed that political orientation influenced the acceptance of the proposed e-ID solution.
3.2 Integrated public acceptance model: GISAM
As indicated above, existing technology acceptance models are not well-suited for the specific research context of this study as they focus on cases where a technological solution has already been established. For example, the TAM and the UTAUT measure intention to use a system and actual usage, whereas in our case, we investigate factors relevant before a system has been established. Since our concern is the ex-ante acceptance of a system, we suggest a new model include factors that might influence acceptance based on previous research, particularly studies related to e-IDs and ePassports.
As a revised and integrated approach to address the public acceptance of government information systems, we propose a new model – the GISAM. GISAM stands for “government information system acceptance model”, and it consists of several factors derived from the literature that can be assigned to four dimensions: (i) individual socio-demographic characteristics, (ii) individual competencies and values, (iii) information system characteristics, and (iv) trust and political orientation. The model is illustrated in Figure 1.
Government information system acceptance model (GISAM).
4 Study Design
In the following, the research design will be presented, starting with an overview over the ballot proposal on electronic proof of identity, before the data, method and operationalization are presented in detail. The research design consists of a multi-stage procedure with a quantitative and a qualitative analysis. Appendix A.3 provides a schematic overview of the research design.
4.1 Ballot proposal on electronic proof of identity
The case selected for the analysis of public e-service acceptance was a referendum on the introduction of electronic proof of identity and the adaptation of an associated law in Switzerland held on 7 March 2021. The Federal Act on Electronic Identification Services (e-ID Act) was intended to regulate how people would be uniquely identified through e-ID so they could order goods or services efficiently and securely online. The proposed law specified that possessing an e-ID would be voluntary, and examples from everyday life were used to highlight the advantages, such as when applying for an official document (Swiss Federal Department of Justice and Police EJPD, 2021) (see also Appendix A.2.). Furthermore, the Federal Council would have been responsible for ensuring that the individual’s personal information was correct according to existing registers before forwarding the request to the issuer (i.e., a private company, a canton, or a municipality), therefore providing a high level of assurance of the e-ID. The goal was to act quickly by collaborating with different partners (Swiss Federal Department of Justice and Police EJPD, 2021). The law also regulated data protection between all partners and defined the purpose of data use and data sharing (Swiss Federal Department of Justice and Police EJPD, 2021).
Legislation on electronic proof of identity (e-ID) was subject to a ballot proposal, which was defeated by a clear majority – with 64.36 per cent against and only 35.64 per cent in favour of the proposal. Voter turnout was 51.27 per cent, equating to 2,762,625 votes cast (Swiss Federal Department of Justice and Police EJPD, 2021).
4.2 Data, method and operationalisation
The data for this study was sourced from the VOX dataset, which is conducted by gfs.bern after every federal vote.3 Gfs.bern, a survey institute specialising in opinion formation in Switzerland and direct democratic processes, conducts a representative survey on behalf of the Federal Chancellery. This survey, known as the VOX survey, involves around 3,000 randomly selected voters. Respondents are asked about their participation, reasons for approval or rejection, the opinion-forming process, and socio-demographic characteristics through both online and paper questionnaires. The VOX survey is a collaboration between gfs.bern and the political science departments of the universities of Bern, Geneva, and Zurich.
Our analysis employed logistic regressions with Bayesian approaches, using MLwiN 3.02 for data analysis. The dependent variable was the respondents’ voting decisions, represented as a dichotomous variable indicating either a “yes” or “no” vote on the e-ID proposal. As shown in Table 1, both the national data and our sample indicated that 36% voted in favour and 64% against the proposal, with no differences between the sample and national results. The independent variables were operationalizations of those captured in the GISAM. After data cleansing, the final dataset included 2,092 individual responses.
Table 1
National results on the electronic identity compared to the sample results.
| NATIONAL RESULTS | SAMPLE RESULTS | |
|---|---|---|
| per cent (N) | per cent (N) | |
| “Yes” votes | 36% | 36% |
| “No” votes | 64% | 64% |
| Total N | 2,762,625 | 2,092 |
Source: Own representation; data for the national results from the Swiss Confederation, Federal Chancellery.
In addition to quantitative regression analysis, we content-coded responses to an open-ended question about the reasons for voting decisions. This qualitative analysis complemented the logistic regression models by capturing dimensions related to “information system characteristics”. Due to the nature of the dataset, some variables, specifically privacy concerns and performance expectancy, required qualitative analysis for measurement. Response behaviour for open-ended questions can vary since they do not restrict respondents to a specific set of options (Fowler Jr. & Cosenza, 2009). Analysing open-ended questions is appropriate when the list of possible answers is extensive, as is the case in this study (Fowler Jr. & Cosenza, 2009). By incorporating both open- and close-ended questions, we ensured comprehensive capture of all relevant dimensions. For the qualitative analysis of open-ended reasons given for their voting decisions, only those individuals who indicated they had voted in favour of the e-ID proposal and disclosed the reason behind their voting decision were included.
It is important to note that the variables were sourced from an existing dataset, limiting detailed adjustment of their operationalization. Appendix A.1 provides the exact assignment and operationalization of the individual variables.
5 Results
Our findings consist of (i) quantitative logistic regression and (ii) the content-coding of open-ended questions, which are presented in turn.
5.1 Results regression analysis
The quantitative analysis consists of eleven variables, four of which were found to be systematically related to voting decisions. The results of the logistic regression analysis are shown in Table 2. The results are presented along the four GISAM dimensions. An overview of the variables, their operationalisations and summary statistics can be found in Appendix A.1.
Table 2
Voter acceptance determinants of electronic proof of identity.
| DIMENSION | DETERMINANTS | MODEL 1 | CI | |||
|---|---|---|---|---|---|---|
| MEAN | S.D. | 2.5% | 97.5% | |||
| Individual socio-demographic characteristics | Constant | 7.398 | 5.163 | –5.949 | 16.716 | |
| Age | –0.003 | 0.003 | –0.008 | 0.004 | ||
| Gender (ref. male) | –0.190 | 0.168 | –0.519 | 0.140 | ||
| Education level low (ref. medium) | 0.604 | 0.305 | 0.003 | 1.199 | ||
| Education level high (ref. medium) | 0.124 | 0.185 | –0.238 | 0.490 | ||
| Individual competencies and values | Subjective norms related to digitisation | 0.064 | 0.063 | –0.059 | 0.185 | |
| Digital literacy/Attitude towards digitisation | –0.459 | 0.063 | –0.562 | –0.337 | ||
| Attitudes towards information system characteristics | Perceived risks | –1.053 | 0.118 | –1.282 | –0.823 | |
| Trust and political orientation | Trust in government | 0.026 | 0.046 | –0.065 | 0.116 | |
| Trust in private businesses | 0.506 | 0.041 | 0.428 | 0.588 | ||
| Political orientation | 0.160 | 0.049 | 0.064 | 0.256 | ||
| Political interest | –0.125 | 0.133 | –0.390 | 0.132 | ||
| DIC: | 1007.298 | |||||
Notes: Dependent variable is the voting decision at the ballot (dichotomous variable “yes/no”). Models ran with MLwiN 3.02 through MCMC estimation. DIC = deviance information criterion. Bold: 95%-credible interval does not contain zero (systematic relationship). Logit-model; posterior mean, standard deviations (S.D.) and 95% credible interval of log odds, based on Bayesian estimation (100,000 iterations, burn-in: 500, thinning: 1).
Dimension 1: Individual socio-demographic characteristics
As shown in Table 2, no systematic relationship was found between the variables from the dimension “individual socio-demographic characteristics” and the voting decision.
Dimension 2: Individual competencies and values
With regard to the second dimension, individual competencies and values, the variable subjective norms related to digitisation did not have a systematic impact on voting in favour of the introduction of the e-ID. In contrast, the results indicate that digital literacy/attitude towards digitisation plays a crucial role in supporting the e-ID. Digital literacy/attitude towards digitisation was measured by the agreement to a scale between “digitisation connects society and simplifies life (= 1) and digitisation promotes anonymity and brings social problems (= 6).” The findings indicate that reservations about digitisation have a negative impact on acceptance. As shown in Figure 2, the predicted probability of voting “yes” decreases with an increased distrust of digitisation. The probability of voting in favour of e-ID introduction when holding all other variables constant was nine per cent (95%-CI 0.06–0.12) among those who fully agreed with the statement that “digitisation promotes anonymity and brings social problems.” When in full agreement with the opposing statement that “digitisation connects society and simplifies life,” and holding all other variables constant, the probability of voting “yes” was 48 per cent (95%-CI 0.40–0.56).
Attitude toward digitisation and probability of voting “yes”.
Note: Agreement on a scale between “digitisation networks society and simplifies life” (= 1) or “digitisation promotes anonymity and brings social problems” (= 6).
Dimension 3: Attitudes towards information system characteristics
The third dimension, information system characteristics, was captured by the variable perceived risks. The variable was measured according to agreement with the statement, “the use of the e-ID is recorded at a private company and stored centrally. This creates a potential for fraud.” The results suggest a systematic negative relationship between the perceived risk and the probability of accepting the e-ID introduction: the stronger the agreement with this statement, the lower the probability of accepting e-ID. Accordingly, trust in private providers seems to play a crucial role in voter acceptance of e-ID (Table 2).
Dimension 4: Trust and political orientation
Regarding the fourth dimension, “trust and political orientation”, two out of the four variables were found to have a systematic impact on the voting decision. The results show that trust in private businesses is decisive in e-ID acceptance. The higher the confidence in private electronic identity providers, the higher the probability of voting “yes”. Keeping all other variables constant, the mean probability of accepting the ballot proposal when having the lowest trust in private providers was seven per cent (95%-CI 0.05–0.09). In contrast, holding all other variables constant, the mean probability of accepting the ballot was 87 per cent (95%-CI 0.80–0.92) among those with the highest trust in a private provider’s anti-fraud assurances for their acceptance, as shown in Figure 3.
Trust in private electronic identity providers and probability of voting “yes”.
Note: Self-rating trust in private providers of digital ID cards on a scale between no trust at all (=0) and complete trust (=10)
Another factor found to impact the voting decision systematically was political orientation. However, this relationship is less linear than the other systematic influencing factors. Accordingly, people with strong left-wing views seemed to oppose the ballot proposal, whereas people with a right-wing orientation were relatively equally divided. For both variables, trust in government and political interest, no systematic relationship could be found.
5.2 Coding of the results
Coding was used to analyse in greater depth the main reasons why the vote was accepted. The VOX dataset contains an open-ended question about why the proposal for the Federal Act on Electronic Identification Services (e-ID Act) was accepted/rejected (depending on the stated voting decision). To examine reasons for acceptance, only those respondents who reported voting “yes” in the ballot were considered. The open-ended responses given to that question represent the data being coded. Before coding, a pretest of 100 open-ended responses was conducted to estimate the inter-coder reliability between two coders. Based on the pretest results, modifications and clarifications of the coding rules were applied. As a result, the coding of the two variables (Figure 5 & 5) was based on coding by a single coder, and 497 open-ended responses were processed.4
The first step sought to examine how many people based their voting decision on considerations related to e-ID. Therefore, the open-ended reasons for accepting e-ID were coded dichotomously, each open-ended response being coded either 0 or 1, depending on whether the answer was directly linked to e-ID. As shown in Figure 4, most reasons (57%) for voting in favour of e-ID were not directly related to e-ID, while 41 per cent were (N = 497). Examples of open-ended responses that were coded as related to e-ID included “there is a need for E-ID in Switzerland”, “simplified and uniform identity verification”, “legally standardised foundation for general security in electronic handling”, and “simplified log-ins, secure online shopping”. In contrast, examples for reasons that were coded as having no link to e-ID included “recommendation government” and “we must move forward with regards to digitisation processes”. This dichotomous coding was done in a first step to discover how prominent the E-ID itself was for the reason of the vote. For further coding, however, all answers of those who had voted yes were coded.
Coding results for a link between reason to accept and e-ID.
Note: Bars represent percentages (N = 497).
Coding results of reasons for accepting e-ID.
Note: Bars represent percentages. The question read: “What was the main reason you accepted the proposal for the Federal Act on Electronic Identification Services (e-ID Act)? (N = 497).
To be able to depict the GISAM holistically, a second coding step was then performed. Here, the open reasons for the acceptance decision were coded based on the variables influencing technology acceptance in the literature and shown in the GISAM model. The variables were operationalised as six coding categories: (i) “performance expectancy: standardisation”, (ii) “necessity digital progress: international competition”, (iii) “necessity digital progress in general”, (iv) “privacy concerns/perceived risk/data security”, (v) “performance expectancy: positive evaluation of e-ID without reasons”, and (vi) “performance expectancy: simplification/advantages of e-ID itself”.
In addition, six other coding categories derived from the control variables were added – digital literacy/attitude towards digitisation, subjective norms related to digitisation, political orientation, trust in the internet, trust in private businesses, and trust in government. Finally, the coding categories “not applicable” and “no reason against” were added to the code book
As shown in Figure 5, the most frequent reason (30%) for accepting the e-ID proposal relates to the perceived necessity of digital progress in general. Examples of this category of responses were, “digitisation must be driven forward”, “finally a step in the direction of digitisation”, “an important step for the digital future”, and “digitisation should not be held back”. The second most common category given as a reason for a “yes” vote relates to performance expectations regarding the simplification or benefits of e-ID. A reason pertaining to this category was given by 21 per cent of respondents, and examples for this category include “facilitation of online services”, “facilitated presentation of my personal data”, and “that you do not have to have a lot of code ready for various e-businesses”. Around 9 per cent of respondents’ answers contained aspects of privacy concerns, perceived risk or data security like “secure identification on the internet” or “less identity fraud”.
Interestingly, none of the responses fell into the category “trust in the internet”, which is a component of the “individual competencies and values” dimension of the GISAM. Similarly, only around one per cent of the open-ended responses could be assigned to “subjective norms related to digitisation” and “political orientation”.
6 Discussion
Besides many studies investigating e-government adoption success factors in situations where a solution has already been implemented, it remains unclear what factors lead to prospective acceptance of e-government in cases where no tangible solution yet exists. Existing models measure the actual design of information systems by investigating usability factors. However, we have consistently argued that acceptance can also refer to support for an information system before its introduction. At this point, information system characteristics (such as intuitive usability) are entirely unknown to the target groups. Therefore, other factors must be relevant for individuals when accepting or rejecting an e-government solution. Indeed, our findings corroborate the notion that in the context of e-government implementation, as exemplified by the case study of introducing an e-ID system, acceptance is influenced not only by information system characteristics but also by more abstract factors. What seems far more relevant are beliefs and attitudes that promote digitisation processes and the implementation context of such a system.
Figure 6 summarises the integrated findings of the qualitative and quantitative analysis. At first glance, the results of our qualitative and our quantitative study may seem contradictory to some extend. Some systematic correlations that emerged from the quantitative analysis could not be confirmed by the qualitative analysis. However, these differences are partly due to the different methodological approaches of the two methods and partly due to the study design. The primary goal of the qualitative coding investigation was to capture the dimension ‘information system characteristics’ and, in particular, the two factors ‘privacy concerns’ and ‘performance expectations’. These factors could not be covered by the quantitative analysis because of the available data of the questionnaire. The basis of the qualitative analysis was an open-ended question in the survey in which the yes-voters could give the reasons for their voting decision. These open-ended responses were then coded. In each case, only the first answer given was coded. For the quantitative analysis, we drew on existing closed question in the survey. These different ways of collecting data may lead to different results, since in one case respondents are supported by existing items, while in open-ended questions respondents must formulate their answers themselves. Therefore, Figure 6 shows the results including their methodological origin. The results complement rather than contradict each other. The qualitative analysis shows that the factors ‘subjective norms towards digitisation’, ‘performance expectancy’ and ‘privacy concerns’ were mentioned spontaneously by respondents, while ‘trust in private businesses’ and ‘political orientation’ were mentioned only rarely. These factors thus appear to be relevant for the respondents without being at the forefront as reasons for acceptance. This lack of evidence for the relevance of trust and political orientation in the qualitative analysis can be explained by the fact that these two variables are more relevant to rejection than to acceptance of the referendum, and the qualitative analysis only examined the responses of those who voted ‘yes’.
Factors from the GISAM impacting the public acceptance of e-government applications.
Note: Factors that produced systematic results in the quantitative analysis are shown in white boxes and in bold font. Factors from the qualitative analysis are shown in white boxes and in regular font. Factors showing influences in both analyses are marked by +. No influence could be detected for the factors in grey boxes.
As shown in Figure 6, our results suggest all the information system’s characteristics, were relevant for Swiss citizens when accepting the proposed e-ID solution. The factor perceived risk produced systematic results in the quantitative analysis that is supported by the evidence of the qualitative analysis. The other two factors (“privacy concerns” and “performance expectancy”) showed a certain influence in the qualitative analysis. Confirming the information system characteristics, our study supports existing technology acceptance models that emphasize the importance of these factors (Davis, 1985; Venkatesh et al., 2003).
In addition to these factors, our results suggest that more intangible factors such as digital literacy/attitude towards digitisation, subjective norms towards digitisation, trust in private businesses and left-wing political orientation are also relevant for acceptance. These results contradict the general assumption in former technology acceptance studies that “attitudes” are irrelevant to technology adoption (Davis et al., 1989; Venkatesh et al., 2003). However, our results are in line with previous research results in an e-government context that have identified “attitudes” (Al-Hujran et al., 2015; Axelsson & Melin, 2012; Hung et al., 2013; Mahadeo, 2009; Xie et al., 2017) and “value compatibility” (Mahadeo, 2009) as important factors in e-government acceptance. These conflicting views might stem from the voluntary character of e-government; other than in a work setting, people can choose whether to use an e-government solution (Al-Hujran et al., 2015). Our results also indicate that a positive attitude towards digitisation is associated with greater e-ID acceptance, supporting the argument that personal attitude is a critical factor in technology acceptance. This can be influenced by emphasising perceived public value and perceived ease of use of the solution in question, which, in combination, can lead to technology acceptance (Al-Hujran et al., 2015).
Consistent with our results, other scholars have stressed the importance of trust in e-government acceptance (Alzahrani et al., 2017; Tsap et al., 2019). Alzahrani et al. (2017) argued that trust in e-government consists of four dimensions influencing e-government acceptance – individual characteristics, technical factors, government agency factors, and risk factors. In our study, trust in private companies was essential. When governments outsource services to private or public partners, they remain responsible for the outcome and must protect individual rights (Argento & Peda, 2015). Our results suggest that not all citizens trust the government to safeguard their rights in the case of e-ID. Accordingly, acceptance is higher in those individuals who trust not only the government but also private companies. Guenduez et al. (2022) consistently found that in countries where trust in both government and the economy is high, e-credential markets are more likely to be competitive and not solely state-dominated. Trust in an e-ID solution can be enhanced through end-user involvement (i.e., by considering their attitudes towards the e-ID solution), thereby increasing the likelihood of acceptance (Axelsson & Melin, 2012).
Our findings support earlier research results suggesting that individual characteristics such as age and gender have a subordinate role in e-ID acceptance (Kalvet et al., 2018). This further confirms previous research that e-government acceptance is highly complex and has various political, social and cultural influencing factors (Al-Hujran et al., 2015). Knowing the critical factors is particularly important for a country such as Switzerland, where voters can reject a proposal (including e-ID) before it is implemented (Cahlikova, 2021). In this respect, the factors that our study shows to be relevant must be considered before attempting to re-introduce e-ID.
7 Conclusion
This article proposes a revised acceptance model called GISAM that focuses on acceptance exclusively (as opposed to intention to use), with acceptance defined as an internal approval intention of the target group before actual introduction. An empirical study aimed to apply the GISAM to a specific case and thereby examine the acceptance of information systems within the public sector in more depth by using the case of a popular vote on the introduction of electronic proof of identity (e-ID) in Switzerland. Using a survey data set among a representative sample of 2,092 Swiss citizens, factors influencing the voting decision and the open-ended reasons for accepting the ballot proposal were investigated.
7.1 Theoretical and practical implications
Public acceptance of information systems is crucial when implementing new digital information systems in the public sector. Although an extensive strand of research on technology acceptance exists to explain the adoption of e-government, we have argued that existing technology acceptance models share two common shortcomings concerning the public acceptance of e-government initiatives. The first is that current models largely ignore individual factors, such as values or attitudes, which are particularly important in general. The second shortcoming is evident in the operationalisation of acceptance, which is mainly measured by utility variables of the information system (i.e., variables that address the handling of the new IT system) and can only be evaluated after implementation.
In addition to theoretical implications, the study also allows for the derivation of recommendations for practice and academia. The results indicate that various factors influence the acceptance of information systems. Concerning e-ID, the findings testify to an elevated level of engagement with the issue among voters. However, this engagement is directed less towards e-ID and more toward digitisation processes in general. For example, our coding revealed that the main reason for accepting the ballot was not directly linked to e-ID but consisted of the perceived necessity of progress regarding digitisation.
Furthermore, the regression results showed that trust in private companies who would have issued e-ID and individual attitudes towards digitisation were significant factors in e-ID acceptance. For practice, these findings mean that the relevance of the fundamental attitude toward digitisation processes should be anticipated by policymakers.
Firstly, it seems crucial to address individual attitudes towards digitisation. Findings suggest that individual attitudes towards digitisation significantly influence technology adoption in the context of e-ID. Governments should focus on understanding and addressing the concerns, fears, and preferences of individuals regarding digitisation and information systems. For example, this aspect can be specifically addressed in voting campaigns to achieve greater acceptance among those who are critical of digitisation processes. The findings support the importance of information systems providers and their interaction with the state. Decision-makers should therefore emphasise this aspect when planning and implementing new information systems, since – depending on the political context – provision by public agencies can lead to greater acceptance. While this study recognizes the positive aspects of e-government from an organizational perspective, it is important to communicate these benefits to the general public as well. Governments should communicate how the introduction of information systems can improve the efficiency, transparency, accessibility and quality of services in the public sector.
A second practical recommendation that emerges from our results addresses public concerns related to data security and the privacy of personal information as these aspects appear to be crucial for promoting acceptance. Governments should prioritise robust security measures, data protection regulations, and transparent information handling practices. Building trust in the security and confidentiality of information systems can alleviate concerns and encourage acceptance. By implementing these practical recommendations, governments can work towards promoting acceptance of information systems in the public sector, thereby enhancing the effectiveness and efficiency of e-government initiative.
Academia can use the findings of this study to rethink technology acceptance in cases where a proposed solution has not been established yet. This stream of research is rather underrepresented and deserves more attention in the future. We show, that in such cases, other factors that are commonly not considered in traditional technology acceptance models – including individual attitudes – may play a role. Further studies can apply and validate the GISAM to different technology acceptance cases, as the application of the GISAM in this study focussed specifically on e-ID acceptance.
7.2 Limitations
Inevitably, our study has several strengths and limitations. By analysing a vote in a direct democratic environment, we were able to actually investigate the ex-ante acceptance of an e-ID solution and confirm the shortcomings of existing technology acceptance models. However, at the time of the study, more than one-fifth of the Swiss population already possessed the SwissID – the thought predecessor of the e-ID solution. This fact should not be neglected when interpreting our results as these people may have been influenced by this existing identification solution when voting. Another limitation of the study manifests itself in the difficulty of implementing theoretical models empirically due to a reduction of complexity and data availability. While all theoretical models have to rely on simplified assumptions to make the analysis tractable, limited data availability is a constraint of our study. Variables were taken from an existing data set. Therefore, an operationalisation adjustment was only possible to a very limited extent. It should also be noted that the variables in the model were measured using one-item scales. However, some of the variables are likely to be latent in nature.
7.3 Strength and future research
In spite of these limitations, however, our study has great strengths. The direct-democratic context of Switzerland allows for a unique research setting where ex-ante acceptance of an IS can be investigated. Our findings therefore offer valuable insights in IS acceptance cases, where an IS has not been established. The direct-democratic rejection of the e-ID solution underscores the relevance of understanding factors driving e-government acceptance within specific national contexts and from a citizen point of view. While some results are in-line with previous research, others show that in the case of Switzerland, different factors may have played a role. The study therefore enriches the field of literature by highlighting the importance of country-specific factors, while recognizing the factors already found in other studies that seem generally applicable when examining the success of e-IDs.
We offer an integrated model that includes factors that are neglected in other IS acceptance models, such as the Technology Acceptance Model (TAM) or the Unified Theory of Acceptance and Use of Technology Model (UTAUT) and combine them with other research-backed factors to provide a new perspective to IS acceptance research and practice. Another strength is the quality and size of the data that the study is based on. In conclusion, our findings suggest that acceptance takes place much more abstractly than conventional acceptance models assume. Consequently, convictions and values toward advancing digital processes are generally crucial when a solution has not yet been implemented. At the same time, underlying values or personal predispositions – such as digital literacy/attitude towards digitisation or faith in online security – do not appear to play a significant role in acceptance.
In future work, there are several avenues for extending research based on the findings of this study. First, as explained in the limitations section, some of the variables in the GISAM model are latent. These require a multi-item approach in order to measure them accurately. Therefore, it would be interesting for future research to measure the model with more precise variables using multiple items. Secondly, it seems to be a promising approach to investigate more deeply the dynamics of ex-ante technology acceptance in general and in the context of e-government initiatives in particular. Future research should also aim to develop a more nuanced understanding of the multiple influences on public acceptance of e-government applications: This paper investigates factors contributing to the acceptance of the implementation of eID as a distinct government initiative. However, it should be noted that this analysis does not yield definitive insights into the overarching determinants governing acceptance of e-government initiatives as a whole. Consequently, future studies might consider examining multiple e-government initiatives to elucidate the underlying determinants of acceptance, thereby uncovering broader patterns and principles applicable across various contexts.