Unlocking citizen confidence: examining trust and continuance intentions in digital services

The rapid digitalization of government services worldwide has led to an increased focus on electronic government (e-gov) implementations and their effectiveness. It has the potential to empower the public, promote electronic democracy, improve convenience, ensure transparency and quickly adapt to the evolving needs of the public (Dwivedi et al., 2017; Kurfalı et al., 2017; Zhang and Zhu, 2021). E-government, also referred to as digital government or electronic governance, is the use of information and communication technologies (ICT) by government institutions to provide services, share information and facilitate interactions with citizens, businesses and other stakeholders. There is no single agreed-upon definition, but it generally encompasses digital platforms that enhance transparency, efficiency and service delivery (Dwivedi et al., 2017). Adopting and continuously using e-gov services can also contribute to sustainable development by improving service access, minimizing resource use and increasing transparency. Through the adoption of digital solutions, e-gov can mitigate the environmental impact of traditional paper-based systems and enable economic and social progress.

In developing nations, particularly in South Asia, the success of e-gov initiatives is crucial for improving public service delivery and fostering citizen engagement. Pakistan, like many developing countries, has been investing in e-gov infrastructure to enhance public service efficiency and accessibility. However, the mere implementation of e-gov platforms does not guarantee their adoption or continued use by citizens. Pakistan ranks low in terms of online population growth compared to other developing countries (e.g., Democratic Republic of the Congo (126%), Iraq (55%), Egypt (22%), India (23%), Pakistan (17%) (Statista, 2020). In addition, online fraud and mishaps undermine citizens’ trust in government and services. For example, the recent hacking of Pakistani banks and the sale of data on the dark web for an estimated $19.9m (GroupIB, 2019). This further weakens individuals’ trust in government and services. Trust in e-gov plays an important role in determining the intention of citizens to adopt e-gov services and to continue to use them. However, in the Pakistani context, there is a paucity of research that has focused specifically on citizens’ trust in e-gov services. Trust in these systems plays a pivotal role in determining their success and sustainability.

Recent e-government literature from different perspectives has reported on factors that play an important role in service adoption, including trust in government or electronic loyalty to government and trust in technology or the internet (Abu-Shanab, 2014; Abu-Shanab, 2021; Alzahrani et al., 2017; Carter et al., 2016; Xie et al., 2017). Trust positively influences citizens’ intention to adopt e-gov services (Verkijika and De Wet, 2018). In the Pakistani context, a few studies have examined factors influencing citizens toward e-gov in Pakistan (Ahmad et al., 2013; Rehman et al., 2016; Zahid et al., 2022). Although research shows the instrumental value of trust in assessing citizens’ behavior toward e-gov adoption (Dwivedi et al., 2017), little is known about the role of trust in the adoption of e-gov services in Pakistan from the perspective of citizens (Ahmad et al., 2013; Khan et al., 2021; Zahid et al., 2022). Therefore, it is important to investigate and understand citizens’ trust in the continuous use of e-gov services in Pakistan. To examine the determinants of e-gov services adoption and continuance in Pakistan, this study extends the trust model proposed by Carter et al. (2016) by integrating it with the Expectation-Confirmation Model (ECM) developed by Bhattacherjee (2001). This integrated approach allows for a comprehensive examination of both trust factors and user satisfaction in the context of e-gov services. The primary research question guiding this study is as follows:

By addressing this question, we aim to:

• identify key factors that contribute to citizens’ trust in e-gov services in Pakistan;

• examine the relationship between trust, satisfaction and continuance intention in the use of e-gov services;

• explore applicability of the integrated trust-ECM model in Pakistani context; and

• provide insights for policymakers and e-gov practitioners to enhance the adoption and sustained use of e-gov services in Pakistan and similar developing nations.

To answer the research question, an online survey was conducted with Pakistani citizens who have experience using e-gov services. The survey instrument was designed to measure various constructs from the integrated trust-ECM model, including perceived usefulness, confirmation of expectations, satisfaction, trust in government, trust in technology and intention to continue using e-gov services. The research data was analyzed using structural equation modelling (SEM). The findings of this study are expected to stimulate discussion among the e-gov research community, particularly in Pakistan and neighboring South Asian developing countries, to improve their e-gov services and related initiatives.

The remainder of this study is organized as follows: Section 2 provides background and related work; Section 3 outlines the conceptual model and hypotheses; Section 4 details research methods, including sampling procedures and measures; Section 5 presents results; Section 6 offers a discussion with theoretical and practical implications; and Section 7 concludes the paper and suggests future research directions.

Evolution of Pakistan Electronic Government Directorate dates back to its inception in 2002 and has undergone several transformative phases. A significant milestone was reached in 2014 with the merger of the Pakistan Computer Bureau and the Electronic Government Directorate, leading to the formation of the National Information Technology Board (NITB). Notable initiatives include the National Database and Registration Authority, online tax payment systems, electronic driving licenses, land records management and e-visa services, all accessible through a centralized portal (www.pakistan.gov.pk).

Despite making progress in providing online services, Pakistan still ranks relatively low in global e-government development indices. The UN E-Government Development Index ranked Pakistan 148th in 2018 and 153 in 2020 (UN, 2020). The current e-government landscape in Pakistan is primarily characterized by an information stage, with limited fully interactive portals. According to Statista’s, 2020 data, citizens around the world are increasingly raising concerns about online privacy with their governments. In particular, countries such as the USA and Hong Kong each saw an 80% share of respondents expressing concern, closely followed by India at 79% and Pakistan at 67% (Statista, 2020).

Research on e-gov in Pakistan focuses on two main dimensions:

1. citizen-centered e-gov services, which include user access, satisfaction and the availability of services; and

2. government-centered e-gov services, which relate to the government’s role in providing transparent, efficient and secure digital services to citizens.

Various theoretical models have been used in Pakistani e-gov research, including: include technology acceptance model, diffusion of innovation and information systems success model (Akram and Malik, 2012; Javaid and Arfeen, 2017; Kamal et al., 2013; Rehman et al., 2012a, 2012b, 2016; Xie et al., 2017), unified theory of technology acceptance and use (Ahmad et al., 2013; Azam et al., 2013) (see Table 1). Within the citizen-centered e-gov services area, researchers are tackling a wide range of issues including quality of service, availability, usability and privacy concerns. In this context, research examines capabilities, policies and various other organizational aspects. Key constructs examined in these studies include trust in agency, perceived responsiveness, transparency and quality of information (Arfeen and Kamal, 2014; Arshad and Khurram, 2020; Ahmad et al., 2016; Hassan and Lee, 2015; Khan et al., 2021; Zahid et al., 2022; Zahid and Haji Din, 2019).

E-gov implementation in Pakistan faces numerous challenges encompassing both citizen concerns and organizational capacity issues. Citizens face a variety of challenges, including a lack of adequate ICT infrastructure and awareness (Kamal et al., 2013; Rehman et al., 2016; Zahid et al., 2022), a lack of transactional services, outdated information, broken web links and widespread trust issues due to concerns about the security of personal and financial information (Rehman et al., 2012a, 2012b). Similarly, organizational capacity and capability pose substantial challenges. These include governance challenges, political instability, weak policy implementation, a poor regulatory environment, security and privacy concerns, a shortage of skilled personnel to implement ICT projects, resource constraints, capacity building constraints, cultural considerations and the government’s struggle to identify different demographic groups and tailor electronic service delivery to their needs (Hassan and Lee, 2015; Arfeen and Kamal, 2014; Javaid and Arfeen, 2017). Furthermore, online fraud and errors diminish individuals’ trust in government and services, as demonstrated by the hacking of Pakistani banks and the selling of data on the dark web for an estimated $19.9m (GroupIB, 2019). This further weakens individuals’ trust in government and services. Governments must ensure internet-based services are secure and trustworthy to increase adoption and reduce reliance on traditional offline government interactions.

Recent studies in Pakistan have explored e-gov adoption from various perspectives, including marketing, sustainability and its role in addressing challenges posed by the COVID-19 pandemic (Zahid and Haji Din, 2019; Ullah et al., 2021; Mansoor, 2021). However, the overall literature evaluating e-gov services in Pakistan remains limited, reflecting a trend observed in many developing countries (Dwivedi et al., 2017). According to Carter et al. (2016), trust in e-government services plays a significant role in their adoption. Furthermore, studies in the Pakistani context have confirmed the importance of these factors (Zahid et al., 2022). Tufail et al. (2017) explored how citizens’ trust in e-gov services and their satisfaction influence e-service adoption in Pakistan, finding that both constructs play a pivotal role in shaping adoption behavior, although using a different model than the current study.

The importance of trust has also been reported in studies outside of the Pakistani e-gov context. Literature highlighted that trust e-gov is mostly influenced by two factors: trust in government and trust in technology/internet (Abu-Shanab, 2014; Carter and Weerakkody, 2008; Connolly and Bannister, 2007; Horst et al., 2007; Pavlou, 2003; Smith, 2011; Warkentin et al., 2002). Building trust between e-gov and citizens is very important to overcome perceptions of uncertainty and vulnerability (Bélanger and Carter, 2008). In a number of e-services, perceived trustworthiness plays a significant role in influencing intention to use. Examples include tax filing (Bélanger and Carter, 2008) and medical information systems (Zimmer et al., 2010).

Several studies have explored the factors influencing citizens’ adoption of e-government services. These include ease of use, perceived risk, reliability, relative advantage, trust, image, facilitating conditions, cultural differences, lack of service and security awareness, self-efficacy, etc. (Ahmad et al., 2013; Aladwani, 2013; Arshad and Khurram, 2020; Bélanger and Carter, 2008; Carter et al., 2016; Irani et al., 2009; Janssen et al., 2018). Our model integrates two theoretical frameworks: the Expectation-Confirmation Model (see Figure 1) (Bhattacherjee, 2001) and Carter et al. (2016) trust model.

Expectation-Confirmation Model (ECM) posits that continuance intention is driven by post-adoption satisfaction and perceived usefulness, which are shaped by confirmation of pre-adoption expectations (Bhattacherjee, 2001). In the context of this study, ECM serves as a theoretical foundation for investigating user engagement with mobile government (m-government) and electronic government (e-government) services (Mandari and Koloseni, 2023). ECM to capture how user satisfaction, perceived usefulness and confirmation drive continued usage. ECM key constructs definitions are in  Appendix. According to the theory, individuals initially form expectations regarding a system or service before engaging with it. After using the system, users evaluate whether their prior expectations align with the actual functionality and benefits provided. This confirmation process influences their level of satisfaction and perception of the system’s usefulness, ultimately shaping their decision to either continue or discontinue usage (Bhattacherjee, 2001).

Trust is a critical determinant of e-gov adoption, as it influences users willingness to interact with and rely on digital services. Carter et al. (2016) proposed a trust-based framework that trust in government (TOG), trust in the internet (TOI), disposition to trust (DT) and perceived risk (PR) as predictors of e-service adoption (see Figure 2). Carter et al. (2016) model constructs definitions are in  Appendix. This model suggests that individuals’ trust in governmental institutions plays a crucial role in shaping their perceptions of e-gov reliability, security and credibility. Additionally, trust in the internet reflects users’ confidence in the technological infrastructure, security mechanisms and privacy safeguards that facilitate e-gov transactions. Trust mitigates perceived risk, which directly affects usage intentions. A key contribution of the Carter et al. (2016) model is its emphasis on the interplay between perceived risk and trust formation. Given the intangible nature of online transactions, citizens often perceive digital interactions with the government as risk-laden, particularly concerning data security, privacy and potential misuse of personal information. The model posits that when trust in government institutions is high, citizens are more likely to perceive e-go services as secure and reliable, thereby mitigating risk concerns. Similarly, a high level of trust in the internet reduces skepticism about technological vulnerabilities, increasing users’ confidence in digital service platforms. By focusing on the antecedents of trust and risk perceptions, the Carter et al. (2016) model primarily explains the initial adoption stage of e-gov services. It provides a strong foundation for understanding how trust influences citizens’ willingness to engage with digital platforms but does not explicitly address the factors affecting continued usage beyond initial adoption. This limitation underscores the necessity of integrating additional theoretical frameworks that explain post-adoption behaviors.

While Carter et al. (2016) emphasizes trust and risk factors in the initial adoption of e-gov services, Bhattacherjee’s (2001) ECM provides a well-established framework for understanding post-adoption behavior. The decision to integrate these two models stems from their complementary nature. The integration of these models is theoretically justified based on several factors.

First, trust is not only a determinant of initial adoption but also plays a pivotal role in users’ long-term engagement with e-gov services. Once users adopt an e-gov service, their trust in government and the internet continues to shape their perceptions of system reliability, security and usefulness. By incorporating trust constructs into ECM, the model extends its explanatory power beyond confirmation and satisfaction, capturing a more nuanced understanding of continuance intention. Second, perceived risk, a central construct in Carter et al. (2016), remains relevant in the post-adoption phase. Users may continue to evaluate risks associated with privacy breaches, data security and system reliability even after initial adoption.

ECM, on its own, does not explicitly account for these concerns; thus, integrating perceived risk into the framework enhances its applicability to e-gov services. Third, merging these models aligns with contemporary research on e-gov adoption, which suggests that a multi-theoretical approach provides a more comprehensive understanding of user behavior (Bélanger and Carter, 2008). Given that e-gov services often involve sensitive personal data and mandatory interactions, an integrated approach ensures that both trust formation (pre-adoption) and expectation-confirmation (post-adoption) processes are accounted for in a single explanatory model (see Figure 3). We only take the hypothesis and investigated in this study which were lacking in existing literature (see Table 1). Our model theoretical fusion enhances our ability to understand, predict and improve citizen engagement with e-gov services.

Trust serves as a social lubricant, facilitating trade and lowering friction costs while also broadening the scope of commercial ties (Carson et al., 2003). Trust is a critical factor in the adoption of e-government and other electronic services. Research indicates that trust has a significant influence on the use of e-government platforms (Connolly and Bannister, 2007; Mcknight et al., 2011; Smith, 2011), and perceived trustworthiness has influenced intentions to use various e-gov services (Bélanger and Carter, 2008), health information systems (Zimmer et al., 2010) and so on. Trust in government is a key predictor of e-gov services, and trust is needed when risks are present (Featherman and Pavlou, 2003). Perceived risk mediates the effect of trust on intention (Bélanger and Carter, 2008; Pavlou, 2003; Warkentin et al., 2002). The internet is a medium used to provide all kinds of digital services. These services need to be reliable and secure for transactions. Therefore, we hypothesize that citizens’ trust in the internet plays an important role in e-gov usage:

Citizens’ trust in the internet and service-providing agencies has a symbiotic relationship and a strong influence on service adoption (Warkentin et al., 2002). Trust in the internet and government significantly predicts initial trust in e-government services (Azam et al., 2013). Governments have a responsibility to develop and deliver secure services in a transparent manner. “[…] many people refer to the government as ‘big brother,’ a ubiquitous presence that infringes upon constituents’ personal lives” (Bélanger and Carter, 2008). Trust in government has an impact on technology adoption (Gefen et al., 2003). It is essential that citizens believe in the ability of e-gov to deliver secure services. In addition, communication must ensure that e-gov services are for their benefit and not for citizen surveillance. Citizens associate trust with the breach of their personal and financial information, which discourages them from using and adopting services (Rehman et al., 2012a, 2012b). Therefore, our second hypothesis is as follows:

Trust could be seen as more subjective, given that each person is somewhat different and is inclined to trust in different ways. At the abstract level disposition to trust “is a propensity or tendency to believe or not to believe in the attributes of others in general” (Gefen et al., 2003; McKnight et al., 2004). DI can also be referred to as personality-based trust (Gefen et al., 2003). The disposition to trust can be divided into Faith in Humanity (assuming that others are good-natured and reliable) and Trusting Stance (assuming that better outcomes will result from the transaction). Thus, the following hypotheses are proposed:

In technology adoption, trust serves as a key factor in reducing users’ perceived risks and uncertainties (Carter et al., 2016). In situations where risk and uncertainties are involved, the role of trust is even more pronounced (Pavlou, 2003), influencing individuals’ willingness to make decisions amid uncertainty. Trust, characterized by a balance of perceived gains and losses, is derived from individuals’ risk assessments (Rousseau et al., 1998). Notably, in the domain of e-government services, the perceived risk tends to be higher at the transactional level compared to the informational level (Horst et al., 2007). Trust acts as a mitigating factor, reducing perceived risk and thereby encouraging the utilization of e-services across both public and private sectors (Featherman and Pavlou, 2003; Bélanger and Carter, 2008; Pavlou, 2003). In light of these arguments, we hypothesized that:

ECM describes how user’s thoughts and beliefs, like satisfaction and perceived usefulness, influence their decision to keep using an information system (Bhattacherjee, 2001). According to the theory, continued usage is linked to satisfaction after using the system and its perceived usefulness. These factors, along with whether the system met the user’s expectations, contribute to overall satisfaction. Furthermore, the model suggests that perceived usefulness can be affected by how well the user’s expectations were met. In our study, we focused on four key elements of the ECM: continuance intention, satisfaction, perceived usefulness and confirmation. Confirmation can be perceived as the degree to which a user perceives the actual performance of a system to meet their expectations following its use (Bhattacherjee, 2001). Perceived usefulness represents how much an individual expects getting operational or strategic benefits from using an information system (Bhattacherjee, 2001). Continuance intention refers to a user’s willingness to continue using information systems, services or technology. Satisfaction reflects the amount to which a user feels positively impacted by their earlier encounters with a system (Bhattacherjee, 2001). Extensive research has proved the ECM’s application and validation in a variety of scenarios, including online commerce, mobile wallets and medical information system continuing intention (Chen, 2020; Gupta et al., 2020; Wu et al., 2020):

In this section, we present the research methodology, which includes sampling, data gathering and analysis procedures used.

This study aimed to investigate the perception and trust of e-government services among Pakistani citizens from a user’s perspective. An online survey methodology was used to collect data. The target population consisted of Pakistani citizens with internet access who have used e-government services. To ensure the inclusion of only relevant participants, a filter question regarding prior use of e-gov services was incorporated into the survey. Only respondents who reported using e-gov services at least once were included in the study.

The survey was distributed through various online channels, including personal emails and social media platforms. This approach was chosen to reach a diverse group of e-gov service users across Pakistan. The survey remained open for a period of two months, from November 2022 to January 2023.

A total of 641 responses were received during the data collection period. After data cleaning, which involved removing responses with missing information and those that failed an attention checker question, the final sample consisted of 558 valid responses (86.1% of the original submissions).

The demographic profile of respondents is summarized in Table 2. To assess the representativeness of the sample, a comparison was made with national demographics. For instance, the sample consists of 78.32% males, which is somewhat higher than the estimated percentage of male internet users in Pakistan. This overrepresentation of males may reflect greater internet access among this group. In terms of age, education and employment status, the sample corresponds reasonably well with national statistics on internet use in Pakistan. The alignment of the sample with national internet usage statistics provides confidence in the generalizability of the results, although the slight overrepresentation of males should be considered when interpreting the findings.

The online survey consisted of three sections; the first introduced the objectives of the study second section included the background of the respondents such as demographic information, internet experience, etc. The last section included the trust model of Carter et al. (2016) as a baseline and extended it with ECM (Bhattacherjee, 2001). The use of previously validated instruments helps to improve the reliability of the instrument. These variables have been adapted to the specific context of our research. To validate the questionnaire clarity and interpretability, we conducted a focus group with 10 Pakistani citizens. Participants were asked to assess each items meaning, and their feedback led to revision to enhance respondent comprehension. Subsequently, a pilot study involving 20 Pakistani citizens was conducted to further refine the questionnaire and ensure the validity of its constructs. A detailed explanation of the measured constructs follows.

All the items used to measure trust in internet (3 items), trust in government (3 items), disposition to trust (3 items), perceived risk (2 items), were adopted from Carter et al. (2016). Whereas, the items measure satisfaction (4 items), perceived usefulness (4 items), confirmation (3 items) and continuance intention (3 items) are adopted from Bhattacherjee (2001). Details on the items measured are given in  Appendix. We used a seven-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). While response quality is a concern for all data collection methods, it is particularly critical for online surveys due to their remote nature (Zickar and Keith, 2023). To address concerns about participant attention and data accuracy, we employed attention checks, such as the item “If you are paying attention, respond ‘agree’ to this item.” These checks help gauge participant engagement and comprehension.

We used partial least squares structural equation modeling (PLS-SEM) to evaluate our research model. PLS-SEM is a widely recognized data analysis method, particularly suited for complex studies involving predictive modeling and theory development (Hair et al., 2017). While covariance-based SEM (CB-SEM) is often preferred for strict theory testing due to its provision of global fit indices, we chose PLS-SEM for several reasons. First, our study aims to both test theoretical relationships and predict key outcomes, aligning with PLS-SEM strengths in predictive modeling. Second, our model includes multiple constructs with numerous indicators, making PLS-SEM ability to handle complex models advantageous. Third, PLS-SEM is more robust to violations of multivariate normality, which may be present in our data. Additionally, we followed the two-stage approach to PLS-SEM implementation proposed by Sarstedt et al. (2021), which provides a systematic framework for applying PLS-SEM rigorously. Although CB-SEM offers global fit statistics that are valuable for theory confirmation, recent advancements in PLS-SEM, such as consistent PLS, have enhanced its applicability for confirmatory research. Nevertheless, we acknowledge the limitations of PLS-SEM, particularly the absence of global fit indices, and discuss how this may influence the interpretation of our findings in the limitations section.

Dash and Paul (2021) compared the performance of partial least squares structural equation modeling and covariance-based structural equation modeling in establishing structural relationships. Their findings suggest that both methods can achieve comparable results when applied to processed and standardized data. While CB-SEM has traditionally been the preferred approach, recent studies have demonstrated the versatility of SEM in exploring and assessing different model configurations. Given these advantages, we use SEM for our data analysis and hypothesis testing. Furthermore, SEM is also useful in studies where there are complex causal relationships (Arshad and Khurram, 2020).

In this section, we discuss the results of our study including the measurement and model as well as hypothesis testing.

The use of questionnaires to collect data on external and internal factors may be susceptible to common method bias (CMB), potentially compromise the validity of research findings. CMB is a pervasive issue in behavioral sciences (Nguyen et al., 2024). To minimize the potential for CMB in our study, we implemented several preventive measures prior to data collection. These measures included explicitly informing participants that there were no correct or incorrect answers and ensuring the anonymity of their responses.

We adopted a two-stage approach to evaluate the model relationships (Hair et al., 2014). The first phase assesses data fitness and validity using a variety of established methods, while the second part examine the model structural relationships.

To assess common-method bias (CMB) we applied the Lindell-Whitney marker-variable technique (Lindell and Whitney, 2001; Malhotra et al., 2006). Although the absolute minimum inter-construct correlation in our matrix was −0.025, CMB can only inflate covariances, so a negative value cannot serve as a marker. We therefore used the smallest positive correlation, r = 0.036, as recommended. Subtracting this constant from every non-diagonal entry reduced the mean inter-construct correlation from 0.333 to 0.297 (an 11% drop) and only two marginal pairs fell to zero; none became negative, and all theoretically relevant links (> 0.20) remained virtually unchanged and statistically intact. Because the marker adjustment produced no substantive attenuation of the focal relationships, we conclude that common-method bias is not a threat to our findings.

We assessed the measurement model using standard procedures for PLS-SEM, including examination of indicator loadings, composite reliability (CR), Cronbach’s alpha (α) and average variance extracted (AVE). The factor loading of each variable was computed and found to be greater than 0.7, which is significantly higher than the criteria established in previous studies (Hair et al., 2017). For example, Hair et al. (2017) proposed a factor loading of 0.3 as practically meaningful when the sample size is at least 350. More strict requirements range between 0.32 (poor) and 0.7 (outstanding) (Comrey and Lee, 2013).

Both composite reliability (CR) and Cronbach’s alpha (α) are commonly used to assess the reliability of a measurement scale. We used CR to evaluate convergent validity, and all constructs exhibited CR values that exceeded the recommended criterion of 0.7 (Arshad and Khurram, 2020). CR is a more comprehensive measure of reliability that considers both the average variance extracted (AVE) and the intercorrelations among items. It is often preferred over Cronbach’s alpha, especially when the scale has a complex structure or when items are not perfectly correlated. Cronbach’s alpha values for all constructs exceed the recommended threshold of 0.7, indicating stronger internal consistency among construct variables. Alpha is a measure of internal consistency that assesses the extent to which the items on scale measure the same construct. It is widely used statistic, but it can be sensitive to the number of items and the intercorrelations among them. Furthermore, AVE values are greater than 0.50, which further supports the convergent validity of our constructs.

Table 3 summarizes the factor loadings, Cronbach’s alpha, CR and AVE values for each variable and construct. Our measurement approach is validated and reliable, as evidenced by high factor loadings and good α, CR and AVE values. These findings lend confidence to the measurement of the components and support the robustness of our data analyses.

Table 4 reports the discriminant validity of the constructs based on the Fornell–Larcker method. The diagonal entries represent the square roots of the AVE for each construct, while the off-diagonal entries show the correlations between constructs. According to the Fornell–Larcker criterion, a construct demonstrates discriminant validity if its square root of AVE is greater than its correlations with all other constructs. As shown in Table 4, for every construct, the diagonal value exceeds all corresponding off-diagonal correlations, indicating that each construct shares more variance with its own indicators than with those of other constructs. This confirms that the measurement model possesses adequate discriminant validity.

Following the fit and validity tests on the data, we present the results of the related structural model in the second stage to validate the construct relationships. Prior to analyzing the hypothesis outcomes, we present the model fit indices. The standardized root mean square residual (SRMR) was found to be 0.059, significantly lower than the cut-off value of 0.08. Similarly, the variable inflation factor (VIF) for each predictor was found to be less than 5.0, suggesting the absence of multicollinearity.

For hypothesis testing, we conducted a statistical examination of the underlying structural model using the path coefficient value and significance level. For significance, we set the threshold p-value at less than 5%. This criterion allows us to confidently assess if the correlations between the constructs are statistically significant. To ensure that the significance of the results is not attributable to chance, we used a bootstrap technique using 5,000 subsamples. This bootstrap methodology reduces the possible influence of outliers, yielding more reliable results. Table 5 summarizes the hypothesized correlations and their levels of significance including the computed path coefficients with respective p-values. The p-value determines if the relationships between the given constructs are statistically significant.

Figure 4 highlights that there is a significant positive relationship between internet trust and intention to use e-gov services, supporting our hypothesis H1 (β = 0.213, p-value = 0.000). The hypothesis that trust in government will have a positive influence on the intention to use the e-gov services is not supported, therefore, our hypothesis H2 is rejected (β = 0.005, p-value = 0.885). Hypotheses H3 and H4 are supported, indicating that disposition to trust has a positive influence on trust of the internet (H3, β = 0.892, p-value = 0.000) as well as on trust in government (H4, β = 0.223, p-value = 0.000). Hypothesis H5, which states that perceived risk has a negative impact on citizens’ use of e-gov services, is also supported (β = −0.083, p-value = 0.003). The hypothesis that a higher level of trust of the internet will lead to a lower perceived risk is supported (H6, β = 0.437, p-value = 0.000), but the hypothesis that a higher level of trust in government will reduce the perceived risk of using e-gov services is not supported (H7, β = 0.073, p-value = 0.064). Hypothesis H8 – perceived usefulness will positively influence citizen use of e-gov services is supported (β = 0.485, p-value = 0.000). Hypotheses H9 and H10 which theorizing that perceived ease of use will positively influence citizen use of e-gov services and perceived usefulness are supported (H9, β = 0.108, p-value = 0.009), (H10, β = 0.585, p-value = 0.000). Finally, hypothesis H11 is supported, which states that confirmation has a positive influence on the use of e-gov services (β = 0.162, p-value = 0.000).

To measure the predictive accuracy of the underlying structural model, we calculated R² values. R² values reflect the variance in the dependent variable that is explained by the independent variables that influence it. The literature provides variance cut-off values for R² values to be significant and meaningful. For example, Cohen (2013) suggested 0.26 as substantial, 0.13 as moderate and 0.02 as weak. Falk and Miller (1992) suggested an R² value of at least 0.1 to be adequate. Chin (1998) suggested 0.67 (substantial), 0.33 (moderate) and 0.19 (weak). Recent literature (Hair et al., 2011, 2013) suggested different R² values including 0.75, 0.5 and 0.25 to be substantial, moderate and weak respectively. Our analysis of the R² values shows that underlying structural model can explain 59% of variance in continuance intention, 34.2% of the variance in satisfaction, 79.6% of trust in the internet, 20.9% of perceived risk and 5% of trust on government.

This study investigates citizens’ perceptions of e-gov services in a developing country, focus on trust and how this affects their intention to continue using such services. The research is motivated by the gap in existing literature, which primarily focuses on the adoption of e-gov services in Pakistan, rather than citizens’ continuous use intention. To address this issue, we integrate the ECM (Bhattacherjee, 2001) with the trust model adopted from Carter et al. (2016) to provide a theoretical understanding of the mechanisms that explain this phenomenon.

Our findings reveals that the continuation of e-gov services is influenced by the confirmation of expectations, perceived usefulness and satisfaction. Citizens are more inclined to adopt and use e-gov services when they perceive them as valuable and useful. This underscores the importance of raising public awareness about the benefits of these services to foster their widespread adoption. By using e-gov platforms, citizens can save time and resources by completing government-related tasks from the convenience of their homes. In contrast, traditional in-person interactions with government offices often involve bureaucratic complexities and time-consuming processes. Moreover, the established correlation between perceived usefulness and behavioral intentions underscores the importance of designing e-gov platforms to be user-friendly and accessible. When citizens are aware that an electronic option is available and find it useful, they are more likely to adopt this innovation (Carter et al., 2016; Ahmad et al., 2016).

Continued use of e-gov services requires trust in institutions or trust in the internet. Citizens need to trust that procedures are in place to protect their personal data. People’s perceptions of the safety and security of the internet are important factors in determining whether or not e-gov services are adopted. Governments need to implement trust-building strategies to increase citizens’ confidence, as TOI influences the intention to continue using e-gov. TOG, on the other hand, was not a significant predictor of USE. Despite citizens’ dissatisfaction with government, Cook (2007) claims that people are satisfied with the online services provided by government.

While perceived risk plays a significant role in the adoption of e-gov services, it is crucial to explore specific concerns related to data security and privacy, especially in Pakistan. Trust deficit in the country is exacerbated by incidents such as the hacking of financial institutions and the sale of personal data on the dark web, which undermine citizens’ confidence in digital platforms. Citizens’ concerns about how their personal and financial data are handled by government platforms represent a substantial barrier to continued usage. E-gov services in Pakistan must ensure the highest levels of data protection, using encryption, regular audits and transparent communication regarding how sensitive information is processed and stored. Addressing these security and privacy concerns can directly mitigate perceived risk and enhance trust in e-gov services. In addition to trust-related factors, broader socio-economic and infrastructural issues significantly influence the continued use of e-gov services in Pakistan. For example, the digital divide in Pakistan, marked by varying levels of internet access and digital literacy, directly affects citizens’ ability to benefit from e-gov services. Many rural areas have limited access to the internet, and even within urban populations, internet literacy remains a challenge. Lack of technical proficiency in navigating online services can lead to frustration and abandonment, particularly among older adults and lower-income groups. Policymakers need to address these infrastructural barriers by investing in nationwide digital literacy programs and ensuring equitable access to affordable technology and high-speed internet, particularly in underdeveloped regions. Furthermore, the R-squared value of 5% for trust in government indicates that other, unmeasured factors may influence this relationship. These results suggest that future research should investigate additional socio-political factors, such as perceptions of governmental accountability or media influence, which may impact trust in e-gov. Another way of understanding the insignificance of this construct may be related to the legal, economic, social and political environment. TOG is insignificant in developed countries where public sector corruption is low (Carter et al., 2016). Surprisingly, in Pakistan, a developing country where corruption permeates almost every aspect of daily life, TOG was reported to be insignificant in the use of e-gov services. A study conducted by Khan et al. (2021) reported that a higher level of trust among citizens is associated with the intention to use social media for e-gov services in Pakistan. Trust is a personal inclination that cannot be influenced by government agencies. There is a positive relationship between satisfaction and citizens’ intention to continue using e-gov services. This relationship is also supported by other studies (Bhattacherjee, 2001; Veeramootoo et al., 2018; Venkatesh et al., 2011).

Although PR is an important factor in the uptake of e-gov services, it is vital to examine the specific concerns surrounding data security and privacy, particularly in the Pakistani context. The lack of trust in the country is further compounded by incidents such as the hacking of financial institutions and the sale of personal data on the dark web, which serve to erode citizens’ confidence in digital platforms. The concerns of citizens regarding the handling of their personal and financial data by government platforms constitute a significant obstacle to continued usage. It is imperative that e-gov systems in Pakistan guarantee the utmost levels of data protection, using robust encryption, periodic audits and transparent communication regarding the processing and storage of sensitive information. The mitigation of these security and privacy concerns can directly reduce the perceived risk and enhance trust in e-gov services. In addition to factors related to trust, broader socio-economic and infrastructural issues exert a significant influence on the continued use of e-gov services in Pakistan. For instance, the digital divide in Pakistan, characterized by disparate levels of internet access and digital literacy, directly impacts citizens’ capacity to reap the benefits of e-gov services. A significant proportion of the population in rural areas has limited access to the internet, and even within urban populations, there is a lack of digital literacy. This can lead to frustration and abandonment of online services, particularly among older adults and those in lower-income groups. It is therefore essential that policymakers address these infrastructural barriers by investing in nationwide digital literacy programs and ensuring equitable access to affordable technology and high-speed internet, particularly in underdeveloped regions.

In this study, the association between the constructs of satisfaction and continuance usage intention is revalidated in the context of e-gov services. Overall, this study suggests that trust is a critical factor in the continued use of e-gov services and that government needs to take proactive measures to build trust among citizens. E-gov service providers should also be aware of citizens’ different levels of trust and tailor their services accordingly to increase public confidence in government and the internet.

This study has several implications for subsequent research. It makes significant contribution to theoretical understanding by integrating Carter et al. (2016) trust model with Bhattacherjee (2001) expectation-confirmation and continuance model. By integrating these two theoretical frameworks, our model provides a more comprehensive understanding of the intricate relationships among the key factors influencing e-gov continuance use intention and trust. This research extends our understanding of online trust in the e-gov context by hypothesizing the differential effects of “trust in government” and “trust in technology.” Our study is among the first to expand Carter et al. (2016) model by integrating trust as an overall antecedent of continuous use of e-gov services. This integrated model allows us to comprehensively assess citizens’ perspectives on e-gov services in Pakistan. Notably, our integrated model accounts for 60% of the variance in citizens’ intention to use e-gov services, thus substantially contributing to the existing literature. This enhanced perspective offers valuable insights into the decision-making processes of individuals when considering the use of e-gov services. The broader scope of the model not only advances to the existing literature but also provides practical guidance for practitioners seeking to improve the effectiveness of e-gov initiatives.

Our results show that citizens’ trusting beliefs are an overall antecedent of continuous e-gov services use. This finding aligns with the notion that trusting beliefs influence perceived usefulness and perceived ease of use in the e-commerce context (Gefen et al., 2003). Both findings suggest that when users receive e-services, their attitudes and perceptions toward these services are influenced by trust. As government agencies provide an increasing range of information and e-services, it is crucial to recognize that investing in building citizens’ trust is essential for the use of all these services. However, further research is needed on how trust affects different types of e-gov services.

Pakistan, as a developing country, has made significant efforts to provide various e-gov services to its citizens. However, most of these efforts requires further enhancement to ensure effective service delivery and facilitate a smooth adoption process. A key contribution of this research is its comprehensive examination of citizen trust in the context of e-gov services, which, to the best of our knowledge, has not been extensively studied in Pakistan. Furthermore, our study introduces and tests a novel research model in this context.

To ensure the robustness of our findings, we collected data from a diverse pool of citizens representing different age groups, levels of education and employment status. This diverse sampling approach provides us with more comprehensive insight into citizens’ perceptions of e-gov services compared to studies conducted only among students or smaller groups. Our results show that the factors TOI, TOG, DI and confirmation significantly predict the intention to continue using e-gov services. Notably, our findings are consistent with previous studies on trust in e-gov services conducted in both developed and developing countries (Carter et al., 2016). However, in the context of Pakistan, the research stream on e-gov has largely ignored the trust factor and service providers have not adequately considered it (Carter et al., 2016; Javaid and Arfeen, 2017; Rehman et al., 2012a, 2016).

In our sample, we ensured that each citizen had already used available e-gov services. We deliberately did not focus on a specific service, as our aim was to capture citizens’ general perceptions and experiences of e-gov services in Pakistan. This approach provides us with a more holistic view of the current state of adoption of e-gov services and the key factors influencing the ongoing usage process. This comprehensive perspective enhances the generalizability of our findings across various e-gov services in Pakistan.

Our research has implications for government agencies that provide, or plan to provide, online services. A primary challenges these agencies face is motivating citizens to continue using their e-gov services. Our findings highlight the critical role of trust in encouraging users’ sustained engagement with digital government platforms. Furthermore, different IT attributes of e-gov services play different roles in users’ intention to continue using them. In particular, trust-building mechanisms—both online and offline—are vital for the success of any e-gov initiative. Given that users’ perceptions of e-gov services are largely shaped by their trust beliefs, governments must implement effective strategies to foster trust.

This research offers practical insight that can help improve e-gov services to satisfy citizens and gain their trust. Factors like convenient access, perceived usefulness and tangible benefits are essential in influencing citizens’ continued use of these services. In developing countries such as Pakistan, local governments play a significant role in conveying the advantages of e-gov services to the public. A major obstacle to adoption in these contexts is a lack of awareness, as previously identified (Ahmad et al., 2013). This challenge align with findings from developed countries, where citizens appreciate the relative benefits of e-gov services after becoming aware of them (Carter et al., 2016).

To effectively communicate the benefits of e-gov services, service providers should emphasize the comparative advantages over traditional methods—such as faster, cheaper and more convenient service delivery. Equally crucial is ensuring transparency and trustworthiness in e-gov platforms is equally critical. Service providers must be transparent and proactive in explaining how these services function and how citizens’ data is securely managed. To raise awareness and build confidence, governments can implement marketing campaigns, engage local authorities and use social media. These efforts can help address concerns regarding the perceived impersonal nature of digital services, as well as potential security and privacy issues.

Trust emerges as a key driver of citizens’ intentions to continue using e-gov services. It is paramount to ensure platforms are secure, privacy concerns are taken seriously, and reliability is signaled through well-designed services is paramount. E-gov providers must be attuned to the nature of trust and its influence on public confidence in both government and the internet. While some citizens may naturally be more hesitant and cautious about using e-gov services, it is incumbent upon providers should engage directly with this segment to address their concerns. Positive experiences with e-gov services can foster trust and confidence, leading to further engagement.

Active engagement and responsiveness by e-gov providers are crucial indicators of a safe and trusted environment. Our study offers practical guidance for improving e-gov services to gain citizens’ trust and increase adoption. By focusing on convenience, perceived usefulness and effective communication of the benefits, as well as addressing citizens’ security and privacy concerns, e-gov services can further enhance trust. Prioritizing transparency, security and reliability in both service design and communication will contribute to building a trustworthy e-gov platform. Furthermore, ongoing engagement with citizens is essential to fostering a secure and reliable digital government ecosystem. To enhance user experience and trust, we recommend the following strategies:

• implement robust security measures and clearly communicate them to users;

• regularly update and improve e-gov platforms based on user feedback;

• provide clear, accessible information on how personal data is handled and protected;

• offer multiple channels for user support, including online chat and helplines;

• conduct regular audits to ensure service quality and identify areas for improvement;

• collaborate with local community leaders to promote e-gov services and address concerns; and

• develop user-friendly interfaces that cater to various levels of digital literacy.

By implementing these recommendations, government agencies can create more trustworthy, efficient and user-centric e-gov services, ultimately leading to increased adoption and continued use among citizens.

This study provides valuable insights into e-gov services in Pakistan; however, several methodological and theoretical limitations must be acknowledged. These limitations not only contextualize our findings but also offer avenues for future research. Our study faces several methodological constraints that should be considered when interpreting the results. First, the cross-sectional nature of our data collection limits our ability to establish causal relationships and observe temporal changes in e-gov adoption patterns. This limitation is particularly significant given the dynamic nature of technologies adoption and trust development. Our research focused on existing e-gov services and infrastructure, but the continuous emergence of new technologies means that some of our findings may require reexamination in light of technological advancements. A critical limitation concerns our measurement model. Specifically, one of our key constructs (i.e. PR) was measured using only two items, which falls below the recommended minimum of three items needed for statistical identification at the construct level. This limitation may affect the robustness of our findings related to this particular construct and should be addressed in future research.

The geographic scope of our data collection represents another significant limitation. Our study context-specific limitations stem from Pakistan’s unique socio-cultural landscape. While Pakistan comprises four provinces with distinct cultural characteristics, our research design may not fully capture this cultural heterogeneity. The complexity of linguistic diversity and its potential impact on e-gov service adoption remains unexplored in our current study. Our sample was restricted to three major urban centers in Pakistan – Peshawar, Lahore and Islamabad – potentially limiting the generalizability of our findings to other regions and rural areas. This urban-centric sampling approach may not adequately capture the diverse experiences and challenges faced by citizens in different parts of the country.

From an analytical perspective, our study focus on general e-gov services, rather than specific service types, may mask important variations in adoption patterns and trust development across different e-gov applications. Additionally, our analysis does not fully account for the digital divide prevalent in Pakistan, which could significantly influence e-gov adoption patterns and user experiences.

These limitations suggest several promising directions for future research. Longitudinal studies could provide deeper insights into the temporal dynamics of e-gov adoption and trust development. Such studies could track changes in citizen attitudes and behaviors as they gain experience with e-gov services, offering valuable insights into the evolution of trust and adoption patterns. Comparative studies between developing and developed countries could help identify universal factors and context-specific influences on e-gov adoption. Such research could particularly benefit from examining how different economic, legal, political and social contexts shape e-gov implementation and adoption.

Future research needs to investigate technological aspects of e-gov services. This includes investigating the impact of emerging technologies such as artificial intelligence and blockchain on service delivery and citizen trust. Research on user experience and interface design could provide practical insights for improving e-gov accessibility and usability. The role of trust in e-participation and civic engagement represents another important area for future investigation. Research could explore how trust influences citizens’ engagement in online democratic processes and examine the relationship between e-gov service use and broader civic participation. Finally, given increasing concerns about data privacy, future studies should investigate citizens’ perceptions of data protection in e-gov services and examine how privacy concerns influence trust and adoption patterns. This research could help develop best practices for balancing data utilization and privacy protection in e-gov implementations.

While PLS-SEM was chosen for its suitability in handling complex models and its predictive focus, we acknowledge its limitations, particularly in the context of theory testing. Unlike CB-SEM, PLS-SEM does not provide global goodness-of-fit statistics, such as chi-square or RMSEA, which are essential for assessing the overall fit of the theoretical model to the data (Hair et al., 2017). This limitation means that while we can evaluate the local fit of individual paths and constructs, we cannot comprehensively assess the global fit of the entire model. Moreover, for reflective measurement models, PLS-SEM may produce biased parameter estimates compared to CB-SEM, potentially affecting the accuracy of our findings (Sarstedt et al., 2016). Given that PLS-SEM is primarily a prediction-oriented method, its use for strict theory confirmation may be less robust than CB-SEM, which is specifically designed for confirmatory research (Hair et al., 2017). These limitations have implications for the interpretation and generalizability of our findings. Without global fit indices, the overall validity of our theoretical model is somewhat constrained. Additionally, the potential bias in parameter estimates for reflective constructs suggests that our results should be interpreted with caution. To address these limitations, future research could use CB-SEM to test the same model, particularly if larger sample sizes and normally distributed data are available. A robustness check using CB-SEM would provide a more comprehensive assessment of the model’s fit and the relationships among the constructs, thereby enhancing the confidence in our theoretical contributions.

This study has explored the key factors influencing Pakistani citizens’ intentions to continue using e-gov services. Our findings underscore the pivotal role of e-gov in improving the efficiency of public service delivery. The research identified several factors as crucial in understanding citizens’ decisions to use e-gov services in the context of developing countries, specifically Pakistan: disposition to trust, trust in the internet, perceived risk, satisfaction and perceived usefulness. These factors, particularly within the context of developing countries, significantly impact the success of e-gov initiatives. As digital technologies become more accessible and affordable, they offer governments unprecedented opportunities to expand their e-gov offerings. However, the long-term success of these initiatives hinges on citizens’ willingness to not only adopt but also consistently engage with these services. By fostering greater transparency, e-gov services can play a crucial role in reducing corruption and increasing public trust in governmental institutions.

Despite considerable research on e-gov adoption in developed and developing countries, the specific influences of trust and risk in emerging economies have remain underexplored. This study contributes to bridging that gap and lays a foundation for future research. Our findings reveal that trust is a fundamental factor in citizens continued use of e-gov services; perceived usefulness and satisfaction significantly influence the intention to use e-gov services; the interplay between trust in government and trust in technology is critical for e-gov success; cultural and contextual factors unique to developing countries like Pakistan play a vital role in shaping e-gov adoption. Furthermore, our research highlights the need for tailored strategies to build and maintain citizen trust in e-gov initiatives; continuous improvement of e-gov services based on user feedback and changing needs; addressing the digital divide to ensure equitable access to e-gov services; balancing technological advancements with user-centric design and privacy protection.

Addressing the limitations of this study and exploring the proposed research directions will help further our understanding of e-go adoption, enabling the development of more effective and widely embraced e-gov services in Pakistan and other developing nations. As governments worldwide continue to digitize their services, the insights from this study can inform policy-making and implementation strategies, ultimately leading to more efficient, transparent and citizen-centric governance. In conclusion, while e-gov offers immense potential for improving public service delivery and citizen engagement, its success is contingent upon understanding and addressing the unique challenges and opportunities presented in different contexts. This study provides a solid foundation for future research and practical initiatives in this critical area of public administration and digital governance.

Trust of the government (TOG): Composed of the traditional view of trust in a specific entity (Carter et al., 2016):

• TOG1. The e-gov services of Pakistan can be trusted to carry out online transactions faithfully.

• TOG2. In my opinion, the e-gov services of Pakistan are trustworthy.

• TOG3. I trust the e-gov services of Pakistan to keep my best interests in mind.

Trust of the internet (TOI): Trust in the reliability of the enabling technology (Carter et al., 2016):

• TOI1. The internet has enough safeguards to make me feel comfortable using it to interact online with e-gov services of Pakistan.

• TOI2. I feel assured that legal and technological structures adequately protect me from problems on the internet.

• TOI3. In general, the internet is now a robust and safe environment in which to transact with the e-gov services of Pakistan.

Perceived usefulness (PU): Belief that the service enhances task efficiency (Bhattacherjee, 2001):

• PU1. The e-gov services of Pakistan web sites would enable me to complete transactions more quickly.

• PU2. I think the e-gov services of Pakistan web sites would provide a valuable service for me.

• PU3. I would find the e-gov services of Pakistan web sites useful.

• PU4. Using e-gov services of Pakistan would make it easier to do my transactions.

Perceived risk (PR): Risk perception in terms of the trustor’s belief about the likelihood of gains and losses (Carter et al., 2016):

• PR1. The decision of whether to use a state e-government service is risky.

• PR2. In general, I believe using state government services over the internet is risky.

Disposition to trust (DT): Defined as one’s general propensity to trust others. It is composed of two concepts: faith in humanity and trusting stance. (Carter et al., 2016):

• DT1. I generally do not trust other people.

• DT2. I generally have faith in humanity.

• DT3. I feel that people are generally reliable.

Continuance intention to use:

• USE1. I would use the web for gathering information from the e-gov services of Pakistan.

• USE2. I would use e-gov services of Pakistan provided over the web.

• USE3. I would use the web to inquire about e-gov services of Pakistan.

Satisfaction: Affective evaluation of prior service experiences (Bhattacherjee, 2001):

• How do you feel about your overall experience of e-gov services of Pakistan?

• S1. Very dissatisfied/Very satisfied.

• S2. Very displeased/Very pleased.

• S3. Very frustrated/Very content.

• S4. Absolutely terrible/Absolutely delight.

Confirmation: The extent to which users’ expectations align with actual service performance (Bhattacherjee, 2001):

• My experience with using e-gov services of Pakistan was better than what I expected.

• The service level provided by e-gov services of Pakistan was better than what I expected.

• Overall, most of my expectations from using e-gov services of Pakistan were confirmed.

• Attention checker question.

• If you are paying attention, please respond agree to this question.

Note(s): We used seven-point Likert scale