eGovernment began about three decades ago to improve service delivery by making it practical and efficient (Ebbers and van de Wijngaert, 2020). According to Moon (2002), eGovernment includes five interaction levels: information provision, communicating with citizens, online transactions, integration of government agencies, and citizen participation. The basic idea of eGovernment was to replace time-consuming and expensive traditional front-desk channels, such as queues in offices, extensive telephone use, and intensive form-filling (Goodwin, 2010), with quick, accurate, and consistent online procedures (Pieterson et al., 2007). These changes would achieve benefits such as cost-efficient due to the diminished need for transportation to offices and reduced stationery expenses (Venkatesh et al., 2012). Hence, government agencies globally have invested in implementing and enhancing eGovernment systems (Alshehri and Drew, 2010). Although some citizens use eGovernment, most still prefer to interact with the government through traditional means (telephone or front desk), especially in developing countries (Madsen et al., 2020). However, numerous government agencies make some services accessible only online to encourage users to utilize eGovernment. In developed countries, such as Denmark and The Netherlands in Europe, many countries have made one or more services accessible through e-services only (Ebbers and van de Wijngaert, 2020), and citizens seem to have accepted it well. However, regarding developing countries, the same cannot be said. Research reports that overall, despite government efforts, eGovernment still creates significant utilization and adoption challenges (Papadopoulou et al., 2010; Al-Hujran et al., 2015).

eGovernment efficiency refers to improved public services, information transparency, and effective public participation with public sector management online, creating significant cost-savings for both citizens and government operations (Hackney et al., 2007). eGovernment efficiency mainly focuses on efficient service delivery that satisfies most users’ needs (Roman and Miller, 2015). Similar to Moon (2002)’s five eGovernment interaction levels, Holzer and Kim (2006) constructed a model to measure eGovernment efficiency consisting of five categories: privacy and security; usability; content; services; and citizen and social engagement. In the context of eGovernment studies, many researchers have explored eGovernment based on these efficiency measures, for example, usability (Gao and Lee, 2017); privacy and security (Janita and Miranda, 2018); content (Mokone et al., 2018); services (Tan et al., 2013); and citizen participation (Lee and Porumbescu, 2019). In the present study, eGovernment efficiency is considered a non-human environment situation. A non-human environment refers to situations or conditions likely to cause psychological and behavioral effects among individuals (Searles, 1960).

According to Chen et al. (2013), psychological consequences significantly influence individuals’ wellbeing. SWB is essential to overall wellbeing because it comprises cognitive elements of life satisfaction (Diener et al., 1998). It plays a vital role in the positive development of an individual as an indicator, a moderator, a mediator, a predictor, or an outcome (Park, 2004). The theory of core affect by Russell (2003) states that various internal and external factors influence reflexes, perception, cognition, and behavior. Still, people do not have direct access to these causal connections. As a result, core affect can be experienced as a free-floating (mood) state or attributed to a specific cause and, thereby, an emotional episode. SWB is a psychology field that attempts to examine and understand how people evaluate their lives regarding positive and negative emotions (Peterson et al., 2005; Diener, 2009). Negative emotions are related to social, psychological, and behavior problems (Peterson et al., 2005), while positive emotions are related to good adaptation and optimum mental health (Proctor et al., 2010). SWB has three components: positive affect (PA); life satisfaction (LS); and negative affect (NA) (Andrews and Crandall, 1976). Individuals have positive SWB if they experience LS and frequent PA, such as happiness and optimism, and infrequent NA, such as rage, annoyance, and sadness. Feelings associated with PA and LS empower people to overcome complications positively and accomplish what they want in life (Diener and Chan, 2011). NA feelings negatively influence individuals’ wellbeing, creating extra pressure and anxiety both on an individual’s body and mind, which might lead to health issues if the strain becomes long-lasting or overwhelming (Xu and Roberts, 2010). Lawless and Lucas (2011) argued that, wherever there is a positive influence on life satisfaction, there is subsequently greater life expectancy, lower levels of homicides, as well as lower rates of liver and heart-related diseases. Governments control essential services citizens need to generate and secure their economic situations and meet their basic needs. In this context, Diener (2016) advocated that an individual’s levels of SWB are influenced both by internal factors, such as behavior and viewpoint, and external factors, such as the environment in which they live, and, most importantly, their ability to meet their basic needs. Therefore, we argue in this paper that eGovernment experience has an essential effect on citizens’ wellbeing because every situation that affects an individual affects their wellbeing (Diener et al., 1985).

SOBC theory states that different elements of the environmental situation (S) affect the internal states of organisms (O), which influences specific behavioral responses (B), triggering consequences (C) as a result (Davis and Luthans, 1980). According to Whelan et al. (2020), the SOBC model is a more multifaceted mechanism of human behavior that modifies and extends the antecedent stimulus–organism–response (SOR) (Mehrabian and Russell, 1974) and antecedent–behavior–consequence (ABC) (Skinner, 1963) models. These authors further assert the importance of identifying the effects of covert stimuli, such as thoughts, expectations, and dreams, as the crucial mediating role for O and B between S and C. Similarly, recognizing utilization behavior is essential in adjusting the relationship between the environment and people. Figure 1 depicts the SOBC model (Davis and Luthans, 1980).

This study employed the SOBC model for several reasons. First, it is suitable for studying the influence of technology use on citizens’ emotions and behavior because it provides a robust theoretical approach for developing and validating an appropriate research model focusing on people and their environment (Whelan et al., 2020). Second, the SOBC model provides a structured procedure that specifically assists in examining how individuals’ internal state of mind mediates the relationship with their environment, which subsequently impacts their SWB. According to Van Hoorn (2007), SWB is a self-report technique for measuring wellbeing based on physical surroundings and social situations. Tuncer et al. (2005) also asserted that these situations could influence people’s behavior. Cooper et al. (2007) stated that behavior takes place in the form of a process. Therefore, firstly, environments affect people’s psychology, then they react, affecting their wellbeing. Research has highlighted trust as the primary mediator between technology (an online environment) and its utilization (Colesca, 2009). Similarly, the SOBC model uses the environment to determine how it affects people; thus, our study expects trust to mediate the relationship between eGovernment efficiency and utilization. eGovernment studies primarily implemented technology acceptance models (TAMs), such as the unified theory of acceptance and use of technology (UTAUT) (Venkatesh et al., 2011) and TAM (Lin et al., 2011) to explore and explain users’ intentions to utilize e-services and subsequent usage behavior. Thus, by using the SOBC model, this study extends and diversifies the theoretical and practical perspectives in the study of eGovernment dimensions.

This section describes how this study selected and aligned variables with the SOBC theory. We combined prior research insights with the SOBC theory to develop this study’s research model. Malodia et al. (2021) conducted interviews with government officials, citizens (urban and rural), technology experts, senior policy-makers, government agencies, and academic researchers from November 2016 to January 2019. With the evolution of eGovernment literature from 2002 to 2019, the author reported that efficiency, trust, citizen satisfaction, infrastructure, culture context, citizen readiness, and citizen empowerment contribute to the successful implementation of eGovernment. These factors align with the five Moon (2002)’s eGovernment interaction levels and Holzer and Kim (2006)’s eGovernment efficiency factors.

Consequently, this indicates that eGovernment efficiency sums up almost all aspects contributing to eGovernment success. Therefore, this research considers efficiency a robust environment antecedent influencing the relationship between eGovernment and citizens. Similarly, prior research has heavily pointed to trust in different forms as major cognitive processes representing the person variable in regards to e-services utilization and adoption (Colesca, 2009; Schierz et al., 2010; Cheng et al., 2017a,b). Since our main objective is how this affects SWB, wellbeing is fundamental to health and happiness. Having a well-adapted and robust sense of wellbeing can help individuals overcome difficulties and help them achieve their goals in life (Diener and Chan, 2011), and reduce levels of homicides, liver, and heart-related diseases (Lawless and Lucas, 2011). Therefore, this study identified eGovernment efficiency, trust, and utilization as significant factors contributing to SWB. Figure 2 represents our research model, in which we propose the relationships between the covert stimuli of eGovernment efficiency, trust, utilization, and SWB. In our model, eGovernment efficiency represents the situation (S), trust represents the organism (O), utilization represents the behavior (B), and SWB represents the consequence (C). Users measured eGovernment efficiency based on their expectations regarding the five efficiency categories: privacy and security; usability; content; services; and citizen and social engagement (Holzer and Kim, 2006). Within the psychology field, an expectation is a strong belief that something will occur or be the case (Driskell and Mullen, 1990; Martin et al., 2001). Citizens’ expectations and level of satisfaction influence eGovernment use; when these are high, their trust and intention to use eGovernment services also increase (Khan et al., 2020; Rosenberg, 2020). Finally, SWB, this study’s dependent variable, refers to the psychology field that attempts to examine and understand how people evaluate their lives based on life satisfaction, joy, and thinking and feeling that life is going in the right direction (Diener, 2009).

Out of 500 questionnaires shared, a total of 300 respondents completed the survey from the share links and used for the analysis, a response rate was 60%. As the questionnaire fields were validated and there were no missing data, all responses received were usable. Approximately 73% of respondents had tertiary-level education, 19% high-school, 6% junior-school, 2% primary-/elementary-school, and none fell under the never attended school category. Regarding employment, approximately 18% of respondents were government workers, 26% were private-sector employees, 31% were university students, and 25% were unemployed. About 43% of the respondents were female, and 57% were male. The minimum respondents’ age was 25–34 years, and the maximum was 45–54 years. Almost 78% of respondents accessed the internet daily, 19% twice a week, 1% fortnightly, and 2% once a month. Supplementary Appendix 1 shows the respondents’ demographic attributes.

This study’s descriptive statistics results for the 18 items for the four constructs are shown in Supplementary Appendix 2, encompassing the minimum value, range value, maximum value, mean value, standard deviation, skewness, and kurtosis. According to George (2011), when proving a normal univariate distribution, values for asymmetry and kurtosis between –2 and + 2 are considered acceptable. Similarly, Bryman (2012) argued that data the skewness values between –2 and + 2, and kurtosis values between –7 and + 7 are considered normal. This study skewness and kurtosis values follow the normal distribution, and the results are detailed below.

Respondents believe that eGovernment efficiency is characterized by websites or systems that are easy to use, with a clear homepage, sufficient and relevant information, as well as a search option to aid in quick access to the desired output. According to Immordino and Russo (2018), users assess e-services using quality aspects such as the service quality, system quality, and information quality of any e-service; these aspects are crucial concerning e-services’ success. Similarly, prior studies highlighted up-to-date information, usefulness, ease of use, web interface, and search engine as attributes that users consider in assessing competence (El-Kassem et al., 2020; Thu et al., 2020).

Regarding trust, the findings reveal that respondents have confidence in government agencies that take full responsibility for any type of security breach during interactions/transactions on the website and do their best to help users if they have a query. Systems or websites are considered more reliable and convenient than physical offices. Prior research has also indicated that reliability, validity, security, and privacy are essential antecedents of eGovernment trust (Thu et al., 2020; Zhu et al., 2020). Users’ covert or overt attributes, such as beliefs and attitudes related to agencies having the best interests of society and its constituents at heart, significantly impact trust and intention to use e-services (Alryalat et al., 2020; Pérez-Morote et al., 2020).

About utilization, respondents believe that interacting with government websites to receive government services enhances an individual’s or an organization’s social status. In terms of the skills needed to interact with online systems, respondents acquire them from computer technology at their workplace, institution, or home. Respondents also believe that websites provide a wider choice of interactions with different functions than interactions within a physical government office. Prior research has shown that personal outcome expectations act as cognitive abilities influencing users to utilize eGovernment services (Alruwaie et al., 2020). Additionally, Mensah et al. (2020) argued that obtaining the desired benefits positively influences intentions to use eGovernment services.

Finally, concerning SWB, respondents believe that, in most ways, online government services are close to ideal and only if their experience with online governance is excellent. Therefore, online service delivery offers essential services needed in daily lives, and only a few critical things need to be changed in existing platforms to perfect the efficiency. Similarly, prior studies have shown that eudaimonia and life satisfaction are cognitive evaluations of SWB, while happiness and anxiety are effective evaluations (Mouratidis, 2020, 2019). Happiness, anxiety, and satisfaction fall within the three components of SWB evaluation (PA, LS, and NA). Therefore, integrating both cognitive and affective measures of SWB to satisfaction with life domains yields essential insights.

Reliability is a method of assessing the consistency of a measure. This study evaluated the multiple Likert-scale items’ internal consistency using Cronbach’s α coefficient. The accepted Cronbach’s α value is 0.70 and above (Peters, 2014). The reliability analysis results are shown in Table 1. The Cronbach’s α value for each type of factor and the total sample is above 0.7, signifying a strong consistency among the items.

In this study, Bartlett’s test of sphericity was used for the validity analysis, which evaluates whether the data matrix is an identity matrix. The KMO test was used to measure the sample adequacy for each variable in the model and the complete model. If Bartlett’s test of sphericity’s value is less than 0.5, this means that the data has good construct validity. Similarly, if the KMO value is at least 0.50, the sample is adequate, with KMO values closer to 1 representing greater adequacy (Antony and Visweswara Rao, 2007). The results are shown in Table 2. The KMO statistics values are greater than 0.7. The F value of Bartlett’s spherical test equals 0.000, which means that the sample data is adequate and has good construct validity, hence being suitable for factor analysis. Following Jordan and Troth (2020), this study evaluated CMB using Harman’s single factor in SPSS Statistics 23 (33%) and the common latent variable test in Amos 23.00 (30%). The results are shown in Table 3. The tests variance is less than 50% in both tests, suggesting that CMB does not affect this study (Meade et al., 2007).

Factor analysis aims to reduce the number of measurement variables and reduce multicollinearity among the main variables for further analysis through model regressions. Principal component analysis was used to determine the factor loadings using direct oblimin rotation, with the criteria that the extraction eigenvalue should be greater than 1 and the factorial loadings should be 0.5 or above (Hair et al., 1998). The factor analysis results are shown in Table 4. One item from eGovernment efficiency (Q5) failed to meet the minimum cutoff criteria and was dropped, reducing the number of variables from 18 to 17.

SEM was conducted using Amos 23 software based on the factor analysis results. The maximum likelihood estimation method was utilized for the confirmatory factor analysis in a linear equation with the dependent variable Y having three dependent variables. The structural equations causal model is shown in Figure 3. This study established the following model fit criteria: χ²/df < 3; RMSEA < 3; AGFI, GFI, NFI, IFI, CFI, and RFI > 0.90; and PNFI > 0.05 (Little, 2014). The thresholds of the model fit are as follows: χ²/df = 1.67; RMSEA = 0.047; AGFI = 0.911; GFI = 0.935; NFI = 0.900; IFI = 0.952; CFI = 0.979; RNF = 0.935; and PNFI = 0.720. Thus, the model fit criteria are met.

The parameter estimation results (Supplementary Appendix 3) show correlations between the variables. Correlations that are ≤ 0.35 are considered weak correlations, 0.36–0.67 moderate correlations, and 0.68–1.0 strong correlations, with r coefficients > 0.90 very strong correlations (Taylor, 1990). The results are as follows.

The correlation between eGovernment efficiency and trust is positive and strong, supporting H1: r(298) = 0.83, p < 0.001. Respondents develop confidence in public online services depending on their effectiveness and efficiency. Easy to use systems, with a simple interface and without information overload on the home page, make a quick orientation possible. Systems with sufficient information and an effective search engine assist in quick access to the right information. Such systems make respondents feel confident in relying on government agencies to do their best to help users feel safe and secure when accessing services online. Similarly, using TAM, Pérez-Morote et al. (2020) conducted a multiple regression analysis where the variable “trust in government” obtained a significant beta coefficient (β = 0.314, p < 0.00), indicating a positive relationship between trust in government and citizens’ use of eGovernment. Using the mobile-government adoption model (MGAM), a model integrating the UTAUT model and the e-government adoption model (GAM) (Almaiah et al., 2020), conducted SEM path analyses, showing that perceived service quality has a significant effect on trust in mobile-government. Their results indicated a significant positive relationship between perceived service quality and increased perceived trust among users (β = 0.501, p < 0.001).

The correlation between trust and utilization is positive and moderate, supporting H2: r(298) = 0.48, p < 0.001. When respondents feel that the online systems are more reliable than physical government offices and that government agencies take full responsibility for any security breach, it motivates them to use them effectively. These results are in line with those of Zhu et al. (2020) in that trust in technology is positively related to citizens’ intention to use e-services (β = 0.473, p < 0.000). Correspondingly, Mensah et al. (2020) stated that government trust is positively related to the behavioral intention to use. The author utilized a structural model extended from the unified model of electronic government adoption (UMEGA), a parsimonious model based on the eGovernment-specific appropriate context with constructs originating from technology adoption models (Dwivedi et al., 2017), eGovernment services, with the results being positive and significant (β = 0.387, p < 0.05).

The correlation between utilization and SWB is positive and very strong, supporting H3: r(298) = 0.91, p < 0.001. Respondents access public services through government websites to enhance their social status, either individually or for their organization. It provides a wider choice of interactions with different functions from the comfort of their offices/homes compared to interactions within the physical government office. In their integrated model, Pérez-Morote et al. (2020) used expectancy disconfirmation theory (EDT) to prove that technology use (eGovernment and social media) expectations affect satisfaction end-result. Whelan et al. (2020) also applied the SOBC paradigm to depict how social media use as covert behavior influences academic performance as a consequence. Similarly, this paper shows that overt or covert behavior regarding technology use influences SWB.

A multiple linear regression done by Khamis et al. (2003) revealed that the higher the R2 value, the better the model fit. Similarly, Hagen (2015) stated that, in regression analysis, three results are most important. The first is the multiple correlation coefficient (R), which measures the correlation strength between the predicted and observed variable; the higher the R-value, the stronger the correlation (Cook, 1977). The second is the coefficient of determination (R2), which explains to what extent independent variables can explain variation among dependent variables (Menard, 2002). The third is ANOVA, which is the analysis of variance and which tests two essential aspects. First, ANOVA determines whether the regression model predicts an outcome variable or not using degrees-of-freedom and F statistics. Second, it confirms whether the results are statistically significant and whether the results can be generalized to the country’s entire population.

This study adopts a Stepwise multiple regression approach to evaluate whether sector, internet use frequency, the highest level of education, gender, and age socio-demographic variables predict trust. The results are shown in Table 5. In model 1, the highest education level was entered into the regression model, and this significantly predicted trust: β = 0.603, t(298) = 13.06, p < 0.001. Highest education level also explained a significant proportion of variance in trust: R2 = 0.36, F(1, 298) = 170.51, p < 0.001. In model 2, age was entered into the regression model, and this significantly predicted trust: β = –0.198, t(297) = –4.02, p < 0.001. Age increased a significant proportion of variance in trust: R2 = 0.39, F(2, 297) = 97.66, p < 0.001. In model 3, internet use frequency was entered into the regression model, and this significantly predicted trust, β = 0.119, t(296) = 2.18, p < 0.001. Internet use frequency also increased a significant proportion of variance in trust: R2 = 0.41, F(3, 296) = 67.51, p < 0.001. Sector and gender were not entered into the regression model [sector (t = 0.41, p > 0.05); gender (t = 1.22, p > 0.05] as they are non-significant in trust prediction (Fritz and Berger, 2015). The best regression model for predicting trust is model 3, with the highest education level, age, and internet use frequency as predictors; it is the most significant and has the highest R2 value (Li and Pham, 2019). Similarly, integrating five theories being; the digital divide theory, TAM, diffusion of innovation theory, and expectancy disconfirmation theory, Pontones-Rosa et al. (2021) conducted a factor analysis using the principal components method with Varimax rotation and bivariate analysis with contingency tables between the satisfaction variables and the municipal, demographic, socio-economic, and technological variables the ICT access, quality, and use. Their results had high factor loadings or significant weights, highlighting different population profiles according to satisfaction levels with ICT development and confirming their hypothesis.

Model 3 shows that the highest education level coefficient is positive and significant (B = 0.461, p < 0.001), indicating that trust also increases as the education level increases. The age coefficient is negative but significant (B = –0.187, p < 0.001), meaning that as age increases, trust decreases. Finally, the internet use frequency coefficient is positive and significant (B = 0.119, p < 0.001), which means that as internet use frequency increases, so does trust. Courville and Thompson (2001) suggested that, in multiple regression analysis, the use of beta weights only is not enough because they can change when more predictors are added or removed. Therefore, predictor variables cannot independently predict the value of the dependent variable if they are highly correlated, causing multicollinearity (Fox and Monette, 1992). Multicollinearity can be detected using tolerance and variance inflation factor (VIF). A small tolerance and a high VIF value denote high collinearity (Hair et al., 1998). A tolerance value under 0.20 suggests very high multicollinearity (Weisburd and Britt, 2014), and predictors with VIF values greater than 3.3 should be excluded (Diamantopoulos and Siguaw, 2006). Results indicated that multicollinearity was not a concern (highest education level: tolerance = 0.63, VIF = 1.60; age: tolerance = 0.83, VIF = 1.20; internet use frequency: tolerance = 0.68, VIF = 1.48).

This study makes several theoretical contributions. First, we implemented the SOBC model to construct a theoretical concept of the antecedents, mediators, and consequences of eGovernment efficiency concerning citizens’ SWB. Most previous studies have adopted the TAM and UTAUT frameworks to examine the relationship between eGovernment and citizens. However, the present study utilized the SOBC model, a comprehensive model of human behavior, to explore this multifaceted human behavior better. Second, according to the SOBC model, trust is a cognitive process that plays an essential mediating role in explaining the consequences of human behavior. The present study’s results support this concept of the SOBC model; trust mediates the relationship between eGovernment efficiency and utilization, which then triggers effects on SWB. Essentially, when eGovernment is efficient, and utilization is effective, this results in either two of the three SWB components [satisfaction (LS) and positive affect (PA)]; if not, this results in negative affect (NA) (Andrews, 1974; Diener and Emmons, 1984). Third, focusing on these different mediating effects of trust, we emphasize the moderating impact of socio-demographic characteristics in exploring variation in trust. The stepwise multiple regression results show that the highest education level, age, and internet access frequency together moderate trust. Information technology usage has expanded intensively; therefore, gender does not influence performance expectancy (Goswami and Dutta, 2015).

The findings from this study reveal practical suggestions for developing systems that consider users’ feelings and behavior patterns to encourage trusting behavior and ultimately have a more significant favorable influence on users’ SWB. Despite the known orotundity concept that the best way to prevent technology issues is not to use it as much (Whelan et al., 2020), this concept does not apply when vital public services are mandatory online only. Since governments provide constraints in citizens’ day-to-day behavior to shield them from foreign intrusion and regularly cater for their happiness and overall wellbeing (Traill et al., 2014), the same objectives should be considered in eGovernment service delivery. Future research should examine whether interventions intended to make eGovernment efficient incorporate factors such as privacy, security, citizen and social engagement, content, services, and usability (Holzer and Kim, 2006). These factors are essential to strengthen the relationship between the online government environment and trust development. In addition, future research can explore the effectiveness and efficiency of online banking systems that enhance service quality, customer satisfaction, customer trust, and loyalty (Yoon, 2010; Chu et al., 2012) and benchmark insights and merge them into eGovernment systems to enhance efficiency. By so doing, government agencies will be able to accommodate the positive wellbeing of citizens concerning accessing essential public services online. Since demographic characteristics, such as internet use frequency and highest education level, influence trust in eGovernment, exposure and education significantly impact the utilization rate. This study recommends implementing eGovernment service access and using training programs and channels specifically formulated for citizens to affect and increase digital literacy positively. Since more people use social media daily, this could be achieved by posting short videos tutorials on government agencies’ social media pages. Such initiatives could provide more opportunities to citizens and promote the unbiased usage of public services in developing communities.

Just like any empirical research, this study has limitations. First, while this study implemented the SOBC theory to explain the measures and consequences of eGovernment efficiency, the variance explained by the study model is reasonable. Further investigations can examine, validate and argue whether other variables aligning with the SOBC model have more predictive power. For instance, different cognitive processes may mediate the relationship between eGovernment efficiency and SWB.

Second, the study used primary data with appropriate sample data corresponding with the social science statistical rule of thumb. However, this could limit the generalization of the findings because questionnaire responses depend on the respondents’ current mood (Siniscalco and Auriat, 2005). However, Lessa (2019) review indicates that previous research has documented eGovernment challenges in developing countries. This study, therefore, suggests that future research should validate the present research results using secondary data, e.g., reviews and other official records from government agencies and a different SEM methodology such as partial least squares (PLS).

Third, as we gathered data from users in one developing country (Botswana), cultural differences may affect the study’s findings (Polkinghorne, 2005) since different cultures have different habits, online experiences, education structures, and internet frequency. Network infrastructures challenges elevate doubt or lack of trust in utilizing e-services (Samboma, 2019). During data collection, when contacted, some respondents said, “I will fill the survey as soon as I have internet access.” Due to this, we noticed that internet services hindered our survey accessibility which is mainly the same challenge toward accessing other e-services. Ntshwarang et al. (2021) pointed to infrastructure and poor internet access off-campus as the main challenges faced concerning electronic learning in the University of Botswana during the COVID-19 pandemic. Therefore, future research should consider this direction, possibly by comparing two or more developing countries, or even including developed countries, to validate the present study’s results using a bigger sample size.