Several definitions of e-government are currently being applied worldwide. They differ depending on the purpose of the definition. Table 1 below highlights some of the definitions.

Furthermore, the World Bank [12] indicates that e-government is a mechanism for whole-of-government public sector modernization that places the citizen at the center of the reform.

Trust building is a cumulative process where the level of trust in the earlier stages affects the level of trust in the later stages and impacts the development of

longer-term trust relationships [13] . Table 2 presents trust definitions.

In e-government development context, there are several overlapping and consistent factors that impact the building of trust. These factors could be classified in two major categories [2] :

Preinteractional factors:

· Individual behavioral attributes: individual demographics, culture, past experiences, propensity to trust, benevolence, credibility, competency, fairness, honesty, integrity, openness, general intention to use e-services.

· Institutional attributes: organizational reputation, accreditation, innovativeness, general perceived trustworthiness of the organization.

· Technology Attributes: interface design, public key encryption, integrity.

Interactional factors:

· Service attributes: reliability, availability, quality, and usability.

· Transactional delivery atributes: usability, security, accuracy, privacy,

interactivity, quality.

· Information content attributes: completeness, accuracy, currency, quality.

Figure 1 depicts the architecture of e-government with the following layers:

Presentation layer identifies and describes the system users, who require access to government information at different capacities and channels through which information is accessible.

E-government layer: This layer focus on integration of different organization data and services into one stop called government portal.

Business layer provides a functional rather than organizational view of the government’s lines of business processes with the aims of mapping existing and updating the processes as well as managing them. It is imperative that the business processes are simplified and understood by all stakeholders.

Information architecture layer is divided into two components:

· The service classification sub-layer: this layer includes the legacy systems that need to be integrated into online services delivery, new online systems, back-office systems, messaging and directory services.

· The data standardization sub-layer: the elements of this layer include standardization of data and information formats, metadata, and data dictionaries. This layer ensures the integration and interoperability of services.

Technology architecture layer: This layer consists of technical infrastructure such servers, storage backups, networks (LAN, MAN or WAN), extranets, internet, cybersecurity solutions.

Based on data collected from UNDESA, the state of e-government in Cameroon

is presented from 2003 to 2022 as well as the state based on the e-Participation sub-index in the same period in order to illustrate of citizens’ participations and usage in e-government services in the country.

1) E-Government Development Index

The E-Government Development Index (EGDI) is a composite benchmark of e-government development consisting of the weighted average of three independent component indices: the Online Services Index (OSI), the Telecommunications Infrastructure Index (TII), and the Human capital Index (HCI).

Figure 2 depicts the evolution of EGDI in Cameroon, which is evaluated on one (1) as well as its ranking on 193 member states.

Figure 2 indicated that e-government in Cameroon is improving in slow rate and moreover the country’s still remain low for the past two decades.

2) E-participation sub-index

Public participation is a key dimension of governance, and its importance is highlighted in a number of SDG indicators and targets, including target 16.7, which calls for ensuring “responsive, inclusive, participatory and representative decision-making at all levels”.

The use of information and telecommunications technology to engage people in public decision-making and services delivery is an essential part of e-government, and since 2001 the Survey has regularly tracked developments in e-participation as reflected in the relevant features of national e-government portals and websites.

The E-Participation Index (EPI) assesses online participation utilizing a three-point scale that distinguishes between the provision of information (whereby the Government provides information to people), consultation (whereby the Government consults on policy or on services delivery at different stages of the process and possibly provides feedback), and decision-making (whereby the

Government involves people in decision-making).

Figure 3 indicates the scoring of e-participation in Cameroon, which is evaluated on one (1) as well as its ranking on 193 member states.

For the past two decade, Figure 3 shows low scores and ranking of e-participation in Cameroon.

3) Security Risk in Cameroon’s Public Administration

As shown above, Cameroonian government is promoting the use of Information and Communication Technologies (ICT) particularly Internet-technology to enhance the government’s service delivery as well as its internal operations. However, the potential for improvements in service delivery and internal operations come with many of the security risks faced by existing systems as well as with new risks.

In Cameroon, the problem of securing information and computer networks has become important as government agencies increase their dependence on networks or information systems. According to Cameroon’s Government Agency in charge of ICT development and cyber security [23] , the increase in the number of attempts to penetrate the integrity of government networks has been exponential in the past decade as illustrated in Figure 4. These incidents often have a profound impact on organizations, particularly for government agencies; it can be the loss or access to secret or sensitive information.

The electronic information systems today are as complex as the business relationships they need to serve. The words ‘Information Security’ are now familiar at the highest levels of the Public Administration. Information security, when approached from a corporate perspective, is an enabler of traditional business goals in an electronic environment.

Colesca [3] found significant positive impact of some individual factors on trust in e-government such as gender, age, disposition to trust, internet experience, and education [33] . Indicates that Cameroon has wide diversity of culture and the country is considered as Africa in miniature.

In this context, individual cultural perception is also considered. Familiarity with internet is one of the most individual characteristics that influences trust in e-government [2] [32] ; added to the previous factors social class as an antecedent to the confidence in e-government services.

Based on the cited literature review and the Cameroonian context, the following hypotheses are proposed:

H₁: Females have more trust in e-government services than males.

H₂: Younger citizens have more trust in e-government services e-government services than older ones.

H₃: Educational level positively influences citizen’ trust in e-government services.

H₄: Income positively effects citizens’ trust in e-government services.

H₅: Internet experience positively effects citizens’ trust in e-government services.

H₆: Cultural perception negatively effects citizens’ trust in e-government services.

Technology is a vital component for IT-based government service adoption since it affects individual beliefs that interactions with these systems are expected and can be trusted [32] [34] .

Our literature review reveals that technological aspects are Perceived Usefulness, Perceived Ease of Use, Information Quality, and Service Quality.

1) Perceived Usefulness

Perceived Usefulness of e-government is related to the degree to which the e-government services is useful, helpful, more productive and efficient to make citizen’s life easier [35] , while Gefen and Straub define Perceived usefulness as an extent a citizen believes that using e-government system would enhance performance and improves his/her efficiency [36] .

Furthermore, citizens will find e-government services useful if it assists them to search for information they want and usual users evaluate system according to task-oriented outcomes. Perceived Usefulness was found to be significant constructs in several e-government adoption literature [24] [37] .

2) Perceived Ease of Use

Perceived ease of use is the degree that users believe that using the system would be easy and would be free of mental effort [38] , in fact, the ease of use is another major determinant of attitude towards use an information system. In other words, users will have a stronger feeling about the trustworthy of the service provider if the provider delivers a service in an easy-to-use manner. Chiu study results on web-based learning environment show that perceived ease of use is associated certainly with the continuance intention to use [39] .

For instance, if mobile users have increases in the market it might be a good opportunity for a government to offer easy to use mobile government services to enhance citizens’ easy access to e-solutions. Mahadeo believed, during pre-implementation and implementation stages of electronic government services it is likely that perceived ease of use will have a very strong positive effect on citizens’ intention to adopt the new technology [40] .

3) Information Quality

The information system output includes accuracy, extent of completeness, extent of relevance, the substances, also whether the information in time, is measured in Information Quality. In addition, information quality also correlate with how the information is presented to the potential users as it needs to be displayed in a well-organized way, up-to-date, appropriate, accurate, and easy to understand [40] .

When users receive inadequate, unrelated or unclear information regarding the smart government services, it will have a negative influence on their trust level of the services [34] [41] . This implies that clients of e-government essential need complete, realistic, clear and up to date information about the services offered on e-government platforms.

4) System Quality

According to Chatterjee and Gupta, service quality is the customer perception on the quality of hardware and software, how reliable the service is, and how supportive the staff on handling the service. It is important that the smart government customer service is always available and responds quickly to citizens’ concerns [42] .

Furthermore, study carried out by Sepasgozar, et al. also noted how utilization of smart government services can avoid the bureaucratic procedures that sometimes appear when citizens directly visit public services [43] . Poor service quality will lead consumers to mistrust the services. Moreover, service users will be hesitant and unwilling to use such services if they found the service quality experience was beyond their expectation level [34] [42] .

Based on the preceding reviews, the following hypotheses were proposed:

H₇: Technological factors positively influence citizens’ trust in e-government.

H₇a: Perceived usefulness has a positive impact on citizens’ trust in e-government.

H₇b: Perceived ease of use has a positive impact on citizens’ trust in e-government.

H₇c: Information quality positively influences citizens’ trust in e-government.

H₇d: Services Quality positively influences citizens’ trust in e-government.

Governmental factors are related to the level of trust and confidence of the citizens towards the government agencies and their beliefs that the government is capable to provide affective services to their citizens [34] . A Large body of research examined the relationship between government agencies and trust in e-government [35] [44] [45] .

The majority of these studies found a significant and positive influence of trust in government on trust in egovernment. In this study the two dimensions of government agencies suggested by Alzahrani and Maan-Alkhateeb are adopted: government reputation and past experience [1] [32] . Consequently, the following hypotheses are proposed:

H₈: Government agency factors positively influence citizens’ trust in e-government.

H₈a: Government Agency Reputation positively influences citizens’ trust in e-government.

H₈b: Government Agency Experience positively influences citizens’ trust in e-government.

Risk is mainly concerned with the issue of privacy and security [32] . A strong relationship was found between risk and trust in r-government in previous literature [32] [44] .

Privacy in e-services context could be defined as the worry about losing control over personal information, whereas security is related to any unauthorized access to the personal data of the individuals [1] .

1) Privacy

Guaranteed privacy is providing formal assurance or promise to citizens that their online transaction and information stored online are confidential and privacy is maintained. McLeod defined privacy as the belief that personal information submitted into a system will remain private [40] .

Citizens’ trust in e-government websites can be affected by the existing privacy policy on e-government websites and those policies should guaranteed citizens’ data protection and the transaction made online [35] . According to Romania context, privacy was found to have the greatest influence on trust in e-government [2] .

2) Security

Guaranteed security is the imaginary gate that brings confidence to the information system or website users in terms of information availability, integrity and confidentiality [35] . Security is a necessary key for the citizen to trust and continue use e-government websites or services. The infrastructure and platform used in these e-government websites or information systems are related to open technologies such as Internet [45] .

Additionally, people will only embrace e-government services if users feel trustworthiness in government websites [24] . Therefore, a user who trusts the system security expects his/her information will not be compromised by hackers and security malware might not use him/her as a zombie to support malicious work such as email spammers or denial of service attacks.

Moreover, the e-government architecture is very complex and has complex integration mechanism where data are shared from a different institution, therefore it’s a responsibility of government to restrict the utilization of private information and secure such information from unintended parties and protect citizen privacy. Based on the preceding reviews, the following hypotheses were proposed:

H₉: Risk factors negatively influence citizens’ trust in e-government.

H₉a: Privacy concerns negatively influence citizens’ trust in e-government.

H₉b: Security concerns negatively influence citizens’ trust in e-government.

1) Data Collecting Tool

The questionnaire was designed as a tool used in collecting data with the aim of assessing the proposed model and to examine the relationships between the constructs. The questionnaire is documented based on the demographics of the respondents. All questions were designed to measure the create variables, which are applied to a five-point Likert scale, varying from 1 (“strongly disagree”) to 5 (“strongly agree”).

Furthermore, the questionnaires were translated from English to French as the two languages are official in Cameroon.

To ease data collection and limit resources available, the data collection tool was deployed electronically through Google online survey questionnaire. Questions’ items were selected from the proposed model that is based on [38] [40] .

2) Sampling and data collecting

A random sample was drawn from Cameroonian citizens living in Yaoundé since it is a metropolitan city; Citizens originating from all the ten regions of the country were included in the sample. The questionnaire was distributed physically and electronically through social media: Facebook, and WhatsApp Groups where the citizens could click on the link to have the online questionnaire.

However, with the low culture of online responding, few citizens responded to online questionnaire on their own. Therefore, the hybrid method was adopted and surveyors could assist the respondents physically or electronically through WhatsApp calls for them to respond to the online questionnaire.

The data collection process was completed in six weeks as from March 13, 2023. Figure 8 presents an example of data captured.

Our data analysis is based on the implementation of the hypothesis-testing algorithm shown in Figure 9. The algorithm is based on Simple linear regression and multiple linear regression statistical concepts [46] [47] [48] , which are applied to determine the explanatory factors (independent variables) of citizens’ trust in online government services (dependent variable).

Multiple regression is an extension of simple linear regression. It is applied when one wants to predict the value of a variable based on the value of two or more other variables. The predicted variable is called the dependent variable (or sometimes, the outcome, target or criterion variable). The variables used to predict the value of the dependent variable are called the independent variables (or sometimes, the predictor, explanatory or regressor variables).

In general terms, Linear regression is a modelling technique for analysing data to make predictions. In simple linear regression, a bivariate model is built to predict a response variable (Y) from an explanatory variable (X). In multiple linear regression the model is extended to include more than one explanatory variable ( X 1 , X 2 , ⋯ , X p ) producing a multivariate model as our research model proposed in Figure 6. Figure 9 showed the Hypothesis-testing algorithm applied.

1) Statistical Significance (Sig.)

a) Simple Linear Regression (SLR)

The simple linear regression model is given by the following equation:

Y i = β 0 + β 1 X i + e i (1)

where β₀ is the intercept, β₁ is the slope of the regression line, and e_i is the residual error. Y_i and X_i are respectively dependent and independent variables.

The parameters (β₀ and β₁) are estimates drawn from a distribution of possible values of Y_i and X_i input into SPSS Software for computing the model in Figure 9.

The standard error of the estimate shows us the spread of this distribution, and the Sig. enable us to know whether these estimated paramaters (β₀ and β₁) are statistically different from zero. The mathematical model of statistical significance (Sig. or p-value) is presented in subsection.

If the estimate for the parameter (β₁) could be zero, then it could be that there is in fact no relationship or a zero coefficient and a flat line of best fit. A value which is not statistically significant is indicated by a p-value greater than 0.05 (reference). For this research model, if p < 0.05 then the estimates of the parameters are statistically significant and inferred that there is an association between the variables (Y_i and X_i).

b) Multi Linear Regression (MLR)

Multiple linear regression extends simple linear regression to include more than one explanatory variable. In both cases, the term ‘linear’ is used assuming that the response variable is directly related to a linear combination of the explanatory variables ( X 1 , X 2 , ⋯ , X p ).

The equation for multiple linear regression has the same form as that for simple linear regression but has more terms:

Y i = β 0 + β 1 X 1 i + β 2 X 2 i + ⋯ + β p X p i + e (2)

In a model with p explanatory variables, each explanatory variable has its own β_coefficient.

Multicollinearity occurs when 2 or more predictors ( X 1 , X 2 , ⋯ , X p ) in one regression model are highly correlated. Typically, this means that one predictor is a function of the other. If multicollinearity is serious and it is ignored, the numerical estimation will be problematic, and the estimated parameters ( β 0 , β 1 , ⋯ , β p ) will not reflect the reality. In order to reduce Multicollinearity effect within the scope of this research, principal component analysis of explanatory variable ( X 1 , X 2 , ⋯ , X p ) is applied with the aim of discerning possible structural effects.

The process of MLR is identical to that of SLR within SPSS Software by including additional explanatory variable in the specification stages. The software compute the parameters ( β 0 , β 1 , ⋯ , β p ) and statistical significance (Sig. or p-value) for each explanatory variable (X_p).

1) Cronbach’s Alpha

Before embarking on data analysis, the internal consistency of the data should be tested. The indicator used for this is Cronbach’s alpha defined as follows:

α = k k − 1 × ( 1 − ∑ i n s i 2 s 2 ) (3)

where α represents Cronbach’s alpha,

k is the number of items in the scale or test.

s i 2 is the variance of each item i.

s² is the total variance of the scores of all the items.

The result of Cronbach’s alpha will be a value between 0 and 1. A value closer to 1 indicates better internal reliability of the scale or test, which suggests high consistency between items.

2) Analysis of Variance (ANOVA)

Analysis of variance (ANOVA) is a statistical method used to compare the means (averages) of several independent groups. In this research, ANOVA will be used to assess the effect of socio-demographic factors and other factors on citizens’ trust in e-government services as presented on our research model. The following represents the summary of the one-way ANOVA:

a) ANOVA Assumptions:

The one-way ANOVA tests the null hypothesis indicates that the group means are equal. The null hypothesis (H₀) is represented as follows:

H₀:

μ 1 = μ 2 = ⋯ = μ k (4)

where μ 1 , μ 2 , ⋯ , μ k represent the averages of the k groups.

b) Calculation of intergroup sum of squares (SCI):

Calculate the intergroup sum of squares, which measures the variability between group averages. The SCI formula is as follows:

SCI = ∑ i n ( n i ∗ ( X ¯ i − X ¯ ) 2 ) (5)

where n i is the size of group i, X ¯ i the average of group i and X ¯ is the overall average.

c) Degree of freedom:

Calculate the degrees of freedom for the SCT, SCI and SCE. Degrees of freedom are used to calculate statistics of value test. The formulas for the degrees of freedom are:

DDL tot = n − 1 (6)

DDL int = k − 1 (7)

DDL err = n − k (8)

where n is the total number of observations, and k is the number of groups.

d) F statistic:

Calculate the F-statistic by dividing the intergroup variance by the intragroup variance, weighted by the corresponding degree of freedom. The formula for F statistic is:

F = SCI DDL int SCE DDL err (9)

Then, conduct a significance test by comparing the value of the F statistic to a critical F value with the appropriate degrees of freedom. The higher this value, the more meaningful the results. Significance indicates whether the result is obtained by chance or due to the independent variable.

e) Significance test:

The value of “Sig.” or p-value is known as the significance, which indicates whether the difference between the groups is large enough to be taken into account due to the independent variable rather than being random. Generally, a p-value (probability of significance) of less than α (level of significance) is a predefined threshold set by the researcher and It is usually 0.05.

For the scope of the research, α is considered to be 5%; meaning that the risk or probability that the difference is random is less than 5%. The lower the value of “Sig.” or p-value, the higher the level of significance.

In broader terms, p-value is a statistical measure that helps researchers determine if their hypothesis is correct. This helps to determine the significance of the results. The null hypothesis is a default position that there is no relationship between two measured phenomena.

Application of p-value

P-value is a statistical measure that helps researchers determine if their hypothesis is correct. This helps to determine the significance of the results. The null hypothesis is a default position that there is no relationship between two measured phenomena. It is denoted H₀. An alternative hypothesis is the one you would believe if the null hypothesis is found to be false. Its symbol is H₁ or Ha.

The p-value is a number between 0 and 1. There are charts, spreadsheets, and statistical software to help calculate the p-value. The level of significance (α) is a predefined threshold set by the researcher. It is usually 0.05. A very small P-value, which is below the significance level, indicates that you reject the null hypothesis. The P-value, which is above the significance level, indicates that one fails to reject the null hypothesis.

To test each of the hypotheses (H₁…H₆) respectively associated to demographic characteristics (gender, age, educational level, income, Internet experience and cultural perception), a separate Analysis of variance (ANOVA) test is conducted where the testing value is statistical significance (Sig.) also known as p-value described in sub section 3.2.2 where “Sig.” is between 0.000 and 1.000.

The variable has effect or the associated hypothesis is accepted when “Sig.” ≤ 0.05. The variable effectiveness increases when “Sig.” approaches 0.000. Table 5 summarizes the results.

In our data collection process the variable occupation (H₀) is included which was not initially in our proposed model and based on the calculated F and p-value, H₀ has effect on trust in e-government services since its p-value is less than the threshold (0.05). Based on the results shown on Table 5, similar results

recorded for H₂ through H₅ meanwhile the associated variables for H₁ and H₆ have no effect on trust in e-government services since their respective calculated p-values were not significant (Sig.).

Following our proposed hypothesis-testing algorithm, hypotheses H₇, H₈ and H₉ with respective associated variables: technological factors, governmental factors and risk factors are analyzed based on a two-step linear regression. Firstly as separate influence of each variable in the construct is examined to test the sub-hypotheses and secondly, multi linear regression for each hypothesis.

1) Hypothesis H₇ (Technological Factor)

The hypothesis (H₇) with construct as technological factor consists of four sub-hypotheses (H₇a through H₇d) with associated variables as perceived usefulness, perceived ease of use, quality of information and quality of service. The results of the simple linear regression of each of these sub-hypotheses are shown on Table 6.

Table 6 illustrates that H₇a, H₇b, H₇c and H₇d are supported and the variables associated with these sub-hypotheses are found to have a positive influence on citizens’ trust in e-government since for each case, p-value (sig.) is less than 0.05. Information quality has the strongest influence (standardized β = 0.137), whereas service quality has the weakest influence (standardized β = 0.079). Detail illustration is shown on Figure 11 below.

2) Hypothesis H₈ (Governmental Factor)

The hypothesis (H₈) with construct as governmental factor consists of two

sub-hypotheses (H₈a and H₈b) with associated variables as past experience and government reputation. The results of the simple linear regression of each of these sub-hypotheses are shown in Table 7 below.

Table 8 illustrates that the two sub-hypotheses (H₈a, H₈b) that composite H₈ are supported and the variables associated with these sub-hypotheses are found to have a positive influence on citizens’ trust in e-government since p-value is less than 0.05 for both variables. Government reputation has a stronger influence (standardized β = 0.203) than Past experience (standardized β = 0.061). Figure 12 below illustrates the results.

2) Hypothesis H₉ (Risk Factor)

The hypothesis (H₉) with construct as risk factor consists of (H₉a and H₉b) with associated variables as security and privacy. The results of the simple linear

regression of each of these sub-hypotheses are shown in Table 8.

Table 8 illustrates that the two sub-hypotheses (H₉a, H₉b) that composite H₈ are supported and the variables associated with these sub-hypotheses are found to have a negative influence on citizens’ trust in e-government since p-value is less than 0.05 for both variables. Security has a stronger influence (standardized β = −0.057) than privacy (standardized β = −0.026). Details are illustrated on Figure 13 below.

Multiple linear regression is applied to examine the combined influence of technological factors, governmental factors, risk factors and their interactions on

citizens’ trust in e-government. Tables 9-11 summarize the results:

Table 9 shows that the determination factor R² is equal to 4.5%, which means that the examined constructs explain about 4.5%, of the change in citizens’ trust in e-government.

a. Predictors: (Constant), Governmental Factors, Technological Factors, Risk_Factors

b. Predictors (Independent Variable): (Constant), Governmental Factors, Technological Factors, Risk Factors.

As shown on Table 10, the model is globally significant (F = 5.522 and p-value = 0.001). Thus, the factors considered entirely indicate citizens’ trust in government services. At 5% threshold of the p-value, the coefficients of the three factors all seem significant. Thus, technological, government and risk factors are the elements that affect citizens’ trust in government services.

Table 11 illustrates the influence of each independent construct, the largest influence is for the governmental factors (standardized β = 0.172, p-value = 0.001), followed by technological factors (standardized β = 0.097, p-value = 0.004), then risk factors (standardized β = −0.085, p-value = 0.002). Figure 14 depicts the results.