An Empirical Study of Factors Influencing the Perceived Usefulness and Effectiveness of Integrating E-Learning Systems during the COVID-19 Pandemic Using SEM and ML: A Case Study in Jordan

The purpose of this research paper is to identify and test the factors influencing the perceived usefulness and perceived effectiveness of adopting an e-learning system from the perspective of teachers in public and private schools as well as the United Nations Relief and Works Agency for Palestinian Refugees in the Near East (UNRWA) in Jordan during the first wave of the COVID-19 pandemic in the academic year 2019/2020. Based on the findings and best practices, the study intends to make appropriate recommendations to decision-makers. Its significance stems from the use of scientific tools of research and investigation, and it aims to ensure the quality and effectiveness of Jordanian schools’ e-learning systems. The study’s hypotheses were verified by electronically collecting 551 questionnaires from teachers in Jordan. To test the study hypotheses, the empirical validity of the research model was set up, and the data were analyzed with SPSS version 21.0. Structural equation modeling (SEM), confirmatory factor analysis (CFA), and machine learning (ML) methods were used to test the study hypotheses and validate the properties of the instrument items. Nineteen variables and one mediating variable were studied. The study found that independent variables pertaining to technology (relative advantage, compatibility, top management support, communication technologies, competitive pressure, technology competence, information intensity, and work flexibility) and moderating variables pertaining to the teacher’s personal income and those pertaining to school (school size, education program, and work sector) had a positive effect on teachers’ perceived usefulness of adopting e-learning systems during the COVID-19 pandemic. On the other hand, independent variables pertaining to technology (complexity and collaboration technology), moderating variables pertaining to the teacher (age, education level, and gender), and moderating variables pertaining to school (educational stage, number of students) were not supported.


Introduction
The world is undergoing significant transformations and rapid development in technological, economic, social, and other fields, which have resulted in the emergence of new concepts, such as electronic-learning (e-learning) and distance learning. The advancement of Information and Communication Technology (ICT) has encouraged many the effects of physical distancing and increased social media knowledge and use. The fifth study [8] concentrated on distance learning education before and after and during lockdown of COVID-19. The sixth study [9], investigated the difficulties faced by Chilean teachers during SARS-CoV-2, while [10] concentrated on the communication problem within the context of university education during COVID-19. Furthermore, [11] evaluated online education from students' perspectives.

Research
Focus on [19] e-learning system for pre-school [39] e-learning and digitalization in primary [20] online learning: class size in K-12 [24] online class size and instructor performance [27,28] online class size [29,30] online teaching on faculty load [40] e-learning motivation: developing countries

Research Hypothesis Development
The research hypotheses of the current study were developed based on previous literature commonly referenced in the e-learning arena, such as the TAM suggested by [1], the Innovation Diffusion Theory proposed by [2], and the factors affecting the evaluation of e-learning systems' success as studied by [3]. In this section, the paper presents 20 hypotheses. There are three major elements in this study: technology, school, and teacher. The three elements and the interrelationships among the variables will be explained further in the next sections.
The independent variables pertaining to technology are: relative advantage (RA), complexity (CX), compatibility (CP), top management support (TM), communication technologies (CT), collaboration technology (CL), competitive pressure (CM), technology competence (TC), information intensity (IN), and work flexibility (WF). The moderating variables concerning the teachers are age, education level, gender, and personal income. The moderating variables relating to school are the schools' size, education program, work sector, educational stage, and number of students. Furthermore, the mediating variable is perceived usefulness (PU), and the dependent variable is perceived effectiveness (PE).
According to [41] and based on the work of [2], RA is defined as "the degree to which an innovation is perceived as being better than the idea it supersedes" (p. 229). Additionally, the same source indicated that RA is the strongest predictor of the rate of adoption of the innovation. Moreover, according to [42], RA positively affects the users' intention to use the system among different participants. The following hypothesis is proposed based on the previous research: H1. Relative advantage (RA) has a positive effect on teacher's perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
According to [41] and based upon the work of [2], CX is the level of difficulty with which an innovation is perceived to be understood and used. Consequently, according to [31], faculty members may be challenged to change their teaching methodology. CX is the one of the "characteristics of innovations" listed by [41] and based on the work of [2] that can foretell "the rate of adoption of innovations." The rate of adoption has been defined by the same source as "the relative speed with which an innovation is adopted by members of a social system." According to the review study conducted by [41][42][43], CX is negatively correlated with the rate of adoption. Based upon the preceding research, the following hypothesis is proposed: H2. Complexity (CX) has a negative effect on the teacher's Perceived Usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
CP, as stated by [41], is "the degree to which an innovation is perceived as consistent with the existing values, past experiences, and needs of potential adopters" (p. 15). Furthermore, Ref. [44] stated that "the compatibility of the organization values, information technologies and infrastructure related to the new adopted systems in addition to the existing internal and external processes increase the acceptance of the new adopted information systems". Refs. [42,43,45], citing the work of [22,46], confirmed that CP has a significant positive and direct effect on PU and behavioral intention. Based upon the preceding research, the following hypothesis is proposed: H3. Compatibility (CP) has a positive effect on the teacher's perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
Two studies, Refs. [32,47], argued that TM is the most important success factor. The same source classified success factor into "Must-Have Factors" and a "Nice-to-Have Factors." Furthermore, management support was divided into "Top management support to employees" and "Management assistance to employees." Additionally, Ref. [33] stated that "The TM and consistency is critical to implementation of any project." In addition, Ref. [48] adopted the same idea. Therefore, the following hypothesis is proposed: H4. Top management support (TM) has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
Refs. [2,41] argued that communication is the second element of the diffusion of innovations process. Communication is defined by [41] as "a process in which participants create and share information with one another in order to reach a mutual understanding" (p. 5). CT has also been discussed by many researchers from many aspects: availability, speed, effectiveness, and resource. According to [49] and based upon the research of [50,51], "the availability of several channels of communication facilitates the constant monitoring necessary for such an interactive and flexible learning experience." Moreover, [3] claimed that to provide a good coverage for the educational system quality, the institutions must grant effective CT. Ref. [13] explained, "Communication resources such as discussion boards enable learners to participate in collaborative learning with other students and with educators. Through an online course, students can share ideas at anytime from anywhere" (p. 149). Hence, based upon the preceding research, the following hypothesis is proposed: H5. Communication technology (CT) has a positive effect on teacher's perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
CL enables learners and educators to collaborate, carry out discussions, and interact when presenting ideas and questions using such media as text, pictures, sound, and animation. According to [13], "Communication resources such as discussion boards enable learners to participate in collaborative learning with other students and with educators. Students can share ideas at anytime from anywhere through the online course" (p. 150). Two studies, [52,53], stated that "the use of ICT tools such as laptop computers, electronic pads, smart phones, along with the broadband internet, interactive Web 2.0 technologies and cloud applications have enhanced both, teaching and learning in the schools". Furthermore, Ref. [13] argue, "Communication resources such as discussion boards enable learners to participate in collaborative learning with other students and with educators" (p. 150). The studies conducted by [54,55] defined user satisfaction as "a measure of the discrepancy between a user's expectations about a specific information system compared to the perceived performance of the system" (pp. 163, 248). One study [56] argued that if an information system meets users' needs their satisfaction will increase. Based upon the preceding research, the following hypothesis is proposed: H6. Collaboration technologies (CL) has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
According to [57], CP is the level of competitiveness between the organizations that operate in the same business field by improving performance, services, and products to win out in competition and overcome other competing organizations.
Competition is fierce among educational institutions, with institutions striving to deliver programs, courses, activities, and surroundings, as well as electronic services, such as e-learning platforms, tools, and services [58]. Educational institutions are competing to provide the best content in their respective e-learning environment to attract as many students as possible, especially during the COVID-19 lockdown. According to [14,25,26], educational institutions are distinguished for developing the best courses content using the e-learning platforms and online teaching techniques and tools to achieve the courses' intended learning objectives, as if the students were in the classroom. Such competitive pressure has a significant impact on the e-learning environment. Therefore, the following hypothesis is proposed: H7. Competitive pressure (CM) has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 crisis.
TC is synonymous with Computer Self-Efficacy (CSE). Refs. [34,59] stated that TC is the teachers' ability to effectively use technology in the classroom. Furthermore, TC refers to the teachers' knowledge of current and emerging learning systems and technologies, as well as how they can be used to support and improve the learning process. The same term has also been used to mean self-efficacy. Ref. [60] defined self-efficacy as: "In context of computer usage, Computer Self-Efficacy (CSE) is defined as one's belief about his/her ability to accomplish a particular task using a computer" (p. 238). Furthermore, Ref. [59] stated, "Technology self-efficacy refers to pre-service teachers' perceptions of their ability to use technology effectively in the classroom" (p. 78), while [34] stated that "students with higher self-efficacy gain better performance in contrast to those with lower self-efficacy in Internetbased settings" (p. 222). They also stated that "the Internet Self-Efficacy (ISE), which examines learners' confidence in their general skills or knowledge of operating Internet functions or applications in the Internet-based learning condition" (p. 222). Furthermore, Ref. [15] found that students' self-efficacy has a strong and direct influence on the students' capabilities and confidence while using e-learning systems. Based upon the preceding research, the following hypothesis is proposed: H8. Technology competence (TC) has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
Information intensity (IN) refers to the volume and quality of information provided by the e-learning environment. The use of audio, video, text, animation discussion, assignment, quizzes, and exams enrich the e-learning environment, yet it takes a toll on the hardware and software. According to [42,61], such elements need a large volume of information to support and improve the students' cognitive access. Thus, the following hypothesis is advanced: H9. Information intensity (IN) has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
According to [62], WF in E-learning environment is limited to "media representations", and it "provides a flexible cognitive support using different media representations" (p. 174). Refs. [13,35] limited the WF in e-learning environment to time and place. Furthermore, Ref. [23] stated that e-learning course flexibility is one of the six factors affecting learners' perceived satisfaction. In addition, Ref. [36] stated that "content feature and interaction significantly affect performance expectations in a blended e-learning system (BELS)" (p. 155). In fact, Ref. [63] listed flexibility as an advantage of e-learning, as it provides the students with time flexibility, place flexibility, and effort management. Therefore, the following hypothesis is proposed: H10. Work flexibility (WF) has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
According to the TAM model suggested by [1], PU influences the success, acceptance, and use of new technologies. Ref. [64] defines PU as "The degree to which the user believes that using a particular system has enhanced his or her job performance" (p. 51). This study [64] states that PU represents the degree of work improvements related to the performance and productivity of the users after the adoption of information systems, arguing that PU is one of the most important factors that must be considered in assessing the validity of information systems' success. According to [59], PU is a major factor that influences technology integration in impacting pre-service teachers' technology self-efficacy. Moreover, Ref. [21] claimed that the PU positively influences the students' satisfaction with the e-learning courses. The studies [3,15,37,42,59,65] claimed that PU has a significant positive effect on using and accepting the e-learning systems. Therefore, the following hypothesis is proposed: H11. Perceived usefulness (PU) positively influences teachers' perceived effectiveness (PE) of adopting e-learning systems during the COVID-19 pandemic.
According to [44,48], firm size is one of the critical factors related to adopting information system. Moreover, Ref. [66] argued that large organizations are eager to adopt new technological innovation more than small and medium organizations. Thus, the following hypothesis is proposed: H12. School size has a positive effect on teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic.
There are two kinds of educational programs in Jordan: national and international. The Ministry of Education (MOE) develops the national program and for its curriculum uses the Tawijihi stream, which is the General Certificate of Secondary Education Exam [67]. International schools, as defined by [68,69], are those that offer a variety of international curriculums and assessments, such as the International General Certificate of Secondary Education (IGCSE), International Baccalaureate (IB), and Scholastic Assessment Test (SAT). International schools are private schools and are not supported by the government. Refs. [68,70] state that international schools have cross-cultural staff and students.
Schools in Jordan provide various options: government schools offer national programs, whereas private schools offer a variety of national and international programs. The difference between the two programs is that the national program is taught in Arabic, and the international program is taught in English. Thus, the following hypothesis is advanced: H13. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of the educational program.
In the context of this study, work sector refers to public and private schools in Jordan. Private schools, as described by [71], are owned, managed, and funded independently without any assistance from the Jordanian government. On the other hand, public schools in Jordan are owned, managed, operated, and funded by the government. Though the UNRWA owns, operates, and funds schools, only Palestinian refugees are admitted to them. During the COVID-19 pandemic, all Jordanian schools utilized distance learning through the use of different technologies, such as e-learning systems, collaborative platforms, and even instant messaging apps.
According to [16], new learning environments, such as augmented reality, are more effective in public schools than in private schools. The study also found that students in private schools are more motivated to use augmented reality learning environments than students in public schools. According to [38], implementing digital mobile e-learning systems in public and private schools improves school management efficiency. Furthermore, Refs. [12,18,38] claimed that e-learning improves student learning quality and increases student learning effectiveness in both public and private schools. Thus, the following hypothesis is advanced: H14. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of work sector.
Students' educational phases vary across nations. Jordan has three levels of education: pre-school, basic education, and secondary education [72]. Pre-school education is imparted in kindergarten schools to children aged three to five years. Basic education (grades 1-10) is followed by two years of secondary academic or vocational education (grades 11-12). Basic and secondary education are provided free of charge in public schools. In a study conducted by [17], the authors developed an android tool to help develop both cognitive and psychomotor skills for pre-school children, consequently, developing Kids Training e-Learning System (Kotel's). Furthermore, [19,73,74] cited the positive influence of e-learning on pre-school education. Ref. [75] even found that 70% of the secondary school students can finish all e-learning program assignments. Researchers have also pointed out that employing technologies, such as short messages (SMS), messenger, and Skype motivates students to study online.
According to [39], "most students felt that e-learning helps students to have access to a limitless amount of material; shows connections between subjects; develops critical thinking; and supports students' manner of learning" (p. 56). The study also states that "the majority of instructors believed that e-learning is easier and more successful; that it helps to further strengthen teachers' computer abilities; and that it brings out the best in students" (p. 56). Students and instructors believe that e-learning lets teachers and students to share responsibilities for learning and accomplishment. Thus, the following hypothesis is proposed: H15. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of the educational stage.
A study by [20], referencing [24,28], found that the number of students or class size is an environmental aspect that is critical for structuring online courses. Ref. [27] found that decreasing class sizes had substantial and favorable benefits on the academic results in subjects, such as mathematics, physics, chemistry, earth science, and biology. Furthermore, Refs. [20,[28][29][30] stated that class size is closely connected with teacher workloads, teaching styles, practices, class relationships, and student accomplishment.
H16. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differ among the study's respondents in terms of the number of students.
Instructors' background characteristics, such as age, gender, income, and educational level have been gathered to investigate their impact on the adoption and use of e-learning systems. According to [76,77], there is a link between the gender and age of teachers and their aspirations to use computer technology.
According to [78,79], a teacher's higher education degree is an important element that impacts comprehension and efficient and effective usage of computer systems. Other experts, however, disagree, claiming that there is no association between the age, gender, and educational level of users and the impact of utilizing computer systems [80].
Furthermore, the influence of income level on technology adoption has been extensively researched by [40,81,82]. The researchers in [82] discovered a substantial association between personal wealth and the PU of adopting and integrating various technologies in their everyday life. Thus, the following hypotheses are put forward: H17. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of age.
H18. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of teacher education level.
H19. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of gender.
H20. Teachers' perceived usefulness (PU) of adopting e-learning systems during the COVID-19 pandemic differs among the study's respondents in terms of the characteristics of personal income. Figure 1 depicts the study's model, which shows the independent variables, the mediating variable, the moderating variable, and the proposed association between them.

Research Methodology
The methodology used in this research is presented in this section. The research data collection tool, research population and sampling, construct and measurement items, and research methods are all included.

Data Collection, Population, and Sampling
This study is based on a national project conducted in collaboration with the Ministry of Education (MoE) and the University of Jordan. It has been approved by the Faculty of Information Technology and Systems of the University of Jordan and the Ministry of Education-Research and Development based on a proposal and questionnaire submitted to both entities. Thus, the questionnaire and research process were approved by both parties. Furthermore, for construct validation, the questionnaire's content was modified according to the practice of Jordanian educational culture context and based on the results of a pilot study and feedback from six professional academic staff members in this field. The survey instrument was reviewed by a panel of six academic researchers in the areas of education and e-learning to guarantee face validity. Consequently, several questions were modified, and the revised questionnaire was used for pilot testing on teachers in Jordan. Indeed, a pretest was conducted with 25 teachers to check the ease of comprehension of

Research Methodology
The methodology used in this research is presented in this section. The research data collection tool, research population and sampling, construct and measurement items, and research methods are all included.

Data Collection, Population, and Sampling
This study is based on a national project conducted in collaboration with the Ministry of Education (MoE) and the University of Jordan. It has been approved by the Faculty of Information Technology and Systems of the University of Jordan and the Ministry of Education-Research and Development based on a proposal and questionnaire submitted to both entities. Thus, the questionnaire and research process were approved by both parties. Furthermore, for construct validation, the questionnaire's content was modified according to the practice of Jordanian educational culture context and based on the results of a pilot study and feedback from six professional academic staff members in this field. The survey instrument was reviewed by a panel of six academic researchers in the areas of education and e-learning to guarantee face validity. Consequently, several questions were modified, and the revised questionnaire was used for pilot testing on teachers in Jordan. Indeed, a pretest was conducted with 25 teachers to check the ease of comprehension of the questions. Some revisions were made, resulting in an easily understandable survey questionnaire.
The required empirical data for the current research were gathered from teachers in the field located in all governorates of Jordan. According to the data produced by the MoE in 2019, there are 136,062 teachers working in public, private, and UNRWA schools. According to the Morgan Table data, a minimum of 384 teachers is the minimum size of the statistical sample of this study [83]. Indeed, after removing the deficient surveys, 551 valid questionnaires were returned from teachers in Jordan, which reached the suggested guidelines of [83][84][85] regarding the appropriate sample size. The questionnaire was prepared in Arabic and English and distributed electronically using email, WhatsApp, and Google forms. To reach them, a web link to the questionnaire was sent to potential respondents during the period between 5 April and 5 June 2020. To authenticate the respondents' responses, the questionnaire was distributed to teachers through schools' principals, and teacher's syndicate research and development department hence the involvement of the MoE. Teachers were also given the choice to participate by agreeing to this information, or to not participate, and could quit the questionnaire at any moment. All participants voluntarily subscribed to the study, and the data were analyzed anonymously. The researchers did not formally ask teachers for written consent.

Constructs and Measurement Items
To explore the relations among the research variables, a 5-point Likert scale was used that ranges between strongly disagree = 1 and strongly agree = 5; reliability and validity analyses have been conducted; and descriptive analysis has been used to describe the characteristic of the sample and the respondent to the questionnaires besides the independent and dependent variables. In addition, structural equation modeling (SEM) analysis was used to examine the research hypotheses. The measured constructs and the items measuring each construct are shown in Table 2.  WF1: Using E-learning systems during the COVID-19 pandemic provides the possibility to complete the teaching process with a more flexible schedule and working hours. WF2: Using the E-learning systems during the COVID -19 pandemic requires more effort and time to complete the required tasks than teaching students in a classroom. WF3: Using the E-learning systems during the COVID -19 pandemic requires planning in a suitable place and environment to complete the teaching tasks.

Research Methods and Data Analysis
SEM and confirmatory factor analysis (CFA) are research approaches employed in this paper. Since the current research investigates a research model with multiple relationships, it employs SEM, which is a multivariate statistical analysis technique that is used to analyze structural relationships. According to [86], SEM is to utilize factor analysis and multiple regression analysis to analyze the structural relationship between measured variables and latent constructs. To validate the qualities of the instrument items, CFA was used.
In this section, the paper presents the demographic profile of the study, descriptive analysis of the study, measurement model of the study, structural model of the study, and hypotheses outcome. Table 3 presents the demographic data of the respondents, showing that most of the respondents work in schools with 250 and more teachers (51.2%); 96.4% work in the national educational program; 72.1% in the government sector; 67.9% in the primary school, 43.2% ranged from 30 to 40 years; most of them are female (71%), and 62.4% of the respondents earn less than USD 750 per month.   Table 4 indicates how remote teaching technologies were used in distant education. We used ratio estimation to estimate the actual value of a population feature within an acceptable range because most teachers used more than one instrument.  Table 4 shows that 41.74% of the teachers adopted the instant messaging applications such as WhatsApp to send the videos, files, assignments, and examination papers to the students during the first wave of the COVID-19 in Jordan; 19.6% of the teachers used collaboration systems, such as Zoom or Microsoft Teams; and 16.88% of the teachers used the public government platform (Darsak) for teaching and examining the students. The survey also shows that 12.34% used e-learning systems, and 9.44% used free educational applications, whereas none of the teachers used email in the teaching process during the first wave of the COVID-19 in Jordan.

Descriptive Analysis
The mean and standard deviations were computed to describe the replies and attitudes of the respondents toward each topic in the survey. According to [85,87], while the mean represents the data's central tendency, the standard deviation measures dispersion and provides an indicator of the spread or variability in the data. The following formula was used to calculate the level of each item based on [88]: (highest point on the Likert scale − lowest point on the Likert scale)/the number of levels utilized = (5−1)/5 = 0.80, where 1-1.80 represents "very low", 1.81-2.60 represents "low", 2.61-3.40 represents "moderate," 3.41-4.20 represents "high", and 4.21-5 represents "very high". Thereafter, the items were ordered based on their means. Table 5 demonstrates the results.

Measurement Model
CFA was performed to verify the properties of the instrument items. The measurement model shows how latent variables or hypothetical constructs are evaluated in terms of observed variables and represents the validity and reliability of the observed variables' responses for the latent variables [83,89]. Table 6 illustrates the different types of goodness of fit indices used for assessing the current research model. Since the initial CFA model showed an acceptable fit, no items were eliminated, and the results showed that the chisquare (χ 2 ) value of the model was 2278.972, with 674 degrees of freedom (p < 0.05), which entails that the measurement model fit the data. Furthermore, the other model fit indices used for this study were the χ 2 /df (2278.972/674 = 3.381; threshold less than 3 for serious consideration or less than 5 for acceptable criteria), the Incremental Fit Index (IFI) of 0.89, Tucker-Lewis Index (TLI) of 0.87, Comparative Fit Index (CFI) of 0.89, the Goodness-of-Fit Index (GFI) of 0.89, the Adjusted Goodness-of-Fit Index (AGFI) of 0.90, the Normed Fit Index (NFI) of 0.92, and the Root Mean Square Error of Approximation (RMSEA) of 0.066. Based on these fit indices, the measurement model appeared to fit the sample data well [83,84,90]. Table 6 demonstrates the results.  Table 7 shows the factor loadings, Cronbach's alpha, composite reliability, and average variance extracted (AVE) for the variables. All the indicators of the factor loadings exceed 0.50 and thus constitute evidence of convergent validity [89,91]. Indeed, while the measurement reached convergent validity at the item level because all the factor loadings exceeded 0.50, all the composite reliability values exceeded 0.60, demonstrating a high level of internal consistency for the latent variables. In addition, as each value of AVE exceeded the threshold of 0.50 stated by [83,89], convergent validity was demonstrated.

Structural Model
The SEM analysis showed that RA, CP, TM, CT, CM, TC, IN, and WF significantly affected PU; thus, H1, H3, H4, H5, H7, H8, H9, and H10 were accepted. Additionally, PU positively and significantly affected perceived effectiveness (PE); therefore, H11 was accepted. However, CP and CL did not affect PU; thus, H2 and H6 were rejected. Moreover, the coefficient of determination (R 2 ) for the research endogenous variables for PU and PE were 0.345 and 0.447, respectively, which indicates that the model does moderately account for the variation of the proposed model. The results are summarized in Table 8.

Moderating Hypothesis Results
The moderating hypothesis results are discussed in this section for H12 through H20, using ANOVA Analysis and t-test. The independent variables are school size, educational program, work sector, educational stage, number of students, age, teacher education level, gender, and personal income.
For H12, ANOVA test was employed to investigate if teachers' PU of adopting elearning systems during the COVID-19 pandemic differs among the study's respondents in terms of school size. The results of the ANOVA, shown in Table 9, indicate that there are no significant differences regarding school size. The t-test findings for H13 are provided in Table 10 and reveal a significant difference ascribed to PU. For PU, the mean scores for the international program are greater than those for the national program. For H14, ANOVA was used to determine whether instructors' PU of adopting elearning tools during the COVID-19 epidemic differed by job sector among the study's respondents. Table 9 shows the results of the ANOVA, which reveal a significant difference ascribed to job sector. The Tukey post-hoc test also revealed significant differences between the private and public groups.
For H15, the ANOVA test was used to check whether instructors' PU of adopting e-learning tools during the COVID-19 epidemic differed based on educational stage among the study's respondents. Table 9 shows the results of the ANOVA, which reveal no significant difference in favor of educational stage. Moreover, Table 9 reflects the same results for H16, H17, H18, and H19, which is attributed to the number of students, age, educational level, and gender, respectively.
For H20, the ANOVA was employed to investigate whether the PU of adopting elearning systems during the COVID-19 pandemic differs among the study's respondents in terms of personal income. The results of the ANOVA test, shown in Table 9, indicate that there is a significant difference attributed to personal income. Furthermore, the Tukey post-hoc test showed significant differences between the three groups (i.e., less than USD 750, USD 750-less than USD 1500, and USD 1500 or more).

Machine Learning Techniques Validation and Prediction
Machine learning methods have been applied as contemporary technologies in a variety of fields [93,94]. Additionally, other studies [95][96][97][98][99][100][101] used triangulation methods such as these to validate and verify the results in addition to SEM. The research [102] used 19 machine learning techniques. Five Machine Learning (ML) classification methods are evaluated in this research, which transform inherited data from a dataset's input into the required output pattern [93,103]. The five ML models used to develop and evaluate models for e-learning dataset application are: Artificial Neural Network (ANN) [104], Linear Regression [105], Sequential Minimal Optimization algorithm (SMO) for Support Vector Machine (SVM) [106], Bagging using REPTree model [107], and Random Forest [108]. The back-propagation method is used by ANN to calculate the differences in output values between the projected and actual values. The weights and bias parameters of the ANN design are then modified using the error to reduce the difference between the actual and predicted value. The output of the linear regression model is a polynomial function with weighted coefficients for the independent variables, and it depends on the target labels. The training phase involves a series of operations that update the coefficients of the linear function from the training dataset. The SMO method uses the Sequential Minimal Optimization algorithm to update the weighted vectors of the SVM model. The SMO algorithm discovers the minimal values in a sequence of iterative operations to reach the optimal values. Using a random sample of the instances and features from the training set, the bagging technique creates numerous REPTree models, with the average value of the trees predicting the outcome. The Random Forest (RF) is a collection of connected decision tree (DT) models created using a random selection of training data instances and attribute subsets for each sub-tree model. The model's final output is the average value of the DT trees.
The 10-fold cross-validation technique is used in the evaluation methodology to confirm that the model is capable of accurately predicting the desired values. The 10-fold cross-validation method is used in the evaluation phase. This approach chooses 10% of the dataset for testing and 90% for training in a sequential manner (the remaining nine folds). A classifier model is created and assess how well it operates in each procedure. Then, a visual representation of the performance average is displayed.

ML Results and Discussion
This study investigates aspects that influence the problems and validates certain integration techniques. To understand the relationship between the factors (or inputs) and the problems, ML techniques as intelligent methods extract inherited meaningful information from datasets. However, to assess the performance of ML models, we need two datasets. The datasets are from model 1, which has PU as a dependent outcome and ten parameters (RA, CX, CP, TM, CT, CL, CM, TC, IN and WF) as independent inputs. Model 2 dataset studies the influence of PU as input to PE as a dependent variable.
The experimental results are shown in Figure 2 using the evaluation metrics R 2 and Mean Square Error (MSE). The R 2 and MSE values are displayed on the y-axis, and the models are displayed on the x-axis. The expected effect of the independent variables on the dependent variable is shown by the R 2 statistic (target). The MSE determines the average difference between a model's predicted and actual output values, as shown in Figure 3.

Findings and Discussion
The aim of this study is to identify and test the factors influencing the quality and effectiveness of the e-learning system from the perspective of teachers in public and private schools as well as UNRWA in Jordan during the COVID-19 pandemic. The findings and best practices suggest appropriate recommendations for decision-makers. In this sec- With R 2 values of 75.45% and 75.19%, respectively, the SMO and linear regression sequential models perform reasonably well on two database models. Other non-linear ML techniques that produce convergent results include ANN, Bagging REPTree, and Random Forest. The findings indicate that the PU factor, PE of 75.45%, R2, and 90% MSE in model 2 have a weak relationship. In model 1, the ten factors reflect how perceived usefulness affects the perceived effectiveness on adopting e-learning systems during COVID-19 pandemic 75.45% R 2 value and approximately 90% MSE value. The ability of the ML techniques to validate results is to anticipate the actual target from the independent inputs.

Findings and Discussion
The aim of this study is to identify and test the factors influencing the quality and effectiveness of the e-learning system from the perspective of teachers in public and private schools as well as UNRWA in Jordan during the COVID-19 pandemic. The findings and best practices suggest appropriate recommendations for decision-makers. In this section each result is discussed according to the order of the hypotheses. Furthermore, each hypothesis is discussed with supporting findings of this research as well as research from previous studies.
The findings of this study indicate that H1 is supported. As a result, the Relative Advantage (RA) of adopting distant e-learning systems during the COVID-19 epidemic influences overall satisfaction. This result supports the findings of [42].
According to [41][42][43], Complexity (CX) is negatively correlated with the rate of adoption. However, in this study, we find that H2 is not supported. Table 4 showed that most of the teachers in Jordan used the available tools, such as the WhatsApp mobile application, during the first wave of the COVID-19 pandemic, and such systems and applications are not complex and as a result they positively influence the usefulness of the adopted distance e-learning process.
For H3, Compatibility (CP) is defined as the ease with which new adopted systems features interact with the organization's information technologies, infrastructure, and values, hence increasing the acceptability of the new adopted systems [44]. According to the findings of [22,42,46], CP has a considerable positive and direct influence on PU when it comes to adopting distant e-learning systems. Furthermore, the findings of this study validated earlier researchers' findings that CP facilitates the adoption of distant e-learning systems and has a good impact on the effectiveness of the adopted distance e-learning process.
According to [33], Top Management Support (TM) in H4 is a vital aspect for every project. Additionally, Refs. [32,47,48] stressed its importance. Furthermore, the findings reveal that it is supported and confirmed by the assumption specified in H4.
For CT, the researchers [3,[49][50][51] argued that the availability of effective CT facilitates the monitoring, interactiveness, and flexibility of the e-learning experience, which is confirmed by the results of testing H5. The study showed that CT is supported and has a positive effect on teachers' PU of adopting e-learning systems during the COVID-19 pandemic.
According to [13,52,53], Collaboration Technologies (CL) are the technologies that enable learners and educators to participate in collaborative learning processes and share ideas anytime from anywhere through the online course. In this study, the results of testing H6 found that this hypothesis is not supported because most of the teachers and students did not use CL, such as Zoom and MS Teams; the percentage of using such systems by the teachers during the COVID-19 first wave in Jordan is less than 20%.
Compatibility (CP) is the major driver in academic institutions to provide the best software and hardware to be used in e-learning environments. The results of this study confirm that H7 is supported, such that CP positively affects teachers' PU of adopting e-learning systems during the COVID-19 pandemic.
For H8, pertaining to Communication Technologies (CT), Refs. [34,59,60] stated that the CT is the teachers' ability to support and improve the learning process effectively by using the current and new learning systems and technologies. This study supports these researchers' claim, as H8 is supported: TC positively affects teachers' PU of adopting e-learning systems during the COVID-19 pandemic.
For H9, pertaining to Information Intensity (IN), Refs. [42,61] argued that IN refers to the large volume of information and the quality of the information provided by the e-learning environment to support and improve the students' cognitive access. The result of the hypothesis testing confirms that IN has a positive effect on teachers' PU.
For H10, pertaining to Work Flexibility (WF), Refs. [23,62,63] listed WF as an advantage of e-learning and as one of the significant factors affecting learners' perceived satisfaction, as it provides the student with time flexibility, place flexibility, and effort management. This study shows that H10 is supported, and that WF has a positive effect on teachers' PU of adopting e-learning systems during the COVID-19 pandemic.
For H11, pertaining to Perceived Usefulness (PU), Refs. [3,21,37,42,59,65] asserted that PU had a significant positive effect on using and accepting the e-learning systems and students' satisfaction with the e-learning system courses. In this study, H11 has been supported.
According to [44,48,66], firm size is one of the critical organizational success factors related to the adoption of the information systems. This study shows that H12 is supported, the school size has a positive effect on teachers' PU of adopting e-learning systems during the COVID-19 pandemic, and large organizations are more eager to adopt new technological innovation than small and medium organizations.
In Jordan, educational programs come in two varieties: national and international. According to the findings of this study, there is a significant difference between them attributable to PU, such that the mean scores for the international program are greater than those for the national program. As a result, teachers' PU varies from program to program; this finding supports H13.
According to the findings of this study, there is a significant difference attributable to work sector, supporting H14. Furthermore, the Tukey post-hoc test revealed statistically significant differences between the private and government schools. As a result, the findings agree with the conclusions of [12,16,18,38].
This study found no significant difference in favor of educational stage. Hence, the finding indicates that different education stages does not differ when considering H15.
Contrary to the findings of [20,24,[27][28][29][30], when ANOVA was used to test H16, no significant difference related to the number of students was found.
This study failed to corroborate the impact of teacher's age reported by [76,77] and denied by [80]. Using the ANOVA test, this study discovered no significant difference by teacher's age. The effect of teacher education level explored in H18, as indicated by [78,79] and contradicted by [80], is also challenged in this study, as no significant difference was found attributable to teacher's education level.
The findings of [80] about teacher's gender was consistent with the findings of this study explored in H19. ANOVA found no significant difference in favor of gender. This is in conflict with the assertions of [76,77].
The findings of [40,81,82] about personal income are consistent with the results of this study. H20 is supported, and ANOVA found a significant difference attributable to personal income.

Conclusions and Implications
In conclusion, this study supported H1, H3, H4, H5, H7, H8, H9, and H10, but not H2 or H6, implying that the independent variables RA, CP, TM, CT, CM, TC, IN, and WF have a positive effect on teachers' PU of adopting e-learning systems, but not CX or CL.
The study also found that H12, H13, and H14, regarding the size of the school, the education program, and the work sector, respectively, were supported and had a positive effect on teachers' PU of adopting e-learning systems during the COVID-19 pandemic. H15 and H16, which pertain to the educational level and number of students, were not supported and thus had no bearing on teachers' PU of adopting e-learning systems during the COVID-19 pandemic.
Furthermore, H17, H18, and H19, referring to the teachers' age, education level, and gender, were not supported by the study; however, H20, pertaining to personal income, was supported. Age, education level, gender, and personal income are the independent factors relevant to the teacher. The following school-related independent variables are school size, education program, work sector, educational stage, and number of students. Furthermore, the dependent variable PE and the mediating variable PU.
The study also corroborated H11, which posited that PU had a significant positive effect on using and accepting e-learning systems and students' satisfaction with e-learning system courses.

Theoretical Implications
The major contribution of this study is a comprehensive model that allowed the measurement of PU among teachers, thus, measuring the PE of adopting e-learning systems during the COVID-19 pandemic. The model was based on the works of [1][2][3]. The second contribution is that the model included 19 independent variables and 1 intermediate variable. The independent variables pertaining to the teacher are age, education level, gender, and personal income. The independent variables pertaining to school are size, education program, work sector, educational stage, and number of students. Furthermore, the mediating variable is PU, and the dependent variable is PE. Therefore, the model reflected most aspects of e-learning. All factors were valid and important measures that contribute to PU and thence to PE on adopting e-learning systems during COVID-19 pandemic. The third contribution is that the variables of the study have been empirically tested. Although some of the variables were tested in previous studies, to the best of our knowledge, this study is the only one that has tested all the variables in the manner reported.
The fourth contribution of this study pertains to education sector management. The research sheds light on harnessing e-learning tools to benefit students by taking advantage of the teacher's view of the technology. As shown in Table 5, work WF is one of the principal factors that influence the teacher's PU, and thus PE.
Thus, this study makes an important theoretical contribution to the field of education technology in the arena of e-learning by measuring the PU of e-learning from teachers' perspective during the first wave of COVID-19 in Jordan. Furthermore, the contribution enriches the models suggested by [1][2][3]. As such, the symmetrical and asymmetrical deliberation of e-learning during the COVID-19 pandemic is contemplated in this research.

Practical Implications
Given that Jordan, like the rest of the world, shifted to e-learning during the first wave of COVID-19, and considering that teachers are the most essential element in the education process, this study is extremely significant. It sheds light on many important and comprehensive factors that influence the PU of e-learning systems and, consequently, PE. The study's results may help the education system (schools, institutes, universities, management, the Ministry of Education, and teachers) improve the education process. The practical contributions of the study are:

1.
Despite Jordan's adoption of e-learning during the first wave of COVID-19, teachers turned to the most widely available tool, WhatsApp (more than 41%), indicating a need for training on collaboration systems, such as Microsoft Teams and the Darsak platform. Proper tool introduction and training are critical.

2.
Providing teachers with appropriate tools (computers, iPads, smartphones), technology (collaboration system software), and communication methods (Internet services) is critical. Since teachers have been overwhelmed by the sudden demand for e-learning.

3.
International programs and the work sector (private education) differed significantly from national programs and governmental sector education. Thus, the standards of both national programs and governmental sector education must be raised.

4.
The study found a link between using and accepting e-learning and top management support. Thus, the top management of Jordan's education sector must meet the demands for e-learning environment. Furthermore, the top management must be educated in and familiar with e-learning. 5.
The study found that a teacher's personal income has a significant influence on PU. Thus, raising teachers' personal income is recommended. 6.
According to the study, CM is one of the most important elements of PU. Therefore, the demand for e-learning increased, particularly during the exceptional circumstance of COVID-19. This includes the education sector's reaction to the availability of e-learning. In other words, the pandemic produced demand, which may be viewed as an opportunity for the education industry to offer and benefit from e-learning in order to move beyond conventional schooling. 7.
Despite the fact that COVID-19 is a global disaster, new technologies have emerged during the crisis. Many apps that were not designed for e-learning, such as ZOOM and Microsoft Teams, were used for e-learning in Jordan. Furthermore, WhatsApp was utilized to create education groups between teachers and students. Jordan's education sector may design its own education software to meet its unique requirements. 8.
As diverse e-learning environments and software are utilized and spread, the Jordanian education system may learn from other international standards and evolve accordingly. Furthermore, the possibility of adequate learning is being extended to rural regions, i.e., a student in a rural area can benefit from proper education offered in better schools. 9.
The education industry can benefit by instilling competition among instructors in the production of high-quality instructional materials, allowing students to be provided with the highest quality knowledge and study materials. 10. Using technology, teachers may share information and study materials as well as learn from one another about the delivery of educational materials, teaching strategies, and examples. 11. During the first wave of COVID-19, monitoring and evaluation tools as well as quality assurance systems were inadequate. Such tools and procedures may be demanded by the education sector and even developed in Jordan.

Limitations and Recommendations for Future Studies
The current study was conducted with teachers from Jordan affiliated with the Ministry of Education, private schools, and the UNRWA. Similar studies can be extended to other Arabic-speaking countries, and comparative studies must be conducted to further enhance the knowledge in this area.
In addition, the study may be extended to reflect the views of students and education management to further explain the results and provide a more comprehensive view. Such research can be conducted during the second and third waves of COVID-19. Further studies can be conducted to analyze the psychological factors pertaining to e-learning among teachers, students, and guardians.
Another future study may include designing and developing e-learning environment according to international standard and educational systems specifically for Jordan's curriculum. The e-learning environment should include monitoring and evaluation tools, quality assurance tools and techniques development.