Repositório ISCTE-IUL

E-learning is a web-based learning ecosystem for the dissemination of information, communication, and knowledge for education and training. Understanding the impact of e-learning on society, as well as its benefits, is important to link e-learning systems to their success drivers. The aim of this study is to find the determinants of user perceived satisfaction, use, and individual impact of e-learning. This study proposes a theoretical model integrating theories of information systems’ satisfaction and success in the e-learning systems. The model was empirically validated in higher education institutions and university centers in Brazil through a quantitative method of structural equation modeling. Collaboration quality, information quality, and user perceived satisfaction explain e-learning use. The drivers of user perceived satisfaction are information quality, system quality, instructor attitude toward e-learning, diversity in assessment, and learner perceived interaction with others. System quality, use, and user perceived satisfaction explain individual impact.

to measure and evaluate the success of e-learning systems (Holsapple & Lee-Post, 2006;Lin, 2007;. As far as we know however, one of the first studies conducted to understand and modulate the e-learning Brazilian reality was Machado-Da-Silva, Meirelles, Filenga, & Filho (2014). In that study the authors found that information quality, service quality, and system quality had direct impact on elearning systems use and satisfaction. Even so, these authors point out that future studies should be conducted, mainly to understand the perceived impacts of e-learning systems in Brazil, such as net benefits, which result from the individual and organizational impacts. Other studies referring to the use of e-learning systems in Brazil are scarce. Some authors who study e-learning in Brazil study learners' satisfaction and use, not measuring the individual performance (individual impacts) (Dias, 2008;George et al., 2014;Machado-Da-Silva et al., 2014;Moreno Jr. & Zaroni, 2015).
The purpose of this article is to achieve a better understanding of satisfaction, use, and success of e-learning in the Brazilian context. Several surveys of e-learning have been conducted, but no study makes use of DeLone & McLean (1992), but instead integrate models of Sun, Tsai, Finger, Chen, & Yeh (2008), and Urbach, Smolnik & Riempp (2010a), and consider that further studies are needed to better understand the reality (Al-Samarraie, Teng, Alzahrani, & Alalwan, 2017). As main contributions of this study, we outline the integration of information systems success theory D&M with e-learning satisfaction theory. Another feature of the study is that it was conducted in several organizational environments of a developing country, in which technology may help to decrease the educational, digital, and geographical divide.

Theoretical Foundation
E-learning is the use of information technology to disseminate information and knowledge for education and training; e-learning emerges as a paradigm of modern education. E-learning comprises the use of the web to access information and knowledge, disregarding time and space (Aparicio, Bacao, & Oliveira, 2014b, 2014a. E-learning is changing in the way it is used through several devices, according to Liu & Hwang (2010). Access to courses through computer networks (e-learning), mobile devices, wireless communications (m-learning) (Amasha & AbdElrazek, 2016), the mobile sensor technologies, and wireless communications are changing the e-learning paradigm. A new system architecture of the learning environment is in progress: contextawareness and ubiquitous learning (u-learning).

E-learning systems studies
The use of virtual learning environments in addition to classroom study (blended learning), were surveyed by Stricker, Weibel, & Wissmath, (2011). These authors compared two groups of students: a group with the support of virtual learning environment (VLE), and a group without contact with the VLE. The students' performance of the VLE support had better results than those having only face to face learning. The research of Sorgenfrei, Borschbach, & Smolnik (2013) points out three major drivers that guide the process of education through e-learning tools: technical and design size, individual motivation, and environment characteristics. According to the authors, each of these drivers will affect the intention of students to take other e-learning courses.
E-learning acceptance predictors were studied by Cheng (2011), who concluded that perceived usefulness, ease of use, perceived enjoyment, network externality factor, system factor, individual factors, and social factors are the main determinants of acceptance of the e-learning systems.
Recent studies found that the encouragement of a higher social ability affects positively the intention to continue using e-learning, by motivating a greater use of communication tools during courses, allowing learners to increase social participation among students (Brahmasrene & Lee, 2012).
Appendix A outlines some of the satisfaction and e-learning success studies. Satisfaction factors identified by Frankola (2001) explain the low rates of satisfaction with the learning: students do M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 5 not have enough time, there is failure in supervision or management of e-learning structure, the lack of motivation, problems with the technology chosen, erratic support to the student, preference for traditional learning, poor graphic design adopted by the platform, and instructor deficiency (lack of knowledge and/or ability to deliver). A study on student satisfaction of e-learning (Wang, 2003) highlighted important determinants such as student interface, learning community, content, and customization. Selim (2003) found that the ease of use of web courses is the main determinant of acceptance as an effective and efficient technology for learning. Selim (2007) described the critical success factors of e-learning as grouped into four categories: trainer, student, information technology, and university support. From literature it is known that students' dimension, teachers, courses, technology, design, and environment determine e-learning success (Sun et al., 2008).
Research shows that attitude of students and instructors toward e-learning can determine success (Liaw, Huang, & Chen, 2007). Their study posed a three-tier technology use model (3-TUM). Figure 1 depicts the evolution of e-learning systems based on the studies in Appendix A. In general, earlier studies focused more on technology itself and on content, but the latest studies reflect that students' attitude and interaction also play an important role in e-learning success.

Information System (IS) success
Information systems success has been studied regarding the End-User Computing Satisfaction (EUCS) developed by Doll & Torkzadeh (1988). DeLone & McLean (1992) proposed one of the most tested IS success models. The 1992 D&M model is composed of six theoretical constructs: system quality, information quality, use, user satisfaction, individual impact, and organizational impact. Afterwards, in 2003, the D&M success model was updated and found theoretical evidence that service quality is also a success determinant regarding positive influence on the use and user satisfaction (DeLone & McLean, 2003). In this model a new construct was included, net benefits, a result of merging the individual impacts and organizational impacts constructs. According to these authors (DeLone & McLean, 2003), system quality corresponds to the technological characteristics, performance, and usability of the system itself. Information quality corresponds to the system's accuracy, validity, and currency regarding the system contents. Service quality relates to the responsiveness and perceived competence of the technological staff. Use is one of the literature success measures, and corresponds to the effective use of a system, therefore full adoption, the first phase of success. User satisfaction is the perceived level of agreeableness toward the entire system. It is measured by the appropriateness and effectiveness. Net benefits are the perceived individual and organizational impacts on tasks/job performance and efficiency.

Theoretical model
This study modulates the e-learning success in the Brazilian context. The proposed model is based on the previous theory of e-learning satisfaction and IS success theory. The proposed research model, Figure 2, integrates two theories, e-learning satisfaction and IS success (DeLone & McLean, 2003;Sun et al., 2008;Urbach, Smolnik, & Riempp, 2010a). These theories have been validated by several empirical studies and are therefore models with solid foundations (Bento, Costa, & Aparicio, 2017).
Collaboration quality corresponds to the web environment features, digital culture and the universal use of the web on various platforms such as smartphones, tablets, and computers (Benbya, Passiante, & Belbaly, 2004;Detlor, 2000;Urbach et al., 2010a;Wang, 2003). Service quality is the e-learning system requirements for efficient service support, which can be measured by points responsiveness, empathy, trust, and security (DeLone &McLean, 2003;Urbach et al., 2010a). Information quality: for the quality of information of an e-learning system some items are needed, such as applicability, comprehensiveness, and reliability. System quality of an e-learning system comprehends functionality, usability, navigability, and the accessibility that users perceive from the usage of an e-learning platform during the course. Learner computer anxiety: anxiety is an internal personal characteristic, stable and durable, as a result of the external environment (Spielberger & Anton, 1976). Instructor attitude toward e-learning is measured by the student perception of the usefulness pointed out by the teacher during the course, compared to face to face learning (Sun et al., 2008). Diversity in assessment is the presence of various assessment methods in the course. Learner-perceived interactions with others comprehend three types of interaction, students with teachers, students with course materials, and students with students (Moore, 1989). User satisfaction is one success measure of the overall level of fulfilment of learners' expectations (Sun et al., 2008). Use measures the actual use of e-learning system by the students to perform their learning tasks, for example, retrieve and publish information and communicate with others.
The individual impact is the degree of benefit perceived by students when using an e-learning system.

Hypotheses
Based on the findings of Urbach et al. (2010a), collaboration quality emerged as a significant determinant on the system usage and also on the user satisfaction. Thus, it creates possibilities for M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 8 co-creation, communities of practice, and collaborative knowledge (Benbya et al., 2004;Detlor, 2000;Urbach et al., 2010a;Wang, 2003). The efficiency of different collaborative features, ease of use, efficiency and comfortability for collaboration, facilitating communication and information sharing on multi platforms (e.g.: LMS, networks, and social media), are essential for supporting collaborating tasks. Therefore, the current research hypothesizes that: Service quality of e-learning systems needs responsiveness, empathy, trust, and security of the supporting staff. According to earlier studies, service quality is essential to satisfaction and use (Chang & King, 2005;Pitt, Watson, & Kavan, 1995;Uppal, Gulliver, & Ali, 2017), and in elearning context service quality impacts positively e-learning usage and students' satisfaction (Aparicio, Bacao, & Oliveira, 2017;Machado-Da-Silva et al., 2014). Our proposal is that service quality has an impact not only on use and on satisfaction, but also on individual performance. Therefore, the current research hypothesizes that: Rich content provides quality of the information regarding its usefulness, understandability, and reliability (DeLone & McLean, 1992). Several studies have found that information quality has a positive impact on the use and satisfaction Machado-Da-Silva et al., 2014;McKinney, Yoon, & Zahedi, 2002;Urbach et al., 2010a;Yang, Cai, Zhou, & Zhou, 2005).
Information quality can also have a direct impact on individual performance (DeLone & McLean, 2002). Therefore, the current research hypothesizes that: System quality of an e-learning system is critical to a good user experience of e-learning (Ahn, Ryu, & Han, 2004). It is also identified as having an impact on performance characteristics, functionality, and usability, among others (McKinney et al., 2002). System quality is the level of ease of use and carrying out of tasks (Elkaseh, Wong, & Fung, 2016;Schaupp, Fan, & Belanger, 2006). Studies by Urbach et al. (2010a) also demonstrate the importance of navigability, accessibility, structure, visual logic, and stability of e-learning systems to ensure a good user experience and learning (Butzke & Alberton, 2017;Tarhini, Hone, Liu, & Tarhini, 2017). Studies demonstrate that system quality has a positive impact on use and satisfaction Urbach et al., 2010a). DeLone & McLean (2002) hypothesize that system quality has a direct and positive impact on individual performance. Therefore, the current research hypothesizes that: (H4a). System Quality has a positive influence on the Individual Impact of e-learning systems.

(H4b). System Quality has a positive influence on the Use of e-learning systems. (H4c). System Quality has a positive influence on the User Satisfaction of e-learning systems.
Students' anxiety toward computers is different from their attitude. Computer anxiety represents beliefs and feelings about computers (Heinssen, Glass, & Knight, 1987). However, learners' anxiety has a negative impact on satisfaction, preventing e-learning success (Sun et al., 2008).
These authors found that the greater is the anxiety, the smaller will be the task performance (Kanfer & Heggestad, 1997). The attitudes of individuals well-adjusted to technology would be more positive and will lower anxiety levels (Barbeite & Weiss, 2004). Therefore, the current research hypothesizes that:

(H5). The Learner Computer Anxiety has a negative influence on the User Perceived Satisfaction of e-learning user.
Instructor attitude toward e-learning corresponds to teachers' reactions about students' problems (Soon, Sook, Jung, & Im, 2000). In an online course, instructor assistance encourages students to continue their studies. Consequently, if a teacher can handle the e-learning activities and responds to students' needs and problems promptly, the satisfaction of learning will improve (Levy & Ramim, 2017). Studies support the impact of this variable on satisfaction (Webster & Hackley, 1997;Sun et al., 2008). Therefore, the current research hypothesizes that:

(H6). The instructor attitude toward e-Learning influences User Perceived Satisfaction of the elearning user.
Diversity in the assessment are the appropriate feedback evaluation methods and mechanisms.
These mechanisms are essential for e-learning users, allowing the follow-up of learning.
According to some studies this considerably influences students' satisfaction (Thurmond, Wambach, Connors, & Frey, 2002;Sun et al., 2008). Therefore, if an e-learning system provides more assessment tools and diverse methods, the level of user satisfaction will be greater. As a consequence, the current research hypothesizes that:

(H7). The Diversity In Assessment has a positive influence on the User Perceived Satisfaction of elearning user.
In a virtual learning environment the electronic interactions between students or between students and course materials can help to solve problems and improve the frequency and quality of the learning process. Arbaugh's (2002) study suggests that when students realize greater interaction with others, there is an increase of user satisfaction. Therefore, the current research hypothesizes that:

(H8). The Learner Perceived Interaction with Others has a positive influence on the User
Perceived Satisfaction of e-learning user.
User perceived satisfaction toward an information system influences the actual usage of a system (DeLone & McLean, 2003). This relationship between satisfaction and use of an information system is sustained by several studies (Costa, Ferreira, Bento, & Aparicio, 2016;Seddon, 1997;Sun et al., 2008;Urbach et al., 2010a). Some studies on e-learning success also support that the more satisfied the students are the more they will use e-learning systems (Aparicio, Bacao, & Oliveira, 2016b;Aparicio et al., 2017;Wang & Chiu, 2011). Therefore, the current research hypothesizes that: (H9). The User Perceived Satisfaction has a positive influence on the Use of e-learning user.
Information systems usage is positively related to individual performance DeLone & McLean, 2002;Tam & Oliveira, 2016;Urbach et al., 2010a). If e-learners' usage perception is aligned with their needs, students can accomplish their tasks in a more effective way.
The more students use e-learning systems, the more they perceive positive individual impacts . Therefore, the current research hypothesizes that: (H10). The Use has a positive influence on the Individual Impact of e-learning user.
The greater the user satisfaction, the greater the individual impact (DeLone & McLean, 1992;Tam & Oliveira, 2016;Urbach et al., 2010a). Studies on e-learning success report that user satisfaction has a significant impact on value (Alsabawy, Cater-Steel, & Soar, 2011), and positive impact on individual performance Piccoli, Ahmad, & Ives, 2001). Therefore, the current research hypothesizes that:   2003). For this reason we included another theory validated for e-learning systems (Sun et al., 2008), as other dimensions, such as learners' attitude toward technology, instructor attitude, assessment, and interaction between learners' satisfaction of e-learning. We included individual aspects of learners to better understand their role in success. We also included collaboration quality in the model (Urbach et al., 2010a) because on-line students tend to be in different physical locations, and that can affect their individual performance.

Methodological approach
As the theoretical model has qualitative approaches to generalization, the quantitative method was chosen, as deemed more suitable (Johnson & Duberley, 2013). The survey collected 301 valid responses. An online survey was conducted with students (undergraduate and graduate) to collect data for the empirical evaluation of our theoretical model. The research design is in line with most studies and best practices.

Construct Operationalization
The research model constructs' operationalization is composed of tested scales of several studies, of both information systems success theory and e-learning systems theory (Appendix A). Each construct of the conceptual model made use of tested and proven measures in order to increase the validity of the study. The latent variables, collaboration quality (CQ), service quality (SerQ), information quality (IQ), use, user satisfaction (US), and individual impacts (II), are operationalized according to items adaptation of Urbach et al. (2010). Learner computer anxiety (LCA), instructor attitude toward e-learning (IATL), diversity assessment (DA), and learner perceived interaction with others (LPIO) are according to the items of Sun et al. (2008).

Survey environment & data collection strategy
The data collection instrument was initially developed in English, according to validated literature scales (Appendix B). Then, the final version was translated to Portuguese, by a professional translator, and then back into English by a different translator to ensure conversion correspondence (Brislin, 1970). The questionnaire was pre-tested by a group of 20, non-distance learners, to certify that it was well interpreted by university students. The sampling strategy undertaken consisted of directly contacting 24 organizations of higher education institutions all over Brazil, requesting collaboration to disseminate the online survey among their students. A hyperlink was provided by email to the coordinators so they could pass it on to the students, asking them to distribute the email to all students. Course coordinators (graduate and undergraduate) in Brazil were contacted.
Participation was encouraged by offering higher education institutions the survey results. When necessary, telephone contact with the course coordinators (and in some cases with principals) was made to clarify the objectives of the study and discussion of new procedures. The data collection process followed a strict path. The research is approved by the university committee, and in an ethical point of view, the universities involved did not oppose the study. The study is anonymous and private, and all the questions concern the study context. At the beginning of the questionnaire, all the respondents were informed about the academic research purpose of the survey, in which they voluntarily agreed to participate. The questionnaire was totally anonymized, no personal information was asked for from respondents, and no tracking systems were employed. As a delimitation, the study was addressed only to higher education institutions, such as colleges, universities, and university centers, either public or private. To minimize bias and obtain respondents, it was emphasized that all the data would be treated with total confidentiality and that the identity of the respondent could not be inferred.
From January of 2015 to June of 2015, 381 responses to the survey were obtained, although due to incomplete answers only 301 were considered valid and complete for analysis. Students' responses included both the classroom mode with e-learning as support (blended learning) and 100% elearning. Learners answered on a seven-point scale, from 1 point -strongly disagree, to 7 pointsstrongly agree. The questionnaire also included queries about general respondents characteristics, such as gender, age, and which e-learning platform they use and general comments ( Table 1). The survey is balanced in terms of male (50%) and female (50%) respondents. The university students M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
14 are respectively 48% male and 52% female (INEP, 2016;OECD, 2015). Most learners in the sample are at a university level. There is no statistically significant difference (p>0.10) between the gender of our sample and the university student population.

Analysis and Results
For data analysis we used the structural equation model (SEM) method. Using the empirical survey data, the measurement properties were evaluated and the hypotheses were tested using the approach of partial least squares (PLS) (Chin, 1998;Wold, 1985). PLS was chosen for data analysis due to its advantages, even if compared to approaches based on the covariance. When the search model is complex, it has a large number of constructs, and the measures are not well established, (Chin & Newsted, 1999;Fornell & Bookstein, 1982). Furthermore, PLS software may be more suitable because it has less-stringent requirements on the distributions (Fornell & Bookstein, 1982;Gefen, Straub, & Boudreau, 2000). Finally, the approach by the PLS is the most suitable for management problems focused on forecasts (Fornell & Bookstein, 1982;Huber, Herrmann, Meyer, Vogel, & Vollhardt, 2007

Measurement model evaluation
Reflective indicators were used to establish the constructs. Following the validation guidelines proposed by Lewis, Templeton, & Byrd, (2005) and by Straub, Boudreau, & Gefen, (2004), models of reflective measurement for one-dimensionality, internal consistency, indicator reliability, convergent validity, and discriminant validity were tested by applying standard decision rules. The traditional criterion to assess the internal consistency is Cronbach's Alpha (CA). All the CA are above 0.700, indicating internal consistency, and the scores of all items of the constructs have the same scope and meaning as defined by Cronbach (1951). An alternative measure for CA is composite reliability (CR) (Werts, Linn, & Jöreskog, 1974). The CR is recommended by Chin (1998) as the preferred measure, as it overcomes some of the deficiencies of the CA. The CA and CR values of all constructs in our model are, as shown in Table 2, above the minimum recommended 0.700 (Nunnally & Bernstein, 1994). The model measures the CR above 0.800, meeting the criteria established by Peter (1979).
We evaluated the indicator reliability checking the criteria that the loadings should be greater than 0.70 (Henseler, 2010;Henseler, Ringle, & Sinkovics, 2009). As reported in Table 2, loadings are greater than 0.7, except two items (Use1 and Use 5), which are lower than 0.7 but greater than 0.4.
Hence, no items in the table were eliminated. The measuring instrument presented good indicator reliability. Convergent validity relates to the level at which individual items are reflected in the construct and converge compared to items that measure different constructs. A commonly applied convergent validity criterion is the average variance extracted (AVE) proposed by Fornell & Larcker (1981). As shown in Table 2, all model constructs have indicators above 0.500, indicating that the variance of the construct is greater than the variation caused by the respective measurement errors and thus indicating that all constructs have adequate validity (Segars, 1997). root extracted for each construct was examined and found to be higher than the correlation between constructs (Table 3). Conceptually, this test requires that each construct represents more of the variance in its indicators than it shares with other constructs (Fornell & Larcker, 1981). A second criterion was used; we compare the loadings with the cross-loadings. We can see in Appendix C that the loadings (in bold) are greater than respective cross-loadings. Consequently, for both criteria the discriminant validity was achieved.

Assessment of the structural model
After validation of the measurement model, the structural model was studied and the possible relationships between the constructs were tested. The results of the test conducted in the structural model are given in Figure 3 and   (Chin, 1998;Cohen, 1988)

Discussion
Most of the hypothesized relationships were verified. Use is explained by collaboration quality, information quality, and user satisfaction. Users' satisfaction is explained by information quality, e-learning system quality, instructor attitude toward e-learning, diversity in assessment, and learner interaction with others. Individual impacts on e-learning usage are determined by use of elearning systems, user satisfaction, and system quality. Although collaboration quality does not determine user satisfaction or individual impacts, and service quality determines none of the elearning success dimensions, learner computer anxiety was found significant to satisfaction (Table   4).
The study indicates that collaboration quality positively influences e-learning systems' use (H1b), and that collaboration quality of e-learning therefore systems appears to be an important success factor. If available, collaborative features are used by the users, achieving a greater overall satisfaction with e-learning. Therefore, providing additional collaboration capabilities and improving existing ones may directly increase use and user perceived satisfaction, and hence the individual impact. Similar results were found in employee portal usage and e-learning usage (Urbach et al., 2010a;Wang, 2003). Results indicate that service quality (H2) has no significant impact on user satisfaction, use, and individual impact. This finding is consistent with the results reported by other authors (Chiu, Sun, Sun, & Ju, 2007;Choe, 1996;Urbach et al., 2010) Another study also found that the access to resource contents predicted success (Bandeira, dos Santos, Ribeiro, & Neto, 2016). Hypotheses H4a and H4c are validated; system quality is positive and statistically significant on user perceived satisfaction, and on individual impacts. Similar results were found in e-learning studies, and employee portal, in ERP usage satisfaction. In these studies system quality also had a positive impact on user satisfaction, and system quality was not significant in these systems' use Costa et al., 2016;McGill & Klobas, 2005;Urbach et al., 2010 a). Machado-da-Silva et al. (2014) found no significant impact of system quality on use, or on satisfaction. The impact of learner computer anxiety on satisfaction (H5) was not found significant, which is at odds with Sun et al.'s (2008) findings. One reason for that is that students today might not feel as anxious toward technology usage. It is likely that the last generations are more familiarized with digital platforms, or because 99% of the respondents had at least a two-year college degree.
Instructor attitude toward e-learning, diversity assessment, and learner perceived interaction with others have a positive impact e-learner satisfaction (H6, H7, and H8). These findings corroborate Sun et al.'s (2008) results. Many assessment methods allow the instructors to establish the effects of learning, and different aspects of education can be more effective. As for the students, diversified rating methods are motivational factors, as evidenced by the efforts of students, engaging them in e-learning activities. Communication functionalities may also allow instructors to engage more students, and students themselves can interact more easily with their peers. Results show evidence that user satisfaction has a positive impact on e-learning systems use (H9). Similar results were reported in other studies (Urbach et al., 2010a;Wu & Wang, 2006). E-learning systems use and user satisfaction have a positive impact on individual performance (H10, H11), and these findings are consistent with various studies Urbach et al., 2010a;Wu & Wang, 2006). The significant impact of user perceived satisfaction on individual impacts supports the suggestion that user perceived satisfaction can serve as a valid substitute for individual impact (Iivari, 2005;Piccoli, Ahmad, & Ives, 2001). Our study demonstrates that M A N U S C R I P T A C C E P T E D ACCEPTED MANUSCRIPT 21 collaboration, and information and system quality are determinant factors of e-learning systems success, and that instructor attitude, diversity in assessment, and learner interaction with others, are also determinants for e-learning success.

Conclusions and implications
This article presents a theoretical background that includes IS success and e-learning systems satisfaction and success. Based upon theory, a model was proposed and validated empirically in Brazilian universities. This study demonstrates that use and user satisfaction are interdependent, and both have a positive impact on individual performance. The hypotheses developed (information quality, system quality, instructor attitude toward e-learning, diversity in assessment, and learner perceived interaction with others) explain the user perceived satisfaction.
Collaboration quality, information quality, and user perceived satisfaction are important drivers for e-learning use. Collaboration quality, service quality, information quality, system quality, user perceived satisfaction, and use explain the individual impact. This Brazilian e-learning success model explains 52% of the variation of individual impacts.
This study presents two theoretical implications, as it contributes to information systems theory.
Our model combines the information systems success theory of DeLone & McLean (1992 with e-learning systems satisfaction theory (Sun et al., 2008) and collaboration quality (Urbach et al., 2010a). As another theoretical contribution, this model validates information systems success theory for the case of e-learning systems usage in the context of Brazil.
The practical implications of this study bring insights to e-learning systems designers and providers. One such implication derived from this study is that e-learning platforms should provide technological features to enable a collaboration environment, an important aspect in e-learning systems success. According to our findings, stakeholders would benefit if considering collaboration modules in the platforms. As an example, technological platforms should allow the articulation of communication and collaboration between students, thereby influencing use and learners' satisfaction. This study also implies that information quality has a significant impact on use and satisfaction, such as course contents. Content should be retrievable, useful, understandable, interesting, and reliable. Institutions should design various ways of self-M A N U S C R I P T

A C C E P T E D ACCEPTED MANUSCRIPT
22 assessment through quizzes, tests, and other ways of testing knowledge. Thus, providers would increase overall success level by investing in the contents of the course. From this study we also learned that if e-learning systems provide a variety of ways for learning assessment, and if learners interact with each other, it will lead to an increase of satisfaction. We found from this study that the perception of individual performance is due to the learners' perceived system quality. If the system is easy to navigate and well-structured in terms of content and functionalities, it will increase satisfaction and usage of e-learning systems.

Limitations and future research
The     (Frankola, 2001) Low satisfaction rates with e-learning . Students do not have enough time . failure in supervision or management of e-learning structure . lack motivation . problems (Wang, 2003) Model for measuring satisfaction of asynchronous e-learning student Highlighted important variables such as: . student interface . learning community . content . customization (Selim, 2003) Evaluate the acceptance of the courses on the web by students, based on the TAM, and formulated the CWAM The ease of use of the courses on the web is the main determinant as to acceptance as an effective and efficient learning technology (Chiu et al., 2007) Model: DEDT usability, quality, value, and the decision to continue the e-learning were highlighted (Lee & Lee, 2005) Study on the success of e-learning system for courses and modules Synthesize the ECM, TAM, the TPB, and the flow theory to the hypothesis of a theoretical model to explain and predict the intentions of users to continue to use e-learning (Roca, Chiu, & Martínez, 2006) The perceived performance component is decomposed in perceived quality and perceived usability. Study based on the EDT, and proposed the DEDT The intention of continuity is determined together: . perceived usefulness . information quality . confirmation . service quality . quality of the system . perceived ease of use . cognitive absorption (Liaw et al., 2007) Study of the students' and instructor's attitudes in connection to e-learning. Developing the 3-TUM Conclusions in levels: . level 1 / layer of individual experience and quality system . level 2 / affective and cognitive layer . level 3 / behavioral intention layer (Wang, Wang, & Shee, 2007) Studies on the success of e-learning system for courses and modules Show the structure of the scale factor, reliability, content validity, criterion validation, data analysis of 206 respondents sample (Selim, 2007) Study of the critical factors of success in e-learning Four categories: .
(3) information technology . (4) support from university (Levy, 2007) Compared the dropouts and persistent e-learning students, and raised two constructs: (1) academic locus of control, and (2) student satisfaction with e-learning The results demonstrated that student satisfaction with e-learning is a key indicator in the decision to abandon the course of e-learning (Shee & Wang, 2008) With the WELS growth, users are recognized as essential as satisfaction influences the adoption of systems Development of methodology based on student satisfaction and their applications in multi-criteria evaluation of web-based e-learning system (Roca & Gagné, 2008) Study for understanding the intention of continuity of e-learning in the workplace were based on SDT, and the study expanded the TAM with the elearning services In the proposed model of perceived usefulness, perceived playfulness and perceived ease of use are expected to be influenced by the perceived autonomy support, perceived competence, and relational perception. The study also helped to examine the effects of motivational factors affecting the TAM constructs (Sun et al., 2008 (Lee, 2010) Evaluated the SAT with the CI, and synthesized ECM, TAM, TPB, and the flow theory to the possibility of a theoretical model to explain and predict the intentions of the users and the continuity use of e-learning New variables that affect the intention of continuity of e-learning users, such as: satisfaction, usefulness, attitude, concentration, subjective norm, and perceived behavioral control (Paechter, Maier, & Macher, 2010) Expectations and experiences of students in e-learning and related learning objectives and satisfaction of the course Showed five important factors in the learning process: .
(1) instructional design, learning materials, and friendly electronic environment .
(2) availability of interaction between students and students with instructors/teachers . (3) possibility of exchanging and sharing among students .
(4) encouragement of individual learning . (5) improvements based on the analysis of the cognitive and emotional outcomes involved in learning (Lin & Bhattacherjee, 2010) Based on the theory of rational action and theory confirmation of expectation generated a model The model features six constructs: .
(3) perceived ease of use . (4) perceived usefulness . (5) attitude . (6) intention to continue (Hassanzadeh, Kanaani, & Elahi, 2012) Broadening and deepening the study of e-learning models, formulated MELSS Include new variables such as quality educational system and compliance objectives (Aggelidis & Chatzoglou, 2012) Study on the success of e-learning system for courses and modules Build on existing body of knowledge, testing past models, and suggesting new conceptual perspectives on how the EUCS is formed between the users of the hospital information system (Chow & Shi, 2014) Study on the level of satisfaction of using e-learning systems Theoretical Support Using a seven-point scale 1, strongly disagree; 7, strongly agree, the variables are to be measured by asking students to rate their perception on elearning systems. Strongly disagree 1 -2 -3 -4 -5 -6 -7 Strongly agree Collaboration Quality

CQ1
Our e-learning system enables an easy and comfortable communication with my colleagues. Urbach et al., 2010 CQ2 Our e-learning system supports an effective and efficient sharing of information with my colleagues.

CQ3
Our e-learning system enables a comfortable storing and sharing of documents with my colleagues.

CQ4
Our e-learning system allows me to easily and quickly locate my colleagues' contact information.
Service Quality

SerQ1
The responsible service personnel are always highly willing to help whenever I need support with the e-learning system.

SerQ2
The responsible service personnel provide personal attention when I experience problems with the e-learning system.

SerQ3
The responsible service personnel provide services related to the e-learning system at the promised time.

SerQ4
The responsible service personnel have sufficient knowledge to answer my questions in respect to the e-learning system.

IQ1
Examples are retrievable documents, course news, process descriptions, and course-specific information.
The information provided by e-learning system is useful.

IQ2
The information provided by e-learning system is understandable.

IQ3
The information provided by e-learning system is interesting.

IQ4
The information provided by e-learning system is reliable.
System Quality SysQ1 Please assess the system quality of the e-learning platform. The e-learning system is easy to navigate. SysQ2 The e-learning system allows me to easily find the information I am looking for.

SysQ3
The e-learning system is well structured.

SysQ4
The e-learning system is easy to use.

Learner
Computer Anxiety

LCA1
Working with a computer would make me very nervous Sun et al., 2008 LCA2 Computers make me feel uncomfortable LCA3 Computers make me feel uneasy and confused Instructor Attitude Toward e-learning

IATL1
Compared to traditional classrooms, you feel that your instructor considers useful a web-based learning technology.

Diversity in Assessment DA1
The e-learning offers a variety of ways to assess my learning (quizzes, written work, oral presentation, etc.) Learner Perceived Interaction with Others LPIO1 I learned more from my fellow students in this e-learning system than in other courses

LPIO2
The instructor frequently attempted to elicit student interaction LPIO3 It was easy to follow class discussions User Perceived Satisfaction

US1
How adequately does the e-learning system support your area of study? Urbach et al., 2010 US2 How efficient is the e-learning system?

US3
How effective is the e-learning system?

US4
Are you satisfied with the e-learning system on the whole?

Use
Use1 Please indicate the extent to which you use the e-learning system to perform the following tasks: Retrieve information. Use2 Publish information.

Use3
Communicate with colleagues and teachers.

Use4
Store and share documents.

Use5
Execute course work.