Paper— An Exploration of Factors Influencing Self-Efficacy in Online Learning : A Systematic Review An Exploration of Factors Influencing Self-Efficacy in Online Learning : A Systematic Review

— This review examines 12 years of research by focusing on the following question: What are the factors that influence self-efficacy in an online learning environment? There has been a plethora of research concerning self-efficacy. However, few works have focused on the sources of self-efficacy in online-learning environments. Systematic searches of numerous online databases published between 2005 and 2017, which covered factors influencing self-efficacy in online learning context, resulted in the investigation of 25 studies. The data were extracted, organized and analyzed using a narrative synthesis. Results revealed that various factors improved self-efficacy and provided evidence of significant sources of self-efficacy in the context of online learning. Moreover, the investigation provides guidance for further research in designing online learning environments to enhance the self-efficacy of learners.


Introduction
Technological advances and easier access to the Internet have led to an increase in online learning compared with traditional learning environments. Online learning offers learning experiences with technology, which provides accessibility, connectivity, flexibility, and ability to promote interactions among learners. As the number of online-learning users continues to increase, there is a need to understand how students can best apply learning strategies to achieve academic success within the online environment.
Self-efficacy is the belief in one's capabilities to organize and execute the requisite actions required to produce particular results [1]. Beliefs about self-efficacy determine level of motivation as reflected in the amount of effort exerted in an endeavor and the length of time devoted to a challenging situation [2]. If persons have a low level of self-efficacy toward a task, they are less likely to exert effort and accomplish the task. Research findings have demonstrated that self-efficacy is a better predictor of academic achievement than other cognitive or affective processes [3]. Therefore, selfefficacy is critical to learning and performance [4]. Student self-efficacy seems particularly important in challenging learning environments, such as an online learn-ing one where students lack opportunities to interact with others and as a result can become socially isolated [5] , [6]. Also, the drop-out rate among students in online learning environments is higher than that in traditional learning environments [7]. Drop-out rate also related with a lack of self-efficacy [8]. Understanding self-efficacy in online learning is critical to improving online education, which can be a key component of academic success in distance education [4]. However, the focuses of the previous studies were mostly on the situation of self-efficacy in online learning. There have been very few works analyzing factors effecting self-efficacy. As a consequence, the objective of the current review is to examine systematically factors that contribute to self-efficacy in the online learning environment, and which have not previously appeared in open literature.

Systematic review method
A systematic review was based on PRISMA guidelines [9]. These structures are the guidelines on the systematic review to compare all the data that matches preset criteria to answer specific research questions: What are the factors that influence selfefficacy in an online-learning environment?

Search strategy
An extensive search strategy of the ERIC, Scopus, and Web of Science online databases was conducted and separated into two key search terms. The strategy search terms are shown in Fig. 1. Factor* OR Influence* OR Effect* OR Affect* OR Role* OR Effect* AND "Self-efficacy" AND "Online learning" OR "e-learning" OR "Distance learning" OR "Mobile learning" OR "Web-based learning" Correlat* OR Predict* OR Relat* AND "Self-efficacy" AND "Online learning" OR "e-learning" OR "Distance learning" OR "Mobile learning" OR "Web-based learning"

Inclusion and Exclusion criteria
Papers for inclusion in the review were limited to publication in the English language between 2005 and 2017. The final search was conducted in September 2017. The study collected only the research concerned with factors that effect self-efficacy in online learning. One reviewer screened titles and abstracts of studies for first selection. After that, all reviewers examined the remaining full texts of studies to determine eligibility for inclusion in the review. Disagreements between reviewers were resolved through discussion of the degree to which articles met exclusion criteria.

Search outcomes
A total of 25 studies were identified for inclusion in the review. Data from the search strategy of online databases yielded 2174 results. After the removal of duplicates, the remaining 1513 records were assessed based on titles and abstracts. Moreover, 1090 records were excluded from reviewing the titles and abstracts because these studies did not meet inclusion criteria. The full texts of the remaining 69 studies were examined, and 25 were considered relevant. The process used to reduce and evaluate the records is illustrated in the PRISMA flow diagram as displayed in Fig. 2.

Data Extraction and Synthesis
A data extraction table was developed to enable collection of information relevant to the review. All data were collated and manually synthesized. Information extracted from each included a study of sample characteristics (sample size, mean age, gender, and researched location), study design, factor measures, self-efficacy measures, and relevant findings. In addition, a narrative summary was provided.

Results and Discussion
The results of the review are presented to explore factors influencing self-efficacy in online learning. Summary of included studies within the systematic review can be seen in Table 1.
-The effect of peer support on internet self-efficacy was significant (# = .36, p < .01).
-Community identity and interpersonal trust influenced knowledge-sharing behavior through the mediation of knowledge sharing self-efficacy. -Perceived responsiveness was observed to influence selfefficacy significantly (path coefficient = 0.20, p < 0.01) -The results revealed the positive influence of the psychological-safety communication climate on self-efficacy (path coefficient = 0.30, p < 0.01) -Self-efficacy partially mediated the relationship between the psychological-safety communication climate and the intention to continue participation (Sobel statistics = 2.07, p = .038). -Social factor correlated with basic internet self-efficacy (r = .29, p < .001) and advanced internet self-efficacy (r = .37, p < .001) -The social factor had positive effects on basic internet selfefficacy (# = .37, p < .001) and advanced internet self-efficacy (# = .29, p < .001) -Social factor played an indirect role in nurses' intention to use web-based continued learning through basic internet selfefficacy. Tang -Class satisfaction -Academic achievement -Learner-learner interaction (r = 0.28) and system quality (r = 0.18) was related to learner computer self-efficacy -Learner-learner interaction (r = 0.59) and content quality (r = 0.45) was related to learner academic self-efficacy -Learner-learner interaction had a significant effect on both computer self-efficacy (! = .28, p < .001) and academic selfefficacy (! = .56, p < .001) -Content quality significantly predicted computer self-efficacy (! = .23, p < .001).
-Self-regulation in interaction between student and content (! = .37, p < .001) had a positive effect on self-efficacy for learning.
-Self-regulation in interaction between student and teacher in online courses (! = .30, p < .001) positively affected self-efficacy for learning. Hong, et al.

Dependent variable(s):
Behavioral intention to use e-learning systems.
mediated the effects of personal experiences and innovativeness in ICT on performance expectation in the case of instructors, and it partially mediated their effect on expectation in the case of learners.

Factors influencing self-efficacy in online learning
The focus of the research question is on the factors that influence self-efficacy in the online learning environment. Self-efficacy perceptions can and do change as a result of environmental, cognitive, and behavioural effects that a person experiences in the course of everyday life [1] , [3]. This study's findings define specific factors that literature reported as having a perceived effect on self-efficacy in the online learning environment. The result of Bates and Khasawneh [21] reported that previous online learning, instructor-acquired skill, instructor feedback, and online-learning system anxiety influenced students' self-efficacy in the context of online learning. These factors are consistent with the sources of self-efficacy introduced by Bandura [1] which states that self-efficacy expectations are based on four major sources of information: enactive mastery experience, vicarious experience, verbal persuasion as well as physiological and affective states. Findings are described on this topic as a set of categories which follow.
Online Learning Experience and Knowledge. Eight studies showed strong agreement on the effect of online learning experience and knowledge on self-efficacy. Choi, et al. [24] revealed that flow experience has a direct and indirect effect via attitude towards e-learning on technology self-efficacy in Enterprise Resource Planning training with a web-based e-learning (ERP) system usage. In a series of studies, Jashapara and Tai [10] , [54] demonstrated that computer experience influenced elearning system self-efficacy. Moreover, these findings suggested that personal innovativeness with information technology (IT) and computer playfulness also influenced e-learning system self-efficacy. Kim and Park [112] investigated factors influencing an individual's behavior to use e-learning through social-cognitive theory by examining the adoption of e-learning by instructors and learners. The results showed that computer experience significantly effected computer self-efficacy for learners. Me-anwhile, the personal innovativeness in the domain of information and communications technology (ICT) and computer experience was also identified as an important factor influencing computer self-efficacy for instructors. Prior, et al. [92] suggested that attitude and digital literacy has a significant positive effect on self-efficacy. Shen, et al. [61] explored the dimensions of online learning self-efficacy. The result demonstrated that online experience measured with the number of online courses was a significant predictor for self-efficacy to complete an online course. Song, Kalet and Plass [113] also examined the effects of medical clerkship students' prior knowledge, self-regulation, and motivation on learning performance in complex multimedia learning environments. The results showed that students with higher prior knowledge about a carotid disease case tended to report higher self-efficacy. Tang, Tang and Chiang [69] proposed an extended expectation-confirmation model (ECM) that explicitly incorporated experiential learning, perceived self-efficacy, and perceived usefulness to examine blog-continuance learning behavior intentions. The results demonstrated that blog learners' confirmation levels affected various learning beliefs, where the effect of perceived self-efficacy was the largest. Enactive mastery experiences are the most influential source of efficacy information because they provide the most authentic evidence of whether one can muster whatever it takes to succeed [1]. Contrarily, the experience of failure will result in recognition of self-efficacy, which leads to a lack of an attempt to complete tasks.
Feedback and Reward. Two studies reported the positive effect on self-efficacy when feedback and reward were presented. The finding from Liou, et al. [88] indicated that members of the Yamol online-test community improved knowledge sharing self-efficacy if they anticipated extrinsic rewards. Wang and Wu [30] suggested that students who received more elaborate feedback significantly increased their selfefficacy. The benefits of feedback and reward are the opportunity to discover whether they achieve their goals in learning.
Online Communication and Interactions. Six studies showed a strong agreement on the effect of online communication and interactions on self-efficacy. Cho and Cho [108] found that online-learner interaction with learner, content, and teacher are likely to demonstrate higher self-efficacy for learning and satisfaction with the course. Lim, et al. [80] also found the effect of learner-learner interaction on the computer and academic self-efficacy. Meanwhile, academic self-efficacy and computer self-efficacy were affected by content quality and system quality. Based on the community of inquiry framework, Lin et al. [76] investigated the relationship among forms of presence, self-efficacy, and training. The results showed that the teaching presence has a positive prediction on social presence, self-efficacy, and cognitive presence. Moreover, self-efficacy is a full mediator between social presence and cognitive presence. Tseng and Kuo [40] showed influences of community identity and interpersonal trust on knowledge-sharing behaviour through the mediation of social awareness and knowledge-sharing self-efficacy. Reychav, et al. [94] investigated the effect of social network on mobile collaboration with a focus on two aspects of social network mechanism, namely eigenvector centrality and network reciprocity. The results indicated that the network reciprocity formed through peer interactions between users in their daily lives can be leveraged when mobile devices are used in collaborative work.
Shen [77] explored the impact of social interaction, perceptual learning, trust, a sense of community and self-efficacy for knowledge sharing among members in community. The empirical results showed that trust between members and perceptual learning has a significant effect to self-efficacy of knowledge sharing in virtual learning community. Vayre and Vonthron [103] reported that the sense of community plays an important role regarding self-efficacy in online education. Zang, et al. [56] reported effects of two environmental-communication factors, namely, psychological safetycommunication climate and perceived responsiveness on self-efficacy. The online communication and interaction not only allow learners to express themselves but also increase opportunities for learners to receive recognition of successful from each other. Online learning does not readily foster opportunities for observing peer success. Vicarious experience refers to one's observation of a role model performing a task successfully. Verbal persuasion can lead to higher self-efficacy by encouragements from others. Therefore, self-efficacy would be reduced if the learners fail to communicate and meet the performance of others. Verbal persuasion has limitations but can be powerful in conjunction with the role models of the individuals. One possibility for addressing the vicarious experience and verbal persuasion in online learning is for users to encourage communication and to share their successful experiences.
Social Influence. Three studies investigated the effect of social influence on selfefficacy. Social factor is defined as an individual's internalization of the reference group's subjective culture, and specific interpersonal agreements that the individual has made with others, in specific social situations [114]. Chiu and Tsai [66] revealed that the facilitating factor of social contexts in the workplace is an influential way of raising nurses' internet self-efficacy. In addition, the social factor plays an indirect role in the nurses' intentions to use web-based continuing learning through basic internet self-efficacy. Chu and Chu [35] proposed the role of collectivism and group potency at group level in predicting individual internet self-efficacy and individual elearning outcomes for people older than forty-five. The results showed that internet self-efficacy fully mediates the relationship between peer support and learners' persistence in e-learning. In addition, collectivism also moderates the relationship between peer support and internet self-efficacy. Chu [32] further indicated that family support had a most significant role in predicting the effects of e-learning, mediated by general and communication internet self-efficacy. In the gender model, men generally relied more on emotional support to enhance their communication-internet self-efficacy, whereas women showed more reliance on tangible support to increase their communication via the Internet. Social support is an important resource that can help individuals improve self-efficacy and handle stress. The last source of information is the direct effect physiological states can have on learners' self-efficacy. When people judge stress and anxiety, they depend on their state of physiological arousal. It is very likely that individuals will succeed if they are not in a state of aversive arousal [1]. In Chiu and Tsai [66] study, a head nurse or co-worker who is successful in utilizing online learning can serve as a role model for nursing staff.
Learner Motivation and Attitude. Three studies indicated that learner motivation and attitude was the main factor affecting the self-efficacy of the online learner. Motivation can be defined as the extent to which persistent effort is directed toward a goal [115]. Motivation can be determined intrinsically by individuals and externally by sources due to situational variables and environmental factors [116]. Hong, et al. [111] proposed that intrinsic motivation of Chinese learning could positively predict online learning self-efficacy. Law, et al. [39] reported that three motivating factors, namely, individual attitude and expectation, challenging goals, and social pressure and competition had a significant and positive relationship with self-efficacy. The ttest was used to compare the mean scores of constructs between male and female students. Male students are apparently more motivated by challenges, and they also showed a higher level of self-efficacy than female students. Wang, et al. [62] suggested that the level of motivation directly influenced the level of technology selfefficacy. Self-efficacy and motivation have a complex interrelationship. It is likely that each influences or supports the other. However, motivation may be strong enough to overcome a weaker sense of self-efficacy.

Limitation
The main limitations are the fact that only published papers written in English between 2005 and 2017 were included in the review process. Most of the selected studies applied survey design. More rigorous research design, such as incorporating a comparison group, is needed to conclude that the reported literature conclusively had an effect on self-efficacy in online learning.

Conclusion
Self-efficacy is the key to success in all activities including online learning. Hence, the understanding of the source of self-efficacy in online learning context is important. As found in this systematic review, many researchers focused on the investigation of various factors that influenced learner self-efficacy in online learning context. These various factors were source of self-efficacy in online learning context as follows: online learning experience and knowledge, feedback and reward, online communication and interactions, social influence, and learner motivation and attitude. Moreover, the results of this review can be guidance in further research for design online learning to enhance self-efficacy of learner.