Australasian Journal of Educational Technology, 2017, 33(1). 114 Learner control, user characteristics, platform difference, and their role in adoption intention for MOOC learning in China Min Zhang, Shuaijun Yin, Meifen Luo, Weiwei Yan Wuhan University, China Massive open online course (MOOC) learning attracts more and more attention in both the practice and the research field. Finding out what factors influence learners’ MOOC adoption is of great importance. This study focuses on learner control, user characteristics and platform difference. Hypotheses and a research model are proposed by incorporating perceived learner control, e-learning self-efficacy, and personal innovativeness in information technology (PIIT) into the original technology acceptance model (TAM). With the empirical data from 214 MOOC learners, the effects of perceived learner control on perceived usefulness and perceived ease of use are confirmed. E-Learning self-efficacy is found to have positive influence on perceived learner control and ease of use. While the effect of PIIT on perceived learner control is not supported, PIIT influences learners’ perception of usefulness and ease of use. Furthermore, a comparison between foreign and native MOOC platforms shows the former should emphasize ease of use more, and the latter should emphasize usefulness more, to enhance their attractiveness. Introduction Massive open online courses (MOOCs), are a development in the area of distance education, and a progression motivated by open education ideals (Yuan & Powell, 2013). Unlike the traditional online teaching and learning systems, which usually allow limited enrolments, MOOCs provide the opportunity for massive participation, the tools of discussion, and the ways of interaction (Pappano, 2012). Thus, students learning in MOOC courses are provided with a greater level of control than they might typically experience in a traditional face-to-face classroom course (Chou & Liu, 2005; Fisher, Wasserman, &Orvis, 2010). Learners now have many MOOC platforms to choose from: Coursera and EdX from North America; Future Learn from England; Iversity from Germany; and XuetangX and ICourse163 from China. In academia MOOCs have also gained wide attention from researchers. A literature review published by Liyanagunawardena, Adams and Williams (2013) divided MOOC research into several research branches including: case studies. introduction of MOOCs, concept of MOOCs, discussion of the education theory, technology details, provider focus, participant focus, and others. Another similar study published by Gasevic, Kovanovic, Joksimovic and Siemens (2014) revealed the main focus for future MOOC research and posited that student motivation, attitude, engagement, and learning success are the foremost current research topics. Extant literature about MOOCs is divided into four categories according to main themes, a) education technology improvement and MOOC platform design, b) learner’s motivation, attitude, and behavioral patterns, c) learner’s performance and learning success, and d) communities and social network learning. Despite the breadth of literature in this area, three critical gaps exist that represent fruitful opportunities for research. First, the reasons why students choose MOOC learning as a new form of e-learning and factors that influence MOOC learners’ adoption intention should be studied further. In the past, researchers concentrated on student retention (Adamopoulos, 2013) and learner intent and behavior (Campbell, Gibbs, Najafi, & Severinski, 2014), but the reason why students adopt or refuse MOOCs is neglected. This is essential for improving education quality. Second, few researchers focus on the effects of learner’s characteristics, especially for learner control, on MOOC acceptance behavior. Research about the relationship between learner control and learning outcomes (Hughes et al., 2013; Karich, Burns, & Maki, 2014), provide a reference for us. Third, comparison research on different MOOC platform characteristics are relatively rare. Foreign platforms usually contain courses taught in English or their native language, and the interface design may be unfamiliar to Chinese students. Website design differences (website layout, spatial features, use of certain colors, preferences for multimedia elements, representation of language and script) existed among different countries (Cyr & Trevor-Smith, 2004). These design differences and language problems may cause different perceptions about MOOC learning and affect students’ MOOC adoption intention. The purpose of this research is to investigate the user characteristics of MOOC learners, Australasian Journal of Educational Technology, 2017, 33(1). 115 their role in the adoption process, and analyze the difference between foreign MOOC platforms and native MOOC platforms. Our research model can be seen in Figure 1. Theoretical background Technology acceptance model Proposed by Davis (1989), the technology acceptance model (TAM) is a process for understanding the acceptance of information technology. TAM suggests perceived usefulness (the degree to which using a particular technology would enhance job performance) and perceived ease of use (the degree to which using the system would be free of effort), are the two main perceptions that influence the attitude, intention, and usage behaviour toward a new technology. TAM is one of the most widely used technology acceptance theories, and also the most commonly used theory in e-learning acceptance studies. Šumak, Heričko and Pušnik (2011) found that 86% of the e-learning acceptance studies reviewed employed TAM as their theoretical basis. Researchers differentiated types of e-learning adoption users into professionals, for example nurses (Cheng, 2014), and employees, teachers, students, and working mothers (Kibelloh & Bao, 2014). In the field of e- learning, TAM could not only be used for predicting adoption behavior, but also for measuring the learning satisfaction, continuous intention to use e-learning (Pereira, Ramos, Gouvêa, & Costa, 2015), loyalty to e- learning (Sánchez-Franco, Peral-Peral, & Villarejo-Ramos, 2014), course outcomes, and academic performance using e-learning (Arbaugh, 2014). The features of technology: interactivity, personalisation, accessibility, mobility, and the choice of media to present the contents (Agudo-Peregrina, Hernández- García, & Pascual-Miguel, 2014); individual characteristics such as personal innovativeness in information technology, computer self-efficacy, and demographic variables (Chow et al., 2013; Thatcher & Perrewe, 2002);the course characteristics of learning resources, course content, tutor quality, and course quality (Persico, Manca, & Pozzi, 2014; Teo, 2014); and other variables such as social influence (or social norm) and flow (Wu & Zhang, 2014), are incorporated into TAM to enhance the understanding of adopting. In summary, TAM has been widely used in predicting e-learning adoption and proved to be robust. Learner control Shyu and Brown (1992) defined learner control as the degree to which students can direct their own learning experiences. Reeves (1993), and Orvis, Fisher and Wasserman (2009)described learner control as an element of learning that allows learners to make important decisions about various features of the learning activity. Although shifting control over the learning process has long been discussed in the pedagogy of teaching and learning, it is not until the information technology and social network become prevalent that the practice of students’ self-controlled learning and personalising learning come into action. The first introduction of learner control in computer-adapted instruction was to explain student performance, and it was assumed that the degree of control the learner was positively related to student performance (Merrill, 1975). A survey by Marshall et al. (2009) found that a great number of teachers thought personalising learning with the help of technology, giving the students more control over the learning process, was very important. Behrend and Thompson (2012) studied the influence of full control and customisation of the designed pedagogical agent (animated characters, in place of the teacher, to inspire motivation and learning in computer-based tutoring systems) in contrast to no choice of designed pedagogical agent and found learner control to be of benefit to learning outcomes. However, research findings seem to be inconsistent when it comes to learning outcomes. Some other studies found no effects on learning outcomes (Lahti, Hätönen, & Välimäki, 2014; Means, Toyama, Murphy, Bakia, & Jones, 2009), or even negative effects of learner control on learning outcomes (Jaggars & Bailey, 2010; Xu & Jaggars, 2014). Despite the inconsistent findings in the relationship between learner control and learning outcomes, leaner control seems to have a positive effect on the perception of the learning process. When conducting a meta- analysis on learner control, Karich, Burns and Maki (2014) identified that learner control within educational technology does not directly lead to increased outcomes for students and had a somewhat higher effect on Australasian Journal of Educational Technology, 2017, 33(1). 116 behaviour variables than the outcome variables. Students would feel competent and be more interested in those activities in which they are allowed to make choices. Kraiger and Jerden (2007) identified learner control as a useful factor high level learning engagement and for e-learning superiority. Orvis et al. (2009) found that learner control has a positive impact on learning satisfaction, and that high-level learner control over the learning environment features would influence the learners’ affective and utility-based reactions to these features. Learner characteristics Vast bodies of psychology and sociology research have explored and identified the role of user characteristics in differentiating the effects on cognitive, affective and behavioural reactions. There are various aspects from which we can have a good understanding of user characteristics when studying the adoption of information technology, such as: the Big Five personality inventory, which includes openness, conscientiousness, extraversion, agreeableness, and neuroticism as five dimensions of personality (Barrick & Mount, 1991); demographic characteristics (age, gender, culture, education level, and social class), personal innovativeness in information technology (PIIT) (Agarwal & Prasad, 1998); and computer self- efficacy (Compeau & Higgins, 1995). The Big Five is used widely in the individual performance, satisfaction and organisation outcome research (Barrick, Mount, & Judge, 2001; Wagerman & Funder, 2007).Orvis, Brusso, Wasserman and Fisher (2010) examined three of the factors and their moderate effects in the relationship between degree of learner control and learning performance. Generally, the Big Five is used to explain the difference in learning performance. Though it has been used to explain the behaviour toward information technology use, small sample sizes delivered varying results. While McElroy, Hendrickson, Townsend and Demarie. (2007) found only openness influenced internet use, the study by Devaraj, Easley and Crant (2008) indicated all the factors except openness mattered in explaining technology acceptance. Venkatesh, Sykes and Ventkatraman (2014) found openness, conscientiousness, and extraversion affected website usage. Karich et al. (2014) suggested a somewhat stronger effect for students in K-12 compared with college students or adult learners. Xu and Jaggars (2014) found that certain kinds of students are more negatively affected in e-learning, especially males, younger students, students with lower level academic skills, and African-American students. The demographic characteristics of gender, age, and education level are often studied to get a better understanding of learning effectiveness. In Xu and Jaggars’ (2014) research, the survey subjects were similar in the demographic characteristics of age and education level. However, as for the adoption behavior intention, the user characteristics concerned with information technology of PIIT, and computer self-efficacy are more often used and their effects are more consistent in the computer technology context. PIIT is a conception that describes one’s willingness to give any new information technology a try (Agarwal & Prasad, 1998). Compeau and Higgins (1995) defined computer self-efficacy as a judgment of an individual’s possible ability to use a computer when he or she uses it in the future. It is developed from the self-efficacy in the social cognition theory. Doubtless, such factors are closely related to the acceptance of new computer related technology. Thatcher and Perrewe (2002) studied the relationship between PIIT and computer self-efficacy as user related features in the context of using computers. Lewis, Agarwal and Sambamurthy (2003) proposed that individual influence was one of the three main aspects that influence one’s belief about technology acceptance. The study surveyed knowledge workers from a university about their perceptions of using course websites in their work, and identified the positive effects of PIIT and computer self-efficacy as individual factors. Ke, Sun, Yang and Sun (2012) differentiated external variables into user characteristics and system characteristics when they studied the acceptance of an e-learning system, and used PIIT and computer self-efficacy as main user features to influence perception of usefulness and ease of use. Hypotheses development and research model Hypotheses from the original TAM model Based on Davis (1989), an individual’s perception of usefulness and ease of use toward the system are positively related to the attitude toward usage, and the attitude has a positive effect on the intention to use, which influences the actual usage behavior. To keep the model parsimonious, we modified the original Australasian Journal of Educational Technology, 2017, 33(1). 117 relationships a little. As Heijden (2004) did when incorporating enjoyment in TAM, and Gefen, Karahannaand and Straub (2003) did when including trust in TAM, we removed attitude and studied the relationships among perceived usefulness, perceived ease of use, and intention to use. Thus, we have the hypotheses 1, 2 and 3. H1: Perceived usefulness of MOOC learning has a positive effect on the intention to use MOOC learning. H2: Perceived ease of use of MOOC learning has a positive effect on the intention to use MOOC learning. H3: Perceived ease of use of MOOC learning has a positive effect on perceived usefulness of MOOC learning. Hypotheses from perceived learner control Learner control which helps form a respective learning experience should aid learning because it increases interest and motivation to learn (Scheiter & Gerjets, 2007). Gerjets, Scheiter, Opfermann, Hesse and Eysink (2009) conducted a test between learner control and time spent in e-learning, and found that learners with a high level of control tended to devote more time to the e-learning environment than those with a low level of control. Hence, one’s perception of learner control (PLC) can cause different levels of engagement, motivation and attitude toward MOOC learning. Justifiably, an individual perception of MOOC learning including its usefulness, ease of use, and intention may be different because of his/her distinct perception of learner control of MOOC learning. Thus, we developed the hypotheses 4, 5 and 6. H4: Perceived learner control of MOOC learning has a positive effect on perceived usefulness of MOOC learning. H5: Perceived learner control of MOOC learning has a positive effect on perceived ease of use of MOOC learning. H6: Perceived learner control of MOOC learning has a positive effect on the intention to use MOOC learning. Hypotheses from PIIT PIIT as a variable measuring one’s possible positive attitude toward new information technology has been studied in many circumstances. For example, Lu, Yao and Yu (2005) studied the adoption of mobile wireless Internet services and confirmed the effect of PIIT on user’s perception of usefulness and ease of use to accept the new technology. As for the e-learning technology, Raaij and Schepers (2008) found a positive effect of PIIT on perceived ease of use when accepting a virtual learning environment. Tan, Ooi, Leong and Lin (2014) studied the influence of PIIT on perceived ease of use and intention, and only found the hypothesis on perceived ease of use to be true. Therefore, it is quite necessary to incorporate PIIT as a determinant variable under the new circumstances of MOOC learning. Besides the effect of PIIT on perceived usefulness and ease of use, we think the positive attitude caused by PIIT will have some influence on the new construct perceived learner control. Thus, we have the hypotheses H7, H8, H9. H7: PIIT has a positive effect on perceived learner control of MOOC learning. H8: PIIT has a positive effect on perceived ease of use of MOOC learning. H9: PIIT has a positive effect on perceived usefulness of MOOC learning. Hypotheses from e-learning self-efficacy Vijayasarathy (2004) studied the factors that influence the usage of the online shopping and employed a factor, internet self-efficacy, that describes “a consumer’s self-assessment of the capabilities to shop online” (Vijayasarathy, 2004; p.751)to substitute computer self-efficacy. As computer technology has become an indispensable part of our lives, we think it is improper to use computer self-efficacy as a factor to influence the perception of MOOC learning. Roca, Chiu and Martínez (2006) studied the effects of computer self- efficacy and internet self-efficacy on perceived ease of use (PEOU), and found that the former had much less influence than the later. Thus we develop a more learning concerned factor, namely e-learning self- efficacy, to replace computer self-efficacy. We define e-learning self-efficacy (ESE) as a learner’s self- assessment of one’s ability to learn online. E-learning self-efficacy is an important individual characteristic, Australasian Journal of Educational Technology, 2017, 33(1). 118 and it is reasonable to assume people with different levels of e-learning self-efficacy will have a different perceptions of perceived learner control. Thus, we have the hypotheses H10 and H11. H10: E-learning self-efficacy has a positive effect on perceived learner control of MOOC learning. H11: E-learning self-efficacy has a positive effect on perceived ease of use of MOOC learning. The TAM model and its extended models have demonstrated robustness in much research, but these relationships may be significant or not significant. Even the significant relationships will not stay the same among different users and technologies. Venkatesh and Morris (2000) conducted research on acceptance and usage behavior toward new technology and compared gender difference among the relationships. They found that the relationships were proven solid among all samples, all male samples, and all female samples, but there existed differences in the strength of the proven relationships. As in the MOOC context, for Chinese learners there are not only domestic platforms such as XuetangX and ICourse163, but also platforms from other countries, such as Coursera and EdX from America, and Future Learn from England. These foreign platforms usually contain courses taught in English or their native language. Generally, these platforms have similar functions. Learners can retrieve a course based on the classified catalogue, or by using the search engine. They choose to participate in a course, and then complete the course by learning from the teacher and other classmates, completing homework and tests. However, a platform like Coursera has an interface that is usually in English, with courses taught in English. Retrieving courses and learning in English is always hard for Chinese learners, even if they have learned English for several years in school. Besides, the website layout, usage of certain colors, and the logic of process varies across different countries. Chinese learners are more familiar with the design features of their own platforms. All these things indicate that Chinese learners probably find their own platforms easier to use and understand than foreign ones. So it is fair to say, for Chinese learners, perceived usefulness affects intention more when using native platforms, and perceived ease of use affects intention more when using foreign platforms. Thus, we have the last hypothesis. H12: The model is robust across different platform, but the strength of the relationship will differ. Specifically, perceived usefulness matters more when domestic platforms are used, and perceived ease of use matters more when foreign platforms are used. Figure 1. Proposed model Study design and survey sample Questionnaire design This research employed a questionnaire survey to acquire data and test the proposed hypotheses. The questionnaire was divided into two parts. The first part asks for basic information of the survey sample and the second part measures the construct from our model. The entire construct is measured using a 7-point H4 H 3 H2 H1 H5 H 7 H8 H9 H1 0 H11 Perceived learner control E-learning self-efficacy Personal Innovativeness in IT Perceived ease of use Perceived usefulness Intention H6 Australasian Journal of Educational Technology, 2017, 33(1). 119 Likert scale, with 1 indicating strongly disagree and 7 indicating strongly agree. Measurements of the variables are listed in Table 1. To avoid the ambiguity and inaccuracy, experts in social surveying and MOOC learning examined all the measurement items. All items were modified and reviewed to agree with the MOOC learning context. Table 1 summarizes the latent variables, their measurement items, and references. All these measurements were originally in English, and developed into Chinese using a forward-backward translation. Table 1 Measurement item Data collection and sample description Coursera is one of the most used MOOC learning platforms around the world and attracts learners from many countries, including China. Besides foreign platforms, learners from China can also use native platforms such as ICourse163 and XuetangX. We chose Coursera (https://www.coursera.org/) and ICourse163 (http://www.icourse163.org/) as our survey platforms, because they resemble each other with their courses are mainly from universities, and both are popular among Chinese learners. We sent emails to learners in the two platforms to invite them to participant in the survey. Construct Item Description Reference Perceived learner control PLC1 I feel in control when searching for courses on this site. Kim, Ferrin, & Rao, 2008 PLC2 I feel in control with the process of participating in courses on this site. PLC3 I feel in control when learning courses on this site. PLC4 Overall, I feel in control when I'm using this website. Personal innovativeness in information technology PIIT1 If I heard about a new information technology, I would look for ways to experiment with it. Agarwal & Prasad, 1998;Tan et al., 2014 PIIT2 Among my peers, I am usually the first to try out new information technologies PIIT3 Overall, I like to experiment with new information technologies E-learning self- efficacy ESE1 I feel confident using the e-learning system Roca, Chiu & Martínez, 2006 ESE2 I feel confident using the internet for learning. ESE3 I feel confident that I have rich experience in e-learning. Perceived usefulness PU1 Using the website for learning improve my quality of e-learning. Davis,1989; Lee et al.,2005 PU2 Using the website for learning improves my efficiency of e-learning. PU3 Using the website for learning is useful. Perceived ease of use PEOU1 Learning to use the website is easy for me. Davis,1989; Lee et al., 2005 PEOU2 My interaction with the website is clear and understandable. PEOU3 Overall, it's easy to use the website for me. Intention to use INT1 I intend to use the website for learning to pursue personal interests or hobbies. Davis, Bargozzi, & Warshaw,1989; Lee et al., 2005 INT2 I intend to use the website for learning to expand upon my formal learning. INT3 I intend to use the website for learning frequently. Australasian Journal of Educational Technology, 2017, 33(1). 120 The subjects of this research are MOOC learners. To obtain effective subjects, this research used the QQ group tool, which is the most popular Chinese instant messaging service. For better learning and communication, MOOC learners studying the same course on Coursera often spontaneously build a QQ group. Chinese MOOC platforms often provide official QQ groups. Sample QQ groups were discovered by keyword searching in QQ, and examining the discussion forum as calls for building a QQ group always happens here. To avoid possible interference, similar courses were selected across the two platforms. After asking for permission to join the QQ groups, ethical approval was obtained from the QQ group administrator. Emails containing the survey link were sent to group members to invite them to participate in the survey. Academic usage and information privacy principles were introduced before survey questions were presented. We had 245 learners respond to the questionnaire. During the process of data cleaning, this research deleted 15 as invalid, from the sample data. Of the 230 sample data was left, and were Chinese learners. A group of 116 used Coursera more often and answered the measurements of latent variables based on their experience of using Coursera. The remaining 98 used ICourse163 more frequently and answered the measurements of latent variables based their experience of using ICourse163. Most of them had heard about two or more MOOC platforms. The detailed demographic data can be seen in Table 2. Table 2 Demographic Data Measurement Item Frequency Percentage Gender male 134 62.62% female 80 37.38% Education Background junior college or lower 8 3.74% undergraduate 129 60.28% postgraduate or higher 77 35.98% Age <=18 4 1.87% 18-22 (included) 111 51.87% 23-28 (included) 63 29.44% >28 36 16.82% Data analysis The tests of both measurement model and structural model were conducted through the partial least squares (PLS) approach (Ringle, Wende, & Will, 2007). Different from covariance-based structural equation modeling method, PLS is a variance-based method and has some relative advantages. It is less strict on the size of the sample and the distribution of the sample data (Hair, Ringle, & Sarstedt, 2013). As our total sample size in the separate platforms is relatively small, the PLS method was used to process data. Measurement model test The factor loadings, average variance extracted (AVE), composite reliability, and Cronbach's alpha of the measurement model for all sample and separated samples are shown below in Table 3 and Appendix A. The composite reliability for all constructs in the entire sample exceeded 0.90.In the sample from ICourse163 all exceeded 0.80 and in the sample from Coursera all exceeded 0.90. According to Fornell and Larcker (1981), all samples demonstrate a reasonable reliability. An AVE for each construct above the recommended threshold of 0.50 indicated over 50% of the variance observed can be explained by the latent constructs, which shows the reliability of the constructs (Hair, 2009). Clearly, the smallest AVE (0.748), which was for PIIT in the sample from ICourse163, reaches the recommended value. Additional evidence on reliability can be obtained by comparing Cronbach's alpha with the recommended value of 0.7. If Cronbach's alpha is above 0.7, then the construct possesses adequate reliability. The lowest score of Cronbach's alpha was 0.829 for PIIT in the sample from ICourse163 and this is much larger than the recommended value. Hence, our data shows adequate reliability. Australasian Journal of Educational Technology, 2017, 33(1). 121 Table 3 Results of reliability ALL Loading AVE CR α PLC PLC1 0.880 0.829 0.951 0.931 PLC2 0.927 PLC3 0.908 PLC4 0.927 ESE ESE1 0.949 0.842 0.941 0.906 ESE2 0.963 ESE3 0.836 PIIT PIIT1 0.879 0.805 0.925 0.879 PIIT2 0.884 PIIT3 0.928 PU PU1 0.915 0.839 0.940 0.905 PU2 0.918 PU3 0.916 PEOU PEOU1 0.945 0.912 0.969 0.952 PEOU2 0.970 PEOU3 0.950 INT INT1 0.933 0.856 0.947 0.916 INT2 0.906 INT3 0.937 Notes: The ALL column presents all data together. Loading refers to factor loading, AVE refers to average variance extracted, CR refers to composite reliability and α refers to Cronbach's alpha. The validity of the measurement items consists of convergent validity and discriminant validity. Convergent validity describes the highly correlated relationships within a latent variable. Factor loading scores greater than 0.7, together with AVE greater than 0.5, demonstrate adequate convergent validity (Wixom & Watson, 2001). With the lowest factor loading being 0.791, and the lowest AVE 0.748, all the constructs are adequate in terms of convergent validity. Discriminant validity means that latent variables are distinct from each other and much less correlated. This means variables should relate much more strongly to themselves than to others. The discriminant validity can be examined by comparing the square root of the AVE with the correlation between constructs (Fornell & Larcker, 1981). If the square root of the AVE of a construct is larger than any of the correlations between the construct and other constructs, then the discriminant validity is conformed. According to Table 4, all the values on the diagonal were larger than the corresponding value off diagonal, so the discriminant validity is proved to be adequate. Thus, our measurement model was reliable, valid, and suitable for the test of the structural model. Australasian Journal of Educational Technology, 2017, 33(1). 122 Table 4 Results of validity ESE INT PLC PEOU PIIT PU ALL ESE 0.918 INT 0.495 0.925 PLC 0.566 0.514 0.911 PEOU 0.601 0.718 0.556 0.955 PIIT 0.538 0.481 0.325 0.500 0.897 PU 0.566 0.724 0.629 0.673 0.552 0.916 G1 ESE 0.901 INT 0.502 0.867 PLC 0.529 0.590 0.903 PEOU 0.609 0.672 0.549 0.950 PIIT 0.387 0.427 0.213 0.351 0.865 PU 0.546 0.728 0.542 0.577 0.513 0.906 G2 ESE 0.926 INT 0.490 0.956 PLC 0.584 0.476 0.919 PEOU 0.606 0.748 0.564 0.958 PIIT 0.624 0.508 0.393 0.596 0.921 PU 0.567 0.724 0.677 0.744 0.574 0.919 Notes: Column ALL represents all samples together. G1 represents samples from ICourse163. G2 represents samples from Coursera. Hypotheses test The hypotheses were tested employing the bootstrapping method. Chin (1998) recommended 500 times sampling for significance testing of the estimated path coefficient. As more sampling times is usually better, we employed 1000 times sampling for all the calculations. The R2 value in the intention construct is an indicator of the explanatory power of our proposed model. The acceptance level depends on the research context (Hair et al., 2013). Venkatesh, Morris, Davis and Davis (2003) compared several models in terms of user acceptance of IT, and found the range of explained variance in user intention was from 17% to 69%. Using the empirical data collected, our model showed a desirable explanation of the intention to use a MOOC platform for learning. As we can see in Figure 2, in the entire sample, 62.2% of variance of the intention was explained by perceived learner control, perceived ease of use, and perceived usefulness. As for the separate samples, Figure 3 and Figure 4 showed that 64.3% of the variance was explained in ICourse163 sample, 62.5% of the variance was explained in Coursera sample. The results show that the proposed model was powerful in explaining learners’ intention. Therefore, the analysis of path coefficient and significance makes sense. The testing result of the entire sample can be seen in Figure 2. In the entire sample, the hypotheses from the original TAM model were all supported. Perceived learner control was found to affect perceived usefulness (β = 0.350, p < 0.001) and perceived ease of use (β = 0.308, p < 0.001), whereas its direct effect on intention was not supported (β = 0.007, ns). Appling the mediation analysis suggested by Kenny (2015), complete mediation was found in the path from perceived learner control to intention by perceived usefulness and perceived ease of use. E-learning self-efficacy significantly influenced perceived learner control (β = 0.549, p < 0.001), while the influence of PIIT on perceived learner control was not significant (β = 0.029, ns). PIIT was confirmed as an important indicator of perceived usefulness (β = 0.266, p < 0.001) and perceived ease of use (β = 0.242, p < 0.01). The variances of perceived usefulness and perceived ease of use were explained respectively 60.0% and 47.0% by their antecedents. Perceived learner control was explained by e-learning self-efficacy of 32.0% in its variance. In summary, except for H6 (perceived learner control on intention) and H7 (PIIT Australasian Journal of Educational Technology, 2017, 33(1). 123 on perceived learner control), the rest of the hypotheses were supported with our empirical data. Figure 2. Entire sample Figure 3. Sample from ICourse163 Figure 4.Sample for Coursera Note: *p < 0.05, **p < 0.01, ***p < 0.001, ns = non-significant 0.308*** 0.344*** 0.414*** 0.44 0*** 0.350*** 0. 02 9n s 0.2 42* * 0.266*** 0.5 49* ** 0.296** Perceived learner control E-learning self-efficacy Personal Innovativeness in IT Perceived ease of use Perceived usefulness Intention 0.007nsR 2=0.320 R2=0.470 R2=0.600 R2=0.622 0.315*** 0.277*** 0.312*** 0.46 8*** 0.321*** 0. 02 0n s 0.1 42n s 0.346*** 0.5 22* ** 0.385** Perceived learner control E-learning self-efficacy Personal Innovativeness in IT Perceived ease of use Perceived usefulness Intention 0.157nsR 2=0.279 R2=0.451 R2=0.511 R2=0.643 0.310*** 0.425*** 0.463*** 0.42 6*** 0.363*** 0. 04 8n s 0.3 39* 0.182ns 0.5 54* ** 0.214ns Perceived learner control E-learning self-efficacy Personal Innovativeness in IT Perceived ease of use Perceived usefulness Intention -0.077nsR 2=0.342 R2=0.505 R2=0.669 R2=0.625 Australasian Journal of Educational Technology, 2017, 33(1). 124 For H12, a comparison of the supported hypotheses and a test of the difference in paired path coefficient were conducted. Figure 3 and Figure 4 show the different results of our two samples. For the sample from ICourse163, besides H6 and H7, the influence of PIIT on perceived ease of use (H8) were not supported. The variances of perceived usefulness and perceived ease of use were explained respectively as 51.1% and 45.7% by their antecedents. Perceived learner control was explained by e-learning self-efficacy of 27.9% in its variance. For the sample from Coursera, besides H6 and H7, the influences of PIIT (H9) and e-learning self-efficacy on perceived usefulness were not supported. The variance of perceived usefulness and perceived ease of use were explained respectively as 66.9% and 50.5% by their antecedents. Perceived learner control was explained by e-learning self-efficacy of 34.2% in its variance. Overall, the proposed model was robust and provided powerful explanation for the variance in the predicted variables. Testing of the difference of the path coefficient of the two sub-samples was conducted using the method recommended by Keil et al. (2000). The calculation process is displayed in Figure 5 below. Figure 5. The calculation process Table 5 and Appendix B display the hypotheses testing results. According to the results, the difference of H5 (perceived learner control on perceived ease of use) and H7 (PIIT on perceived learner control) are not significant. Though the difference of H6 is significant, it should be explained carefully since H6 is not supported in the proposed model. Except H5, H6 and H7, the rest of the hypotheses all are significantly distinct. In particular, the influence of perceived usefulness on adoption intention was stronger in ICourse163 user group than in Coursera user group (βG1-βG2 > 0, p < 0.05). As to perceived ease of use, the effect on adoption intention and perceived usefulness was stronger in Coursera user group than in ICourse163 user group (βG1-βG2 < 0, p < 0.0001).That is to say, perceived usefulness affected intention more when using native platforms, and perceived ease of use affected intention more when using foreign platforms. Hence, it’s reasonable to proclaim H12 is supported based on the empirical data. Spooled = √{[(N1 - 1)/( N1 + N2-2)] x SE12 + [(N2 - 1)/( N1 + N2 - 2)] x SE22} t = (β1 –β2) / [Spooled x √(1/ N1 + 1/ N2)] Where Spooled = pooled estimator for the variance t = t-statistic with (N1 + N2 -2) degrees of freedom; Ni = sample size of dataset for group i SEi = standard error of path in structural model of group i Australasian Journal of Educational Technology, 2017, 33(1). 125 Table 5 Results of hypotheses test ALL β SE t value p H1 0.440 0.085 5.143 *** H2 0.414 0.085 4.926 *** H3 0.344 0.096 3.614 *** H4 0.350 0.069 5.095 *** H5 0.308 0.084 3.697 *** H6 0.007 0.059 0.091 H7 0.029 0.077 0.370 H8 0.242 0.088 2.718 ** H9 0.266 0.071 3.769 *** H10 0.549 0.064 8.590 *** H11 0.296 0.111 2.680 ** R2PLC 0.320 R2PU 0.600 R2PEOU 0.470 R2INT 0.622 Notes: Column All represents all samples together.*p<0.05, **p<0.01, ***p<0.001 Discussion and implication Discussion of the results In this study, the intention to use a MOOC platform for learning was predicted using the extended TAM model. A new construct named perceived learner control, together with two user characteristics (ESE and PIIT), were incorporated into the TAM model. H1, H2 and H3 were supported. The support of all hypotheses among perceived usefulness, perceived ease of use, and intention demonstrated the robustness of the TAM. H4 and H5 were supported, while H6 was not. According to the data analysis, perceived learner control was an important antecedent for both perceived usefulness and perceived ease of use. Previous researchers have studied how learner control could influence the learning engagement, performance and outcome, but few have identified its influence on attitude and intention to use the technology. The more personalized one perceives an e-learning environment to be, the more control of such environment he/she perceives (Wu & Zhang, 2014). The positive influence of perceived learner control of MOOC learning on cognition of the MOOC learning indicates that giving leaners more choices about their learning process makes learners more positive about the technology or service. This is consistent with the underlying motivation of personalization. However, in terms of perceptions toward MOOC learning, the positive influence of perceived learner control on intention was not direct and fully mediated by PU and PEOU. H8, H9, H10 and H11 were supported, while H7 was not PIIT, consisting with previous studies, had a positive effect on both perceived usefulness and perceived ease of use, but its influence on perceived learner control was not supported. Since PIIT describes the willingness to try out a new information technology, its weak relation with the control perception about the learning process is quite reasonable. As for ESE, a more e-learning concerned factor, its positive effect on perceived learner control was supported. A learner more confident in his/her ability to learn online is likely to feel the learning process is more controllable. Consisting with the demonstrated influence of computer self-efficacy and internet self-efficacy on perceived ease of use, the effect of e-learning self-efficacy, a more specific self-efficacy factor on perceived ease of use was also confirmed. Australasian Journal of Educational Technology, 2017, 33(1). 126 H12 was supported. Though the proposed model showed its robustness across different platforms, there existed some differences regarding the path coefficients. For ICourse163 from China, the positive influence of PIIT on perceived ease of use was not significant (β = 0.142, ns). This may be explained by familiarity with the platform. Since the samples are Chinese, they don’t have the potential language problem when using the platform for learning. Hence, the positive attitude caused by high level PIIT didn’t influence the perceived ease of use, for it had already been easy to use. Compared with Coursera from America, larger path coefficients were found on the paths from e-learning self-efficacy to perceived ease of use, from PIIT to perceived usefulness, and from perceived usefulness to intention. This further confirmed the effect of perceived usefulness in predicting the intention. Learners studying on native platforms, put more attention on perceived usefulness of the platform. For Coursera from America, both the positive influence of PIIT on perceived usefulness (β = 0.182, ns) and e-learning self-efficacy on perceived ease of use (β = 0.214, ns) were not significant. Raaij and Schepers (2008) indicated that non-significant relationship between PIIT and PU could be explained by the type of system studied or variable like voluntariness. Path coefficients related to perceived ease of use (H2, H3 and H8) were larger than the other platform. Learners studying on foreign platforms might feel difficulty in using the platform and emphasised the effect of perceived ease of use. Wang and Baker (2014) found students who were English native speakers were more confident in completing the course than non-native speakers. The lack of confidence was likely to be associated with emphasis on ease of use. Chinese learners on foreign platforms have to put more effort into using the platform, because they have to understand the indications and instructions before actually learning. Compared with learners on domestic platforms, they are more concerned with perceived ease of use. Also, it is reasonable that the effect of e-learning self-efficacy on perceived ease of use is shadowed. Theoretical implications From the insights of students, this study provides a different perspective on perceived control. Control in IT/IS acceptance research is mainly related to external control, which influences the perceived ease of use, for example facilitating conditions and organisational support. In the e-learning context, there is another source of control namely learner control. Learner control has been studied regarding its influences on the learning outcomes in the education field. Hence, our study researched the perceived learner control and its effect on perception toward MOOC learning. Furthermore, the perceived learner control is found to influence the intent through cognition of the MOOC learning, that is, perceived usefulness and perceived ease of use. This paper also gives deeper insight into the inner mechanism of learners’ MOOC adoption intention, by combining two learner characteristics (ESE and PIIT) into the TAM model. The effects of learner characteristics (ESE and PIIT) on MOOC adoption intention are mediated by platform characteristics (PEOU and PU). This means that, when it comes to MOOC adoption intention, students will take both their own personal characteristics and platform’s characteristics into consideration. Last but not least, to test the model’s robustness and its different relationship strength, a comparison study of different countries’ MOOC platforms was implemented. Because of familiarity with the native platform, the positive influence of PIIT on perceived ease of use was not significant. When learning on the native MOOC platform, students put more attention on perceived usefulness of the platform. With the international or foreign MOOC platforms, Chinese students felt difficulty in using the platforms, and emphasising the effect of perceived ease of use. Practical implications For MOOC learning platforms providers should focus on the perception of learner control so as to extend their adoption and take into consideration that students have different usage experiences with foreign MOOC platforms and native platforms. For international platforms providers, ease of use is of more importance to keep their attractiveness for students from different countries who may use different languages and be unfamiliar with the platform. Providing a technologically, linguistically, and culturally neutral platform is very important for the globalisation of the MOOC platform. Hence, the providers should pay attention to keep their platform simple, understandable, and easy to use, which includes translation of language, layout, symbols, navigation, the use of color, and so on. For platforms mainly serving native students, it is important to enhance the usefulness of the platform, such as providing courses that are unique Australasian Journal of Educational Technology, 2017, 33(1). 127 and peculiar to native culture. It does help for them to compete with the international platforms in high quality courses. Further, MOOC platforms or researchers and teachers should cultivate or enhance students’ e-learning self- efficacy. According to our results, e-learning self-efficacy has a positive influence on perceived ease of use, sequentially improving MOOC’s acceptance. ESE is an important student characteristics factor. Some foundational training for teachers, or introduction videos displayed on the platform’s homepage about how to use the MOOC correctly and efficiently is necessary and feasible. Last but not least, the MOOC platform should pay attention to stimulating students’ personal innovativeness in IT and inner learning initiative. PIIT positively affects perceived ease of use and perceived usefulness, and then promotes MOOC adoption intention indirectly. In this paper, PIIT measures one’s possible positive attitude toward MOOCs, as a new digital technology. For students with high levels of innovativeness, the MOOC platform designer or manager should meet their need for innovativeness and maintain their learning enthusiasm by providing more innovative functions and novel courses. For students whose personal innovativeness in IT is relatively low, the MOOC platform manager or operator should communicate with them, listen to them, and strengthen the platform’s interest and interactivity to motivate them, improving their personal innovativeness in IT. For instance, the MOOC platform can encourage students to participate in community discussion or course learning, either by encouragement or as compulsory participation. Conclusion and future work The TAM model has been adapted for various application contexts, and e-learning is definitely an important one. Previous studies have researched the TAM model extended by factors such as perceived enjoyment (Lee, Cheung, & Chen, 2005), social influence (Cheung & Vogel, 2013), and perceived playfulness (Liao, Huang, Chen, & Huang, 2015) in e-learning related technology acceptance. This study introduces a new construct, perceived learner control into the TAM model, and identifies its positive influence on perceived usefulness and ease of use. Though both e-learning self-efficacy and PIIT are learner characteristics, the former influences perceived learner control, while the latter does not. As is confirmed in previous studies, ESE as one kind of self-efficacy, positively influences perceived ease of use in the context of MOOC learning, as does PIIT on perception of usefulness and ease of use. Differences exist during the process of deciding between the different MOOC learning platform because learners weigh perceived ease of use and usefulness differently in distinct situations. There are limitations of this article need to be further researched. More foreign and domestic platforms need to be researched to further investigate the generalisation of platform differences. As in this study, the sample learners tended to have high education levels, the generalisations of the results may be reduced. In addition, more work needs to be done to further investigate the antecedent of perceived learner control. As perceived learner control is important in predicting learners’ adoption intention, it is necessary to find out what kind of system characteristics can influence learner’s perception of control. Acknowledgments The authors are grateful to the editors and the referees for their constructive comments. The authors were supported by the National Natural Science Foundation of China [Grant 71203166], and the Fundamental Research Funds for the Central Universities [Grant 2016 Wuhan University project of Humanities and Social Science” Humanities Computing & Semantic Mining”]. References Adamopoulos, P. (2013). What makes a great MOOC? An interdisciplinary analysis of online course student retention. Retrieved from http://pages.stern.nyu.edu/~padamopo/What makes a great MOOC.pdf Agarwal, R., & Prasad, J. (1998). A conceptual and operational definition of personal innovativeness in the domain of information technology. Information Systems Research, 9(2), 204-215. javascript:showjdsw('jd_t','j_') http://pages.stern.nyu.edu/%7Epadamopo/What%20makes%20a%20great%20MOOC.pdf http://pages.stern.nyu.edu/%7Epadamopo/What%20makes%20a%20great%20MOOC.pdf Australasian Journal of Educational Technology, 2017, 33(1). 128 http://dx.doi.org/10.1287/isre.9.2.20 Agudo-Peregrina, Á. F., Hernández-García, Á., & Pascual-Miguel, F. J. (2014). Behavioral intention, use behavior and the acceptance of electronic learning systems: Differences between higher education and lifelong learning. Computers in Human Behavior, 34(15), 301-314. http://dx.doi.org/10.1016/j.chb.2013.10.035 Arbaugh, J. B. (2014). System, scholar or students? Which most influences online MBA course effectiveness? Journal of Computer Assisted Learning. 30(4), 349-362. http://dx.doi.org/10.1111/jcal.12048 Barrick, M. R., & Mount, M. K. (1991). The Big Five personality dimensions and job performance: A meta-analysis. Personnel Psychology, 44(1), 1–26. http://dx.doi.org/10.1111/j.1744- 6570.1991.tb00688.x Barrick, M. R., Mount, M. K., & Judge, T. A. (2001). Personality and performance at the beginning of the new millennium: What do we know and where do we go next? International Journal of Selection & Assessment, 9(1-2), 9–30. http://dx.doi.org/10.1111/1468-2389.00160 Behrend, T. S., & Thompson, L. F. (2012). Using animated agents in learner-controlled training: The effects of design control. International Journal of Training & Development,16(4), 263-283. http://dx.doi.org/10.1111/j.1468-2419.2012.00413.x Campbell, J., Gibbs, A. L., Najafi, H., & Severinski, C. (2014). A comparison of learner intent and behaviour in live and archived MOOCs. The International Review of Research in Open and Distance Learning, 15(5), 235-262. http://dx.doi.org/10.1016/0260-6917(88)90151-7 Cheng, Y. M. (2014). What drives nurses’ blended e-learning continuance intention? Educational Technology & Society, 17(4), 203–215. http://dx.doi.org/10.1080/20932685.2015.1105111 Cheung, R., & Vogel, D. (2013). Predicting user acceptance of collaborative technologies: An extension of the technology acceptance model for e-learning. Computers & Education,63(2), 160-175. http://dx.doi.org/10.1016/j.compedu.2012.12.003 Chin, W. W. (1998). The partial least squares approach for structural equation modeling. In G. A. Marcoulides (Ed.), Modern methods for business research (pp. 295–336). Mahwah, NJ: Lawrence Erlbaum Associates. Chou, S. W., & Liu, C. H. (2005). Learning effectiveness in a web-based virtual learning environment: a learner control perspective. Journal of Computer Assisted Learning, 21(1), 65-76. http://dx.doi.org/10.1111/j.1365-2729.2005.00114.x Chow, M., Chan, L., Lo, B., Chu, W. P., Chan, T., & Lai, Y. M. (2013). Exploring the intention to use a clinical imaging portal for enhancing healthcare education. Nurse Education Today, 33(6), 655-662. http://dx.doi.org/10.1016/j.nedt.2012.01.009Compeau, D. R., & Higgins, C. A. (1995). Computer self- efficacy: Development of a measure and initial test. MIS Quarterly,19(2), 189-211. http://dx.doi.org/10.2307/249688 Cyr, D., & Trevor-Smith, H. (2004). Localization of web design: An empirical comparison of German, Japanese, and United states web site characteristics. Journal of the American Society for Information Science & Technology, 55(13), 1199-1208. http://dx.doi.org/10.1002/asi.20075 Davis, F. D. (1989) Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, 13(3),319-340. http://dx.doi.org/10.2307/249008 Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982-1003. http://dx.doi.org/10.1287/mnsc.35.8.982 Devaraj, S., Easley, R. F., & Crant, J. M. (2008). Research note-how does personality matter? Relating the five-factor model to technology acceptance and use. Information Systems Research, 19(1), 93-105. http://dx.doi.org/10.1287/isre.1070.0153 Fisher, S. L., Wasserman, M. E., & Orvis, K. A. (2010). Trainee reactions to learner control: An important link in the e-learning equation. International Journal of Training & Development, 14(3), 198-208. http://dx.doi.org/10.1111/j.1468-2419.2010.00352.x Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39-50. http://dx.doi.org/10.2307/3151312 Gasevic, D., Kovanovic, V., Joksimovic, S., & Siemens, G.(2014). Where is research on massive open online courses headed? A data analysis of the MOOC research initiative. The International Review of Research in Open and Distance Learning, 15(5),134-176. Gefen, D., Karahanna, E., & Straub, D. W. (2003). Trust and TAM in online shopping: An integrated model. MIS Quarterly, 27(1), 51-90. http://dx.doi.org/10.1287/isre.9.2.20 http://dx.doi.org/10.1016/j.chb.2013.10.035 http://dx.doi.org/10.1111/jcal.12048 http://dx.doi.org/10.1111/j.1744-6570.1991.tb00688.x http://dx.doi.org/10.1111/j.1744-6570.1991.tb00688.x http://dx.doi.org/10.1111/1468-2389.00160 http://dx.doi.org/10.1111/j.1468-2419.2012.00413.x http://dx.doi.org/10.1016/0260-6917(88)90151-7 http://dx.doi.org/10.1080/20932685.2015.1105111 http://dx.doi.org/10.1016/j.compedu.2012.12.003 http://dx.doi.org/10.1111/j.1365-2729.2005.00114.x http://dx.doi.org/10.1016/j.nedt.2012.01.009 http://dx.doi.org/10.2307/249688 http://dx.doi.org/10.1002/asi.20075 http://dx.doi.org/10.2307/249008 http://dx.doi.org/10.1287/mnsc.35.8.982 http://dx.doi.org/10.1287/isre.1070.0153 http://dx.doi.org/10.1111/j.1468-2419.2010.00352.x http://dx.doi.org/10.2307/3151312 Australasian Journal of Educational Technology, 2017, 33(1). 129 Gerjets, P., Scheiter, K., Opfermann, M., Hesse, F. W., & Eysink, T. H. S. (2009). Learning with hypermedia: The influence of representational formats and different levels of learner control on performance and learning behavior. Computers in Human Behavior, 25(2), 360-370. http://dx.doi.org/10.1016/j.chb.2008.12.015 Hair, J. F. (2009). Multivariate data analysis: A global perspective (7th ed.). Upper Saddle River, NJ: Prentice Hall. Hair J. F., Ringle, C. M., & Sarstedt, M. (2013). Partial least squares structural equation modeling: Rigorous applications, better results and higher acceptance. Long Range Planning, 46(1-2), 1-12. http://dx.doi.org/10.1016/j.lrp.2013.01.001 Heijden, H. (2004). User acceptance of hedonic information systems. MIS Quarterly, 28(4), 695-704. Hughes, M. G., Day, E. A., Wang, X., Schuelke, M. J., Arsenault, M. L., Harkrider, L. N., & Cooper, O. D. (2013). Learner-controlled practice difficulty in the training of a complex task: Cognitive and motivational mechanisms. Journal of Applied Psychology, 98(1), 80. http://dx.doi.org/10.1037/a0029821 Jaggars, S. S., & Bailey, T. (2010). Effectiveness of fully online courses for college students: Response to a department of education meta-analysis. Retrieved from http://files.eric.ed.gov/fulltext/ED512274.pdf Karich, A. C., Burns, M. K., & Maki, K. E. (2014). Updated meta-analysis of learner control within educational technology. Review of Educational Research, 84(3), 392-410. http://dx.doi.org/10.3102/0034654314526064 Ke, C. H., Sun, H. M., Yang, Y. C., & Sun, H. M. (2012). Effects of user and system characteristics on perceived usefulness and perceived ease of use of the web-based classroom response system. Turkish Online Journal of Educational Technology - TOJET, 11(3), 128-143. Keil, M., Tan, B. C. Y., Wei, K. K., Saarinen, T., Tuunainen, V., & Wassenaar, A. (2000). A cross- cultural study on escalation of commitment behavior in software projects. MIS Quarterly, 24(2), 299- 325. http://dx.doi.org/10.2307/3250940 Kenny, D. A. (2015). Mediation. Retrieved from http://davidakenny.net/cm/mediate.htmKibelloh, M., & Bao, Y. (2014). Can online MBA programmes allow professional working mothers to balance work, family, and career progression? A case study in China. The Asia-Pacific Education Researcher, 23(2), 249-259. http://dx.doi.org/10.1007/s40299-013-0101-1 Kim, D. J., Ferrin, D. L., & Rao, H. R. (2008). A trust-based consumer decision-making model in electronic commerce: The role of trust, perceived risk, and their antecedents. Decision support systems, 44(2), 544-564. http://dx.doi.org/10.1016/j.dss.2007.07.001 Kraiger, K., & Jerden, E. (2007). A meta-analytic investigation of learner control: Old findings and new directions. In S. M. Fiore, & E. Salas (Eds.), Toward a science of distributed learning (pp. 65-90). Washington, WA: American Psychological Association. Lahti, M., Hätönen, H., & Välimäki, M. (2014). Impact of e-learning on nurses' and student nurses knowledge, skills, and satisfaction: A systematic review and meta-analysis. International Journal of Nursing Studies, 51(1), 136-49. http://dx.doi.org/10.1016/j.ijnurstu.2012.12.017 Lee, M. K. O., Cheung, C. M. K., & Chen, Z. (2005). Acceptance of internet-based learning medium: The role of extrinsic and intrinsic motivation. Information & Management,42(8), 1095-1104. http://dx.doi.org/10.1016/j.im.2003.10.007 Lewis, W., Agarwal, R., & Sambamurthy, V. (2003). Sources of influence on beliefs about information technology use: An empirical study of knowledge workers. MIS Quarterly, 27(4), 657-678. Liao, Y. W., Huang, Y. M., Chen, H. C., & Huang, S. H. (2015). Exploring the antecedents of collaborative learning performance over social networking sites in a ubiquitous learning context. Computers in Human Behavior, 43(2), 313–323. http://dx.doi.org/10.1016/j.chb.2014.10.028 Liyanagunawardena, T. R., Adams, A. A., & Williams, S. A. (2013). MOOCS: A systematic study of the published literature 2008-2012. The International Review of Research in Open and Distributed Learning, 14(3), 202-227. Lu, J., Yao, J. E., & Yu, C. S. (2005). Personal innovativeness, social influences and adoption of wireless internet services via mobile technology. Journal of Strategic Information Systems, 14(3), 245-268. http://dx.doi.org/10.1016/j.jsis.2005.07.003 Marshall, H., Teeman, D., Mundy, E., Lin, Y., Morrison, J., Yeshanew, T., Cardozo, V., & Stoddart, S.(2009). Harnessing technology schools survey 2009 data report – Part 1, descriptive analysis. British Educational Communications & Technology Agency. Retrieved from http://dera.ioe.ac.uk/1548/ McElroy, J. C., Hendrickson, A. R., Townsend, A. M., & DeMarie, S. M. (2007). Dispositional factors in internet use: personality versus cognitive style. MIS Quarterly, 31(4), 809-820. http://dx.doi.org/10.1016/j.chb.2008.12.015 http://dx.doi.org/10.1016/j.lrp.2013.01.001 http://dx.doi.org/10.1037/a0029821 http://files.eric.ed.gov/fulltext/ED512274.pdf http://dx.doi.org/10.3102/0034654314526064 http://dx.doi.org/10.2307/3250940 http://davidakenny.net/cm/mediate.htm http://dx.doi.org/10.1007/s40299-013-0101-1 http://dx.doi.org/10.1016/j.dss.2007.07.001 http://dx.doi.org/10.1016/j.ijnurstu.2012.12.017 http://dx.doi.org/10.1016/j.im.2003.10.007 http://dx.doi.org/10.1016/j.chb.2014.10.028 http://dx.doi.org/10.1016/j.jsis.2005.07.003 http://dera.ioe.ac.uk/1548/ Australasian Journal of Educational Technology, 2017, 33(1). 130 Means, B., Toyama, Y., Murphy, R., Bakia, M., & Jones, K. (2009). Evaluation of evidence-based practices in online learning: A meta-analysis and review of online learning studies. Washington, WA: U.S. Department of Education. Retrieved from http://www2.ed.gov/rschstat/eval/tech/evidence-based- practices/finalreport.pdf Merrill, M. D. (1975). Learner control: Beyond aptitude-treatment interactions. Educational Technology Research & Development, 23(2), 217-226. http://dx.doi.org/10.1007/BF02768379 Orvis, K. A., Brusso, R. C., Wasserman, M. E., & Fisher, S. L. (2010). E-nabled for e-learning? The moderating role of personality in determining the optimal degree of learner control in an e-learning environment. Human Performance, 24(1), 60-78. http://dx.doi.org/10.1080/08959285.2010.530633 Orvis, K. A., Fisher, S. L., & Wasserman, M. E. (2009). Power to the people: Using learner control to improve trainee reactions and learning in web-based instructional environments. Journal of Applied Psychology, 94(4), 960-971. http://dx.doi.org/10.1037/a001497 Pappano, L. (2012). The year of the MOOC. Retrieved from http://edinaschools.org/cms/lib07/MN01909547/Centricity/Domain/272/The Year of the MOOC NY Times.pdf Pereira, F. A. D. M., Ramos, A. S. M., Gouvêa, M. A., & Costa, M. F. D. (2015). Satisfaction and continuous use intention of e-learning service in Brazilian public organizations. Computers in Human Behavior, 46(51), 139–148. http://dx.doi.org/10.1016/j.chb.2015.01.016 Persico, D., Manca, S., & Pozzi, F. (2014). Adapting the technology acceptance model to evaluate the innovative potential of e-learning systems. Computers in Human Behavior, 30(1), 614-622. http://dx.doi.org/10.1016/j.chb.2013.07.045 Raaij, E. M. V., & Schepers, J. J. L. (2008). The acceptance and use of a virtual learning environment in china. Computers & Education, 50(3), 838-852. http://dx.doi.org/10.1016/j.compedu.2006.09.001 Reeves, T. C. (1993). Pseudoscience in computer-based instruction: The case of learner control research. Journal of Computer-Based Instruction, 20(2), 39-46. Ringle, C. M., Wende, S., & Will, A. (2007). SmartPLS 2.0 M3. Hamburg, Germany: University of Hamburg. Retrieved from http://www.smartpls.de./ Roca, J. C., Chiu, C. M., & Martínez, F. J. (2006). Understanding e-learning continuance intention: An extension of the technology acceptance model. International Journal of Human-Computer Studies, 64(8), 683–696. http://dx.doi.org/10.1016/j.ijhcs.2006.01.003 Sánchez-Franco, M. J., Peral-Peral, B., & Villarejo-Ramos, Á. F. (2014). Users' intrinsic and extrinsic drivers to use a web-based educational environment. Computers & Education, 74(5), 81–97. http://dx.doi.org/10.1016/j.compedu.2014.02.001 Scheiter, K., & Gerjets, P. (2007). Learner control in hypermedia environments. Educational Psychology Review, 19(3), 285-307. http://dx.doi.org/10.1007/s10648-007-9046-3 Shyu, H. Y., & Brown, S. W. (1992). Learner control versus program control in interactive videodisc instruction: What are the effects in procedural learning? International Journal of Instructional Media, 19(2), 85-96. Šumak, B., Heričko, M., & Pušnik, M. (2011). A meta-analysis of e-learning technology acceptance: The role of user types and e-learning technology types. Computers in Human Behavior, 27(6), 2067-2077. http://dx.doi.org/10.1016/j.chb.2011.08.005 Tan, W. H., Ooi, K. B., Leong, L. Y., & Lin, B. (2014). Predicting the drivers of behavioral intention to use mobile learning: A hybrid semi-neural networks approach. Computers in Human Behavior, 36(2), 198-213. http://dx.doi.org/10.1016/j.chb.2014.03.052 Teo, T. (2014). Preservice teachers' satisfaction with e-learning. Social Behavior & Personality: An International Journal, 42(1), 3-6. http://dx.doi.org/10.2224/sbp.2014.42.1.3 Thatcher, J. B., & Perrewe, P. L. (2002). An empirical examination of individual traits as antecedents to computer anxiety and computer self-efficacy. MIS Quarterly, 26(4), 381-396. http://dx.doi.org/10.2307/4132314 Venkatesh, V., & Morris, M. G. (2000). Why don't men ever stop to ask for directions? Gender, social influence, and their role in technology acceptance and usage behavior. MIS Quarterly, 24(1), 115-139. http://dx.doi.org/10.2307/3250981 Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425-478. Venkatesh, V., Sykes, T. A., & Venkatraman, S. (2014). Understanding e-government portal use in rural India: Role of demographic and personality characteristics. Information Systems Journal, 24(3), 249– 269. http://dx.doi.org/10.1111/isj.12008 Vijayasarathy, L. R. (2004). Predicting consumer intentions to use on-line shopping: The case for an http://www2.ed.gov/rschstat/eval/tech/evidence-based-practices/finalreport.pdf http://www2.ed.gov/rschstat/eval/tech/evidence-based-practices/finalreport.pdf http://dx.doi.org/10.1007/BF02768379 http://dx.doi.org/10.1080/08959285.2010.530633 http://dx.doi.org/10.1037/a001497 http://edinaschools.org/cms/lib07/MN01909547/Centricity/Domain/272/The%20Year%20of%20the%20MOOC%20NY%20Times.pdf http://edinaschools.org/cms/lib07/MN01909547/Centricity/Domain/272/The%20Year%20of%20the%20MOOC%20NY%20Times.pdf http://dx.doi.org/10.1016/j.chb.2015.01.016 http://dx.doi.org/10.1016/j.chb.2013.07.045 http://dx.doi.org/10.1016/j.compedu.2006.09.001 http://www.smartpls.de./ http://dx.doi.org/10.1016/j.ijhcs.2006.01.003 http://dx.doi.org/10.1016/j.compedu.2014.02.001 http://dx.doi.org/10.1007/s10648-007-9046-3 http://dx.doi.org/10.1016/j.chb.2011.08.005 http://dx.doi.org/10.1016/j.chb.2014.03.052 http://dx.doi.org/10.2224/sbp.2014.42.1.3 http://dx.doi.org/10.2307/4132314 http://dx.doi.org/10.2307/3250981 http://dx.doi.org/10.1111/isj.12008 Australasian Journal of Educational Technology, 2017, 33(1). 131 augmented technology acceptance model. Information & Management, 41(6), 747-762. http://dx.doi.org/10.1016/j.im.2003.08.011 Wagerman, S. A., & Funder, D. C. (2007). Acquaintance reports of personality and academic achievement: A case for conscientiousness. Journal of Research in Personality, 41(1), 221–229. http://dx.doi.org/10.1016/j.jrp.2006.03.001 Wang Y, & Baker R. (2014). MOOC learner motivation and course completion rates. Retrieved from http://www.moocresearch.com/wp-content/uploads/2014/06/MRI-Report-WangBaker-June-2014.pdf Wixom, B. H., & Watson, H. J. (2001). An empirical investigation of the factors affecting data warehousing success. MIS Quarterly, 25(1), 17-41. http://dx.doi.org/10.2307/3250957 Wu, B., & Zhang, C. (2014). Empirical study on continuance intentions towards e-learning 2.0 systems. Behaviour& Information Technology, 33(10), 1027-1038. http://dx.doi.org/10.1080/0144929x.2014.934291 Xu, D., & Jaggars, S. S. (2013). Adaptability to online learning: Differences across types of students and academic subject areas. Columbia University, NY. Retrieved from http://files.eric.ed.gov/fulltext/ED539911.pdf Yuan, L., & Powell, S. (2013). MOOCs and open education: Implications for higher education. Retrieved from http://publications.cetis.ac.uk/2013/667 Corresponding author: Weiwei Yan, yanww@whu.edu.cn Australasian Journal of Educational Technology © 2017. Please cite as: Zhang, M., Yin, S., Luo, M., & Yan, W. (2017).Learner control, user characteristics, platform differences, and their role in adoption intention for MOOC learning in China. Australasian Journal of Educational Technology, 33(1), 114-133. https://doi.org/10.14742/ajet.2722 http://dx.doi.org/10.1016/j.im.2003.08.011 http://dx.doi.org/10.1016/j.jrp.2006.03.001 http://www.moocresearch.com/wp-content/uploads/2014/06/MRI-Report-WangBaker-June-2014.pdf http://dx.doi.org/10.2307/3250957 http://dx.doi.org/10.1080/0144929x.2014.934291 http://files.eric.ed.gov/fulltext/ED539911.pdf http://publications.cetis.ac.uk/2013/667 mailto:yanww@whu.edu.cn https://doi.org/10.14742/ajet.2722 Australasian Journal of Educational Technology, 2017, 33(1). 132 Appendix A Results of reliability G1 G2 Loading AVE CR α Loading AVE CR α PLC PLC1 0.890 0.815 0.946 0.924 0.883 0.845 0.956 0.939 PLC2 0.926 0.932 PLC3 0.870 0.933 PLC4 0.923 0.928 ESE ESE1 0.930 0.812 0.928 0.883 0.958 0.858 0.948 0.917 ESE2 0.942 0.973 ESE3 0.827 0.843 PIIT PIIT1 0.791 0.748 0.898 0.829 0.923 0.849 0.944 0.912 PIIT2 0.879 0.903 PIIT3 0.918 0.937 PU PU1 0.892 0.821 0.932 0.891 0.924 0.844 0.942 0.908 PU2 0.914 0.918 PU3 0.912 0.914 PEOU PEOU1 0.948 0.903 0.966 0.946 0.944 0.918 0.971 0.955 PEOU2 0.959 0.976 PEOU3 0.945 0.953 INT INT1 0.876 0.751 0.900 0.833 0.958 0.914 0.970 0.953 INT2 0.821 0.955 INT3 0.902 0.955 Notes: G1 presents data from ICourse163 learners. G2 presents data from Coursera learners. Loading refers to factor loading, AVE refers to average variance extracted, CR refers to composite reliability and α refers to Cronbach's alpha. Australasian Journal of Educational Technology, 2017, 33(1). 133 Appendix B Results of hypotheses test G1 G2 Test for H12 β SE t value p β SE tvalue p Difference (βG1-βG2) tvalue p H1 0.468 0.115 3.940 *** 0.426 0.133 3.153 ** 0.042 2.442 * H2 0.312 0.090 3.525 *** 0.463 0.137 3.513 *** -0.151 -9.320 *** H3 0.277 0.135 2.105 * 0.425 0.130 3.366 *** -0.148 -8.136 *** H4 0.321 0.131 2.383 * 0.363 0.081 4.486 *** -0.042 -2.860 ** H5 0.315 0.106 2.954 ** 0.310 0.118 2.567 * 0.005 0.323 H6 0.157 0.096 1.778 -0.077 0.078 1.004 0.234 19.632 *** H7 0.020 0.108 0.092 0.048 0.121 0.390 -0.028 -1.767 H8 0.142 0.118 1.127 0.339 0.136 2.514 * -0.197 -11.188 *** H9 0.346 0.098 3.533 *** 0.182 0.089 1.912 0.164 12.796 *** H10 0.522 0.078 6.702 *** 0.554 0.100 5.542 *** -0.032 -2.569 * H11 0.385 0.121 3.243 ** 0.214 0.183 1.179 0.171 7.888 *** R2PLC 0.279 0.342 R2PU 0.511 0.669 R2PEOU 0.457 0.505 R2INT 0.643 0.625 Notes: G1 represents samples from ICourse163. G2 represents samples from Coursera. *p<0.05, **p<0.01, ***p<0.001 Min Zhang, Shuaijun Yin, Meifen Luo, Weiwei Yan Wuhan University, China Introduction Theoretical background Technology acceptance model Learner control Learner characteristics Hypotheses development and research model Hypotheses from the original TAM model Hypotheses from perceived learner control Hypotheses from PIIT Hypotheses from e-learning self-efficacy Study design and survey sample Questionnaire design Data collection and sample description Data analysis Measurement model test Hypotheses test Discussion and implication Discussion of the results Theoretical implications Practical implications Conclusion and future work Acknowledgments References