Evaluation of Online Video Usage and Learning Satisfaction: An Extension of the Technology Acceptance Model

The aim of the study was to examine the determining factors of students' video usage and their learning satisfaction relating to the supplementary application of educational videos, accessible in a Moodle environment in a Business Mathematics Course. The research model is based on the extension of Technology Acceptance Model (TAM), in which the core TAM constructs – perceived usefulness, perceived ease of use, attitude – and internet self-efficacy were included as the explanatory factors of video usage. As regards the determinants of learning satisfaction, beside the core TAM constructs, the role of learning performance, learner-learner interaction, and learner-teacher interaction was examined. Data were collected from 89 students using a questionnaire, on which the partial leastsquares structural equation modelling approach was used to evaluate the research model. The results confirmed that perceived usefulness, attitude, and internet self-efficacy had a direct effect on the video usage. Learning satisfaction was directly influenced by learner-learner interaction, perceived ease of use, and learning performance. Furthermore, the results indicated that video usage had a significant effect both on learning performance and on learning satisfaction. The findings show that the extended TAM model can be applied for predicting the university students' video technology usage and their learning satisfaction regarding the usage.


Article abstract
The aim of the study was to examine the determining factors of students' video usage and their learning satisfaction relating to the supplementary application of educational videos, accessible in a Moodle environment in a Business Mathematics Course. The research model is based on the extension of Technology Acceptance Model (TAM), in which the core TAM constructs -perceived usefulness, perceived ease of use, attitude -and internet self-efficacy were included as the explanatory factors of video usage. As regards the determinants of learning satisfaction, beside the core TAM constructs, the role of learning performance, learner-learner interaction, and learner-teacher interaction was examined. Data were collected from 89 students using a questionnaire, on which the partial least-squares structural equation modelling approach was used to evaluate the research model. The results confirmed that perceived usefulness, attitude, and internet self-efficacy had a direct effect on the video usage. Learning satisfaction was directly influenced by learner-learner interaction, perceived ease of use, and learning performance. Furthermore, the results indicated that video usage had a significant effect both on learning performance and on learning satisfaction. The findings show that the extended TAM model can be applied for predicting the university students' video technology usage and their learning satisfaction regarding the usage.

Introduction
Thanks to the development of video technolgy and the Internet, as well as the development of multimedia reproduction devices, using video as a web-based multimedia device has become more and more widespread. Nowadays, several institutions offer online educational materials including video.
Students can have rapid access to these materials anywhere and any time.
The numerous available technologies have lead to numerous possibilities for their use, therefore the way in which video technologies can be integrated into education is a serious question for both the management and the teachers. As a result it has become indispensible to evaluate the use of educational videos in different way.

161
To create successful strategies, it is essential to estabilish how to use instructional media in the most efficient way with the aim of engancing learning (Koohang & Durante, 2003). For planning an efficient learner-centered electronic learning environment, apart from the cognitive factors, we must have a better understanding of the students' and behaviour as well as the factors influencing them (Abbad, Morris, & Nahlik, 2009;Giannakos, Chorianopoulos, & Chrisochoides, 2015;Grandon, Alshare, & Kwan, 2005;McConville & Lane, 2006). Therefore research into students' feelings and behaviour linked to video usage is needed. These investigations can provide the management with information that lays the foundation of decision-making (Al-Gahtani, 2016), which is essential because of the high production costs of educational videos (Donkor, 2010).
The aim of the present study was, on the one hand, to analyze the correlation of the factors influencing the university students' video usage and their learning satisfaction. On the other hand, with the help of the elaborated model, a theoretical framework is offered that could be suitable for assesing the video technology usage and the learning satisfaction of the university students. The results will contribute to the successful educational video usage for management and teachers alike.
The model used in the present study was based on the Technology Acceptance Model (TAM). The current research was carried out in a higher-level mathematics course at a College of Business located in Hungary. Besides traditional classroom teaching and electronic textbooks, the students had access to online educational videos available in a Moodle environment. The 10-15 minute long, problem-based educational videos had been recorded by the teachers and the step-by-step solution of each task was demonstrated on a split-screen display.
With Moodle, the students were able to communicate electronically with their group-mates and the teachers using forum participation and message sending.

Videos in Education
The exclusive dominance of educational media such as speech and written textbooks lasting for many decades was overthrown by the appearance of visual and audiovisual media and by their integration into education in the twentieth century. There are more and more examples world-wide for using video technology in the area of organized training. The means of usage and the applied videos are varied. The selection of video type, for example, is influenced by the course content or the level of expenditure.
However, Lonn-Teasley (2009) raise a serious question: How does video technology influence teaching and learning?
In the course of evaluation, examining the cognitive factors, such as learning performance, is of primary importance, but analysing the affective factors, such as learning satisfaction and the behaviour linked to it, is also important. (Noel-Levitz, 2011;Palmer & Koenig-Lewis, 2012).
In descriptive attitude studies where students can use videos optionally along with live lectures, the researchers find that students have a positive attitude and they advise such courses supported by videos to fellow students (Copley, 2007;Dupagne, Millette, & Grinfeder, 2009;Hill & Nelson, 2011;Kelly, Lyng, McGrath, & Cannon, 2009;Williams, Birch, & Hancock, 2012). In the case of satisfaction studies unrelated to the means of usage, most of the students are satisfied (Soong, Chan, Cheers, & Hu, 2006;Gosper et al., 2007). There are comperative examinations as well in which researchers point out that the video enhances learning satisfaction compared both to traditional education (King & He, 2006;Schepers & Wetzels, 2007;El-Sayed & El-Sayed, 2013) and to traditional, video-free, text-based elearning (Choi & Johnson, 2007).
The Technology Acceptance Model (TAM) introduced in the next chapter is used by the majority of correlation revealing studies.

Factors Influencing Video Usage -Technology Acceptance Model
The most widely-used model for the explanation of students' technology use is the Technology Acceptance Model (TAM) associated with the name of Fred Davis (1986).
According to the original model, the user attitude has a direct influence on the usage of a new information system, the aim of which is to measure the acceptance on the part of the system user. This attitude toward using is determined by another two variables: perceived usefulness and perceived ease of use. Perceived ease of use, beside its effect on attitude, has a direct effect on perceived usefulness as well. The relationships between TAM-constructs are shown in Figure 1. Copyright 1986 by the Massachusetts Institute of Technology.
Both perceived usefulness and perceived ease of use have a direct effect on system use (Davis, Bagozzi, & Warshaw, 1989;Sumak, Hericko, & Pusnik, 2011;Venkatesh & Davis, 2000). The above-mentioned correlations of constructs and their role in the explanation of behaviour, have been confirmed in several further studies carried out on different samples in different settings. For example, Sumak et al. (2011) in their meta-analysis enlist only educational applications, which evaluate e-learning systems, environments, or technologies and tools. They come to the conclusion that the results are consistent.
In higher education, the TAM model is used mainly for learning management systems. The Moodle system analysis is the most frequent, to which several extended models have been worked out. In the course of extension, numerous external variables that fit into the given educational environment are used. The variables measure the caracteristic features of the system such as perceived convenience (Hsu & Chang, 2013), lack of LMS availability (Alharbi & Drew, 2014), and technical support (Sanchez & Hueros, 2010), or they measure the users' caracteristics and previous experience such as perceived playfulness (Padilla-Meléndez, Aguila-Obra, Garrido-Moreno, 2013); faculty type, academic year, GPA, skills on computer and internet, difficulty reading from computer screen, the number of previously elearning courses (Majdalawi, Almarabeh, & Mohammad, 2014;Al-Assaf, Almarabeh, & Eddin, 2015), LMS usage experience (Alharbi & Drew, 2014), and computer self-efficacy (Sanchez & Hueros, 2010).
Not only external variables can be used for the extension of the TAM model. In the article written by Islam (2013), three outcome constructs (namely perceived learning assistance, perceived community building assistance, and perceived academic performance) are involved in the TAM model revealing the possible outcomes of the Moodle e-learning systems adoption and the use among university students.
Relatively few studies based on the TAM model with the aim of examining video technology acceptance are available. Using the TAM constructions Donkor (2011) estimated the students' acceptance satisfaction regarding video lessons used among students of distance education. Lee and Lehto (2013) identified the determinants of the YouTube behavioural intention with their model in the course of their study into procedural learning through YouTube in a lab setting. For this they used an extension of the TAM model where they took the user satisfaction mediator variable and YouTube self-efficacy, 164 vividness, content richness, task technology fit external variables into consideration beside the usual variables.
Based on the above-mentioned results, the following hypotheses have been set up in this research: H1: Perceived ease of use has a significant positive effect on perceived usefulness.
H2: Perceived ease of use has a significant positive effect on attitude.
H3: Perceived usefulness has a significant positive effect on attitude.
H4: Attitude has a significant positive effect on video usage. According to the claim of Davis and his collegaues Venkatesh & Davis, 1996) when using the TAM it is important to take into consideration the external factors influencing perceived usefulness and perceived ease of use.
Including these factors in the study, special information can be obtained, apart from general correlations. Moreover, with their help, the explanatory power can be increased Mathieson, 1991).
According to Abdullah and Ward's (2016) meta-analysis, self-efficacy is the most commonly used external factor in applications of the TAM in education. Self-efficacy is defined by Bandura (1997) as the belief "in one's capabilities to organize and execute the courses of action required to produce different attainments" (p. 3).
In the course of web-based learning, the student's self-efficacy can be interpreted as the ability to fulfil learning tasks with the help of the e-learning system or technology under consideration (Abbad, 2010).
Those students who have a higher level of self-efficacy for example in operating an e-learning system have a more positive perception of the ease of use of the given system and its usefulness and they are more willing to accept and use the system than their fellow-students who have a lower level of selfefficacy (Abbad, 2010).

165
In this study the role of students' internet self-efficacy is examined with regard to two hypotheses.
H8: Internet self-efficacy has a significant positive effect on perceived ease of use.
H9: Internet self-efficacy has a significant positive effect on perceived usefulness.

Factors Influencing Learning Satisfaction
Satisfaction in general is the feeling of difference between prior expectations and perceived achievement. Keller (1983) defines learning satisfaction as a student's overall positive assessment of his or her learning experience. Satisfaction can be measured only after the learning activity.
In the case of learning using e-learning systems or technology, the most important factors, which have a positive effect on learning satisfaction, are the learner's actual performance (Hui, Hu, Clark, Tam, & Milton, 2008;Liao, Palvia, & Chen, 2009) and perceived usefulness, perceived ease of use and attitude, which are also connected to the given e-learning system or technology, measurable also after the learning activity (Del Barrio, Romero-Frías, & Arquero, 2013;Hui et al., 2008;Lee & Lehto, 2013;Sun, Tsai, Finger, Chen, & Yeh, 2008). All of these are measurable only following the learning activity.
Consequently, with regard to video-based learning satisfaction, the following hypotheses were formulated: H10: Learning performance has a significant positive effect on learning satisfaction. H11: Perceived usefulness has a significant positive effect on learning satisfaction.
H12: Perceived ease of use has a significant positive effect on learning satisfaction.
H13: Attitude has a significant positive effect on learning satisfaction.
A further factor, identified by several researchers, (Abdous & Yen, 2010;Ali & Ahmad, 2011;Bray, Aoki, & Dlugosh, 2008;Lee, 2012;Sahin, 2007;Yukselturk & Yildirim, 2008) is interaction which is an important predictor of learning satisfaction in the course of web-based learning. Interaction is defined as verbal and non-verbal communication between players (Laurel, 1993). There are several types of interaction groupings, which are based on the type of players in them. The most widely-used system was developed by Moore (1989) for distance education and it contains three types of interaction groups: learner-teacher interaction, learner-learner interaction, and learner-content interaction.
In the course of web-based learning, students can exploit several interaction tools, for example instant message, e-mail, forum, discussion board, voice call, video call, voice conferencing, and video conferencing; all of which support interaction with other students and the teacher. Some of these tools can be found intergrated into e-learning systems, like Moodle.
If we examine the role of learner-teacher interaction and the role of learner-learner interaction in learning satisfaction separately, the results will be different. Some researchers claim that learnerteacher interaction has the greatest impact on learning satisfaction (Kuo, Walker, & Schroder, 2010;Kuo, Walker, Schroder, & Belland, 2014;Marks, Sibley, & Arbaugh, 2005;Sahin, 2007) others, 166 however, confirmed that learner-learner interaction has a greater effect on learning satisfaction (Jung et al., 2002;Rodriguez Robles, 2006).
The difference in findings can have several explanations. One of them is that different educational settings, such as different course requirements, need interactions of different levels (Kuo et al., 2010); the other is that interactive preferences may change depending on the type of student (Croxton, 2014).
Based on the results of previous research, it is assumed in this study that interaction has a positive effect on satisfaction. From this it was hypothesized: H14: Learner-teacher interaction has a significant positive effect on learning satisfaction.
H15: Learner-learner interaction has a significant positive effect on learning satisfaction.  The definitions of the constructs used in the present research is summarized in Table 1. Table 1 Construct Definitions

Construct
Definition Source PU Perceived usefulness is defined as the degree to which the user thinks that using the video will enhance his or her performance. (Davis, 1989) PEU Perceived ease of use is defined as the degree to which the user thinks that using the video will be free from effort. (Davis, 1989) A Attitude towards using videos refers to the student's positive or negative feelings associated with video-based learning.
(Fishbein & Ajzen, 1975, as cited in Davis et al., 1989) SAT Learning satisfaction refers to the student's positive feelings toward video-based learning experience. (Keller, 1983) U Video usage is defined as the overall amount of video usage estimated by the user.

Method Questionnaire Development, Content Validity
For the collection of demographic and educational data multiple choice questions were used (4 items), out of which there were one numerical and three nominal-level variables (Table 2).
To ensure content validity, the measurement of the latent variables in the model under study was carried out using a reduced version of the scale employed in other research (26 items). As a result, for the measurement of the constructs, widely-used scales or their simplified variants were used. Some minor changes were made to the formulation so that the items fit into the inter-relation of the research (Appendix 1).

The Population Under Examination
One  Table   2.

Data Analysis
To evaluate the research model and to examine the hypotheses, structural equation modelling (SEM) using partial least squares (PLS) regression was used. The use of PLS path analysis (PLS-SEM) was justified by the low sample size (N=89) of the survey (Henseler, Ringle, & Sinkovics, 2009) as well as the fact that normality was not confirmed in every construct (Chin, 1998).
The analysis was carried out using the SmartPLS 3 program (Ringle, Wende, & Will, 2005). For the examination of construct reliability the Cronbach's alpha measurement of inner consistency with 0.7 limit was used as well as composite reliability (CR) also with 0.7 limit (Fornell & Larcker, 1981;Hair, Black, Babin, & Anderson, 2010). Table 3 shows that in the case of all constructs, both conditions are fulfilled and the scales are reliable.  based on the 0.5 criterion value, suggested by (Fornell & Larcker, 1981). Table 3 shows that the AVE value of each construct is above 0.5, thus the condition of convergent validity are fulfilled. Discriminant validity was proved in this study with the help of correlation matrix and the AVE values (Table 4) based on Fornell and Larcker (1981) criterion. Since, in the case of all variables, the square root of AVE values is greater than the off-diagonal values, discriminant validity between variables is satisfactory (Fornell & Larcker, 1981).

Results and Discussion
The model obtained as a result of the PLS path analysis is shown in Figure 3, on which are displayed the path coefficients, their significance tests (the bootstrapping results) and the multiple determination coefficients (R 2 ). Based on the above-mentioned results hypotheses H1-H5 have been accepted, hypothesis H6, however, has been rejected.
The model role of internet self-efficacy was formulated in the form of two hypotheses. The results harmonize with those of other researchers (Lee et al., 2014;Liang & Tsai, 2008;Rezaei et al., 2008;Roca et al., 2006) according to which internet self-efficacy has a direct significant positive effect on perceived ease of use (H8, β=0.531; p<0.001) as well as on perceived usefulness (H9, β=0.212; p=0.025). Based on this, hypotheses H8-H9 were accepted. From the results it can be concluded that internet self-efficacy has a significant indirect effect on students' attitude (0.412; p<0.001) and on video usage (0.347; p<0.001) as well.
Using the extended TAM model mentioned above, 47.4% of video usage variance was successfully explained. This explanatory power of the extended TAM model is consistent with previous studies. It is considered that TAM consistently explains about 34%-52% of the variance in usage intentions (in case of a specific system) (Venkatesh & Davis, 2000). In accordance with earlier studies (Liu, 2008;Shi, Chen, & Tian, 2011;Tsai & Tsai, 2003;Kuo et al., 2014;Womble, 2007), this study confirms that internet self-efficacy has an indirect positive effect on learning performance (0.087; p=0030) and learning satisfaction (0.233; p=0.001).
The indirect effects between constructs as well as their significance are shown in Table 5.

Conclusion and Implications
The TAM model was extended in the study with the aim of analysing the university students' video usage and the correlations of the factors influencing learning satisfaction relating to the usage. The model provides a useful theoretical basis for understanding the Edutus College students' video usage and learning satisfaction. The findings assist the college leadership and teachers in successful application of educational video.
Furthermore the model, which was validated in the college environment, ensures an integrated theoretical framework which can be used to estimate the university students' usage of video technology and the learning satisfaction connected with the usage, whether it be within the framework of traditional education or of distance learning.
The results of the study have several consequences for further research and practical application. The data to a certain extent support the existence of the basic TAM correlations in a Hungarian college of business studies, thus confirming the robustness and explanatory power of TAM for the use of technology in a higher education setting.
Perceived usefulness and attitude exert the strongest influence on video usage; therefore it is advisable to take measures that could reinforce the positive opinion of students' video usage.
Perceived ease of use did not have a significant direct effect on video usage. The logical explanation for this could be that video usage for educational purposes, accessible online, proves to be easy. Another reasonable explanation could be that students tend to overcome the difficulties of video usage to achieve better results in their studies.
Internet self-efficacy significantly influences video usage, learning performance, and learning satisfaction. It is therefore essential to assess self-efficacy and to use suitable methods for increasing the students' internet self-efficacy. They can be embedded in education if needed.

175
Learner-teacher interaction had a significant effect on learning satisfaction, while the effect of learnerlearner interaction on learning satisfaction was not significant. The possible explanation of the latter correlation on the one hand is that in the type of course learner-learner interaction was not expected, in other words group work was not used. On the other hand the study did not take the quality of interactions and their goal into account. As a result we were not familiar with the type of interaction among students.
Learner-teacher interaction can be inspired via the teacher's intervention and initiative. It could equally be important to guarantee further possibilities for interaction in a Moodle environment, since the intensity of interaction can considerably be influenced by the quantity and quality of communication channels offered by LMS.

Limitations and Future Work
As in every study some limitations should be mentioned in this study as well, a part of which could inspire further research. Firstly, the restrictions are attributed to the application of self-admission measurement, as a result the data obtained were not objective for the examination of the model. Therefore, further research should be carried out in the future in which it would be advisable to combine the video usage data based on self-confession with the data of online students' activities derived from the diary file of the web server.
Secondly, since the research was carried out in a given course of a college, in a given education setting, we should be careful when generalizing the results. It is also advisable to use the model in the future for analysing further courses in other fields of science and offered by other universities.
Thirdly, every effort should be made to increase the explained variance: by increasing Cronbach's alpha, the number of items in each scale, then performing an item analysis to get the highest possible alpha; by increasing the sample size and by making sure that the sample is adequately homogeneous, for which the examination analysis of different individual factors (such as socio-economic background knowledge of ICT and motivational factors) as moderator variables is suggested, using multi-group analysis methods.