Assessing the factors influencing users accessing higher education content on TikTok

Abstract As its number of users increases, TikTok needs to focus on user retention. One such effort involves answering educational needs through social media. The purpose of this study is to analyze the factors affecting the continuance intention of TikTok users to obtain higher educational content. Data from this study were collected through online questionnaires from a total of 629 respondents. Then, the respondents’ data were processed using covariance-based structural equation modeling. This study found that the perceived usefulness and satisfaction affect TikTok users’ continuance intention towards higher educational content. This study suggests that TikTok should develop functionality and features that are more aligned with user preferences for accessing higher educational content.


Introduction
Social media allows people to interact and have the freedom to create, share, and exchange information in virtual communities and networks. Some popular social media applications are TikTok, Facebook, Instagram, and Snapchat (Statista, 2022). TikTok is a platform that provides short video content for all its users (Business of Apps, 2022). By 2021, there were 1.2 billion active Putu Wuri Handayani ABOUT THE AUTHORS Nadia Adilah Rahimullah, Sinta Bela Damayanti, Aisyah Amyra Izra received their bachelor's degree in information system from Universitas Indonesia. Recently, they have taken a keen interest in research on e-learning. Currently, they are working at a private organization as system analyst. Putu Wuri Handayani is a lecture and head of Information System Department in Faculty of Computer Science Universitas Indonesia. She obtained her doctoral study at the Faculty of Computer Science Universitas Indonesia. Her research interest is related to information system/information technology covering topics in e-health, e-commerce, enterprise resource planning, supply chain management, and customer relationship management.

PUBLIC INTEREST STATEMENT
To date, TikTok needs to focus on user retention by continuing to innovate and better meet its users' needs, especially in terms of higher educational content. Thus, this study analyzes the factors affecting the continuance intention of TikTok users to obtain higher educational content. This study used quantitative approach using online questionnaire and qualitative approach using online interview. Result of this study found that the perceived usefulness and satisfaction affect TikTok users' continuance intention towards higher educational content. Further, information quality, system quality, and the enjoyment perceived by users affect user satisfaction regarding TikTok. Finally, the continuation intention to use TikTok is influenced by the confirmation of the users' initial expectations and the level of the perceived usefulness of TikTok for accessing higher education content.
TikTok users worldwide, and this is expected to reach 1.5 billion by the end of 2022 (BBC, 2020). The Covid-19 pandemic has also affected the increase in TikTok application users to date (Backlinko, 2022). According to Statista, there were 22.2 million active TikTok users in Indonesia at the end of 2020, which makes Indonesia the second largest user in the world (Statista, 2022). However, despite the significant increase in users, it was found that TikTok has an uninstall rate of 9.43%, which is much larger than those of other social media platforms, such as Facebook, which has an uninstall rate of 1.6%, and Instagram, which has an uninstall rate of 4% (Hariharan, 2019). Therefore, TikTok also needs to focus on user retention by continuing to innovate and better meet its users' needs, especially in terms of higher educational content (TikTok, 2019).
Several studies have been conducted regarding behavioral intentions and the continuance intention to use a system or social media for e-learning (Cheng, 2020;Pang et al., 2020;Suzianti & Paramadini, 2021). However, studies conducted by Cheng (2020) and Pang et al. (2020) only focused on developed countries, and a study conducted by Suzianti and Paramadini (2021) was limited to respondents who worked as elementary school teachers in Java. Further, Fajardo et al. (2021) conducted experiments related to the use of the TikTok application as a learning tool in e-learning. The results of the study indicate that TikTok application use in education has a positive impact on student learning outcomes because TikTok can create an interesting and motivating learning environment for students (Fajardo et al., 2021). However, the study only used a limited number of respondents, with a total of 65 students (Fajardo et al., 2021). The present study aims to fill the existing research gap and help TikTok increase its user retention by analyzing factors affecting users' continuance intention to use the TikTok social media platform for education in Indonesia. This study uses two widely adopted theoretical models from previous studies to analyze users' continuance intention to use applications for e-learning (Cheng, 2020;Pang et al., 2020), namely expectation confirmation model (ECM) and the information system success model (ISSM). In the field of information systems, ECM has been widely adopted in studies related to application or product use and has revealed positive results with regard to the continuance intention to use an application (Cheng, 2020;Suzianti & Paramadini, 2021). The ISSM is also one of the most commonly used models in the study of various uses of information systems to measure system success (Pang et al., 2020;Putra et al., 2018;Suzianti & Paramadini, 2021;Zulfan, 2018). The results of this study are expected to help TikTok and other social media companies in strategizing and developing applications in the future to improve the services offered in terms of Indonesian education.

The TikTok social media platform
To date, TikTok application has video length limit to 10 minutes (micro video). This micro video is the most popular form of multimedia in the mobile internet domain (Guo et al., 2021). In general, micro videos have three characteristics: The first is their short duration, which makes them easy to spread on social media (Guo et al., 2021). Second, social attributes, such as descriptions, numbers of followers, and numbers of clicks, can facilitate subjective expressions of micro-video semantics (Guo et al., 2021). Third, the videos are generated in real time so that they can capture the emotions and feelings of content creators (Guo et al., 2021). Content that is short, informative, expressive, and to the point is the reason why users choose TikTok (Guo et al., 2021).
Regardless of the strategy that TikTok uses to increase its number of users, it is an application filled with comedy content, knowledge, talent, and more, which entertain users and make them feel addicted (Yang & Zilberg, 2020). A recommendation system that uses artificial intelligence (AI) has also been successfully adopted by TikTok (Grandinetti, 2021). AI algorithms can accurately learn user preferences by providing content recommendations based on previous searches (Bobadilla et al., 2018). The use of AI systems to display content recommendations has been widely used on social networks, such as Facebook, Instagram, and Twitter, to better understand their users' preferences (Kenyon, 2021). TikTok is used as entertainment and as a way to relieve boredom when spending time at home by showcasing creativity and talent, expressing frustration, and exercising (Utami, 2021). In addition, TikTok users also consist of various groups ranging from students, celebrities, and influencers to government officials (Utami, 2021). This means that TikTok is used as a forum for disseminating messages, and it appeals to the wider community, as has been especially observed during the current pandemic (Utami, 2021). However, besides all the advantages provided by TikTok, this application can also trigger a sense of addiction (Smith & Short, 2022). Although there are still concerns about the negative impact of social media, it is undeniable that it can enrich students' learning experiences (Greenhow & Lewin, 2015).
Although TikTok may be more popular as a platform where users upload videos of dancing, lip syncing, and other talents, the sharing of educational content is also starting to increase (Fiallos et al., 2021). In October 2019, TikTok hosted the #EduTok program, which aimed to revolutionize e-learning in India (TikTok, 2019). The program received a positive response, with over 10 million items of content created using the hashtag and 48 billion views (TikTok, 2019). According to an article on the TikTok platform, Senior Vice President Toppr supports organizations in using digital tools to transform India's education industry (TikTok, 2019). A year after the #Edutok program, TikTok created the #LearnOnTikTok (TikTok, 2020) program aimed at residents of the United States, and, to date, it has had 250.7 billion views. In this program, TikTok collaborates with 800 educators, educational institutions, and non-profit organizations to create educational content on TikTok (TikTok, 2020). There are various types of educational content that can be found on #LearnOnTikTok, ranging from tips, tutorials, motivational information, math tricks, unique science experiments, and other educational content (TikTok, 2020). An analysis by Fiallos et al. (2021) of 1,495 TikTok content items with the hashtag #LearnOnTikTok found that the most liked topics came from the fields of medicine and healthcare, food and drink, health, science/chemistry, and technology and engineering. This information shows that there is a great opportunity to increase the dissemination of knowledge from various fields of science quickly and efficiently through the TikTok application (Fiallos et al., 2021).

Expectation confirmation model
The ECM is a theoretical model that only focuses on post-consumption factors. According to Bhattacherjee (2001), it is important to examine post-consumption expectation because there might be changes in the level of user expectations before and after using the system. The ECM includes perceived usefulness to represent post-consumption expectation. Several studies have adopted the ECM to examine behavior and users' continuance intention in the context of education, and it has been proven that perceived usefulness, expectation fulfillment (confirmation), and satisfaction have a significant effect on users' continuance intention (Al-Fraihat et al., 2019;Cheng, 2020;Suzianti & Paramadini, 2021;Zhang et al., 2020).

Information system success model
When implementing a system for a company, it is very important to conduct an evaluation (DeLone & McLean, 2003). Such an evaluation can be conducted using the ISSM, which is useful for determining whether the system created is effective and efficient in delivering information to the recipient (DeLone & McLean, 2003). The use of the ISSM can help identify aspects that still need to be improved as well as the positive impacts that can be obtained from the system (DeLone & McLean, 2003). It includes six dimensions, namely system quality, information quality, usage, user satisfaction, individual impact, and organizational impact, to measure the success of information systems, and these dimensions are said to be interconnected (DeLone & McLean, 2003). System quality measures the success of an information system (IS) technically; information quality measures IS success semantically; and usage, user satisfaction, individual impact, and organizational impact measure the success of an IS in terms of its effectiveness (DeLone & McLean, 2003). The interrelationships of these dimensions are important for measuring, analyzing, and reporting on the success of an IS (DeLone & McLean, 2003).

Conceptual model
The main basis of the proposed research model is the studies conducted by Cheng (2020) and Suzianti and Paramadini (2021). The ECM is the theoretical basis for this study because it has been widely adopted to examine user satisfaction and continuance intention after using a system or product (Cheng, 2020;Gupta et al., 2020;Nguyen & Ha, 2021;Suzianti & Paramadini, 2021). The confirmation, perceived usefulness, satisfaction, and continuance intention factors in the ECM have been shown to have an effect on users' continuance intention in the context of e-learning (Al-Fraihat et al., 2019;Cheng, 2020;Suzianti & Paramadini, 2021;Zhang et al., 2020).
To examine users' technology-focused continuance intention, we combined the ECM and ISSM. In the context of e-learning, information quality, system quality, and service quality factors in the ISSM have been shown to have an effect on continuance intention and user satisfaction (Dreheeb et al., 2016;Mtebe & Raphael, 2018;Pang et al., 2020;Suzianti & Paramadini, 2021). In this study, the ISSM was adopted to identify the influence of system quality and information quality on user satisfaction.
We also added enjoyment and interactivity to examine how satisfied users are with TikTok and how useful they find it to be in an educational context. Interactivity has been shown to increase the effectiveness of the learning process and motivate students to think critically (Munusamy et al., 2019). Enjoyment has also been shown to have a significant effect on perceived usefulness, which significantly affects the continuance intention (Al Kurdi et al., 2020). Although the effect of enjoyment and interactivity on users' continuance intention has been proven (Chao, 2019; Cheng, 2020; Al Kurdi et al., 2020;Weng et al., 2018), only limited studies used these variables. Figure 1 describes the proposed conceptual model.
Confirmation is used to observe the perceived usefulness felt by a user when using an IS (Zhang et al., 2020;Daneji et al., 2019;Suzianti & Paramadini, 2021). According to a study conducted by Zhang et al. (2020), the level of confirmation positively affects perceived usefulness. In addition, Daneji et al. (2019), and Suzianti and Paramadini (2020) stated that confirmation has an effect on perceived usefulness. Besides affecting perceived usefulness, confirmation also affects user satisfaction (Bhattacherjee, 2001). Confirmation is positively associated with user satisfaction in terms of using an IS (Zhang et al., 2020;Daneji et al., 2019;Suzianti & Paramadini, 2020). In addition, Zhang et al. (2020), Daneji et al. (2019), and Suzianti and Paramadini (2020) stated that confirmation has an effect on satisfaction. Therefore, we propose the following hypothesis:

H1: Confirmation (CON) influences users' perceived usefulness (PU) of higher education content on TikTok
Besides affecting perceived usefulness, the confirmation variable also affects user satisfaction (Bhattacherjee, 2001). It is positively associated with user satisfaction with regard to using an IS (Zhang et al., 2020;Daneji et al., 2019;Suzianti & Paramadini, 2020). In addition, research conducted by Zhang et al. (2020) regarding students' sustainable use of virtual labs and remotes; Daneji et al.'s (2019) research regarding the intention to continue using a massive open online course (MOOC); and Suzianti and Paramadini's (2020) study regarding teachers' intentions to continue using e-learning stated that confirmation has an effect on satisfaction. Therefore, this study defines confirmation, with regard to TikTok, as the satisfaction felt by users when using TikTok to watch educational content.

H2: Confirmation (CON) influences user satisfaction (SAT) regarding higher education content on TikTok
Perceived usefulness is defined as the extent to which a person believes that using a particular system will improve his or her job performance (Davis et al., 1989). Basyar and Sanaji (2018) define perceived usefulness as a parameter that measures the extent to which users believe that the use of technology, such as an application, can increase work productivity and provide benefits for its users (Basyar & Sanaji, 2018). Meanwhile, according to Pang et al. (2020), perceived usefulness is the users' perception of the performance of a system or product after using it. The perceived usefulness is positively associated with the continuance intention of users to use a system (Pang et al., 2020;Daneji et al., 2019;Suzianti & Paramadini, 2020). When TikTok provides educational content services that are considered useful and beneficial for its users, it will have a positive impact and increase the sustainability intentions of TikTok users in terms of obtaining educational content from the platform. Therefore, we propose the following hypothesis:

H3: Perceived usefulness (PU) influences users' continuance intention (CI) regarding higher education services on TikTok
Satisfaction felt by users will lead to user continuance intentions (Cheng, 2020; Gummerus et al., 2004;Lee, 2010;Lin & Wang, 2012). In the context of using TikTok for education, if expectations are met, users will feel satisfied, and this will lead to the intention to continue using the application. According to Oliver (1980), satisfaction is the key to building and maintaining long-term users. Dissatisfied users can easily switch to other applications without much consideration (Deng et al., 2010). Several studies have proven the effect of the value of satisfaction on continuance intention in the context of e-learning (Cheng, 2020;Chow & Shi, 2014;Pang et al., 2020;Suzianti & Paramadini, 2021). Thus, we present the following hypothesis:

H4: User satisfaction (SAT) influences users' continuance intention (CI) towards higher education services on TikTok
Satisfaction is defined as a user's attitude or emotion towards an application after interacting with or using the program (Doll et al., 1998). According to Pang et al. (2020), the higher the user's belief that the application can increase productivity and provide benefits, the more satisfied the user will be with the application's performance. This happens because with the increase in productivity, the benefits felt by users meet or exceed the expected benefits (Arcand et al., 2017). Users of the TikTok application will feel satisfied when its performance is considered useful and helps users meet their needs in an educational context. Several studies have proven the effect of perceived usefulness on satisfaction in the context of e-learning (Cheng, 2020;Pang et al., 2020;Suzianti & Paramadini, 2021). Cheng (2020) shows that perceived usefulness is one of the main elements in determining user satisfaction. In addition, Suzianti and Paramadini (2021) also demonstrate how the more useful an application is considered to be for users, the higher the productivity and satisfaction felt by the users. Consequently, we suggest the following hypothesis:

H5: Perceived usefulness (PU) influences user satisfaction (SAT) regarding higher education services on TikTok
According to the ISSM, system quality is an important factor influencing user satisfaction (DeLone & McLean, 2003). If the features of the TikTok application run well and are perceived to be of a high quality, users will feel satisfied. In addition, several studies have proven that system quality affects user satisfaction in the context of e-learning (Dreheeb et al., 2016;Mtebe & Raphael, 2018;Pang et al., 2020;Suzianti & Paramadini, 2021). Pang et al. (2020) found that system quality lies in the interface design, available access speed, and reliability of the available services (Pang et al., 2020). Therefore, it is important for a system to have a clear and neat interface design, efficient access availability, and uninterrupted service to increase user satisfaction. Consequently, a good quality system will increase users' trust in the system, and users will tend to use the system continuously, which will affect user satisfaction (Pang et al., 2020). Hence, we formulate the following hypothesis:

H6: System quality (SQ) influences user satisfaction (SAT) regarding higher education services on TikTok
Information quality can be defined as the extent to which a person feels that the information obtained is organized, current, accurate, relevant, and extensive (Halonen et al., 2009). According to Gao et al. (2012), information quality is the usefulness of information obtained from a system, which will later be used by someone to help assess the quality of the system. Koivumäki et al. (2008) found that significant correlation exists between information quality and user satisfaction.
If TikTok has good quality information that is accurate, timely, needed, and supported by primary and supporting references, as well as being easy to understand, users will feel satisfied and tend to continue using the application. We suggest the following hypothesis:

H7: Information quality (IQ) influences user satisfaction (SAT) regarding higher education services on TikTok
Interactivity is the degree to which a person in the communication process has control and can change roles during an activity (Williams et al., 1988). Within the scope of e-learning, it can be defined as communication that gives users control over the website or application used and the opportunities provided for them to communicate reciprocally (Barreda et al., 2016). Interactivity in a technological environment has a significant effect in terms of increasing one's effectiveness in the learning process and has been recognized as an important factor in virtual learning environments (Cheng, 2020). This is because interactivity encourages students to actively participate in learning and motivates them to better understand meaningful information and think critically (Evans & Gibbons, 2007;Khaled et Wu, 2006). Two-way communication refers to reciprocal communication between the source and the receiver. When these three needs are met, users will feel that the e-learning system is a useful learning tool (Pituch & Lee, 2006). Users will also find the system useful based on its interactive functions (Cho et al., 2009). This certainly proves that interactivity has an influence on perceived usefulness (Chen, 2012). The relationship between interactivity and perceived usefulness is that the more a technology gives users a sense of control over the system used, the more the users will feel that the system can meet their needs and expectations, so they will feel that the system is beneficial for them. TikTok provides various interactive features for users, such as the opportunity to like the content they are interested in, making it easy to categorize videos according to their content and interact with other users. We formulate the following hypothesis:

H8: Interactivity (INT) influences users' perceived usefulness (PU) of higher education services on TikTok
Enjoyment is defined by Alharbi and Drew (2014) as the extent to which a person feels pleasure when using a system, aside from the performance resulting from using the system. According to C.-T. Chang et al. (2017), the enjoyment that a person feels is important in discussions regarding the adoption of e-learning because when a technology has provided a full sense of enjoyment, users will use the technology consistently. This is supported by Shiau and Luo (2013), who say that enjoyment has a strong influence on one's satisfaction and that satisfaction will motivate a person to use a system continuously (DeLone & McLean, 2016;Pereira & Tam, 2021). Many studies have observed that a person's satisfaction is an outcome that is expressed through emotions regarding the use of a technology (Liao et al., 2009;Tan et al., 2007). Student satisfaction is an important aspect that needs to be evaluated in educational programs because it is related to the quality of the system and student performance (Chang & Smith, 2008). In this study, we defined enjoyment as that which a user feels during the process of obtaining information. The enjoyment in question is a sense of pleasure that motivates a user to continue using the TikTok application as a platform to obtain higher educational content. When someone already feels pleasure and enjoyment from using a system, positive feelings are experienced, which can increase user satisfaction. Therefore, we propose the following hypothesis:

Research methods
In this study, a combination of quantitative and qualitative approaches was used (a mixed method). We used an online questionnaire and online interviews to collect the data. To test our questionnaire, we used a readability test and pilot study. The readability test was used to see whether the statements in the questionnaire were easy for the respondents to understand. It was conducted on 12 respondents who had used TikTok to obtain higher educational content in the past year (purposive sampling). The readability test results were used to improve the statements included in the questionnaire. Then, we conducted a pilot study on 30 respondents for 10 days from February 7 to 17 February 2022. The results from the pilot study were then entered into the SPSS application to determine the Cronbach's alpha (CA) value. The result was a CA value of 0.9, so the questionnaire link was distributed through LINE, Instagram, and Twitter. These social media platforms were chosen because they could provide active users and respondents representing various demographics in Indonesia. We collected the questionnaire data from 15 February 2022, to 15 March 2022. We obtained 954 respondents; however, only 629 provided valid and complete responses. A summary of the respondents' demographic data can be seen in Table 1.
This qualitative data collection was carried out through online interviews with 21 respondents who had used TikTok to obtain higher educational content in the past year (purposive sampling). The interviews only focused on the rejected hypotheses, as the accepted hypotheses were already believed to have a strong basis resulting from the conclusions of the quantitative data analysis and previous research. The online interviews were conducted for 2 weeks from March 15 to 29, 2022.

Analysis methods
The processing and analysis of the quantitative data were carried out through a series of stages, and several software packages were used to process the information. AMOS 26 was used for managing and analyzing the questionnaire data. IBM SPSS Statistics 26 was used to calculate the reliability of the questionnaire data. Finally, to obtain the average variance extracted (AVE) and composite reliability (CR) values, we used Microsoft Excel. Moreover, the results of the interview data were processed using the content analysis method, which pays attention to the repetition of words and phrases in the statements of the respondents. The interview data were also used to provide practical implications for TikTok application developers.

Research instruments
This study used an online questionnaire instrument consisting of three parts. The first part was a validation question to ensure that the respondent was a user of the TikTok application. The second part comprised demographic questions. The third part contained questions related to the measurement items based on 38 statements. The answers were categorized based on a Likert scale ranging from 1 to 5 (1 = strongly disagree, 2 = disagree, 3 = neutral, 4 = agree, and 5 = strongly agree). Table A1 describes the measurement items used in this study, and Appendix B lists the interview questions.

Measurement model
The measurement model consisted of a convergent validity test, a reliability test, and a discriminant validity test. The convergent validity test was obtained by calculating the loading factor, the CR, and the AVE. When a model has a loading factor value > 0.7, a CR > 0.7, and an AVE > 0.5, it possesses convergent validity (Hair et al., 2017). Table 2 shows the CR and AVE values.

Structural model
Structural model testing uses the goodness-of-fit (GoF) parameter to estimate the relationship between variables in a predetermined hypothesis (Hair et al., 2017). The criteria considered in GoF are chi-square fit statistics/degree of freedom (CMIN/df), the root mean square error of approximation (RMSEA), the normed fix index (NFI), the comparative fix index (CFI), the goodness-of-fit index (GFI), the root mean square residual (RMR), and the Tucker-Lewis index (TLI). The modification of the research model was carried out to make the GoF criteria value meet the cut-off value. Table 3 reveals how all the GoF criteria achieved good fit. Moreover, the coefficient of determination results show that all the endogenous variables have a strong effect size (PU = 0.996, SAT = 0.997, and CI = 0.852).

Hypothesis testing
This test uses a significance level of 5% and a two-tailed test. Consequently, accepted hypotheses will have a p value < 0.05, and rejected hypotheses will have a p value > 0.05 (Hair et al., 2017). The results of the hypothesis testing can be seen in Table 4, which led to the acceptance of eight of the hypotheses and the rejection of one.
This study found that confirmation has a significant effect on perceived usefulness (H1 is accepted). This result means that users have high expectations of TikTok in terms of obtaining higher educational content due to TikTok having a video recommendation system that is aligned with what its users are interested in and that helps them obtain higher educational content. During the interviews, 2 of the 21 respondents stated the following: "TikTok makes it easier to find higher educational content . . . " (Interviewee 12) and "Yes, I think the TikTok content presented is often informative" (Interviewee 15). TikTok is hosted an educational program in the form of   a competition with the hashtag #SamaSamaBelajar, with the aim of advancing education in Indonesia (Puspita, 2021). Examples of the #SamaSamaBelajar and #LearnOnTikTok educational programs can be seen in Figure 2. In addition, according to TikTok (2020), several universities have joined TikTok to provide higher education content.

Hypothesis
This study also found that confirmation has a significant effect on satisfaction (H2 is accepted). This finding is also supported by the questionnaire data, where 89.35% of the respondents reported feeling satisfied when using TikTok educational services. Through the interviews, it was found that 14 out of the 21 respondents were satisfied with TikTok with regard to obtaining higher educational content. TikTok users are generally more satisfied with videos that are shorter as they can be understood more quickly: "TikTok likes to show educational content in short videos; it makes it easier to digest and match the pace of learning" (Interviewee 6); "I like how TikTok only shows the main idea from a video rather than a complete video" (Interviewee 12); and "Satisfied because it is a short, concise, and clear video" (Interviewee 15). Moreover, this study has shown that the effect of perceived usefulness on continuance intention is significant (H3 is accepted). Respondents found TikTok useful and fun because it can be used to obtain appropriate information and make it easier to share higher educational content: "Yes, TikTok makes it easy to get personalized educational content because there is an algorithm that shows videos according to our preferences" (Interviewee 18). In addition, TikTok helps provide information that is easier to understand because of the short duration of its videos and the visualization of educational content through an approach that is easy to follow and that adds to the knowledge of its users in a fun way. This helps TikTok users recognize the benefits and usability value of continuing to use TikTok for higher educational content. Consequently, H4 is accepted. Based on the questionnaire, the majority of the respondents were satisfied with the performance of TikTok

Figure 2. Educational Content Hashtags on TikTok
(89.4%). In addition, 80% of the interviewees stated that they feel TikTok has met users' expectations in terms of accessing higher educational content.
However, this study did not find a significant effect of satisfaction on continuance intention (H5 is rejected). Several interviewees felt that the TikTok algorithm did not suit their needs. The TikTok application algorithm is very random, causing the application to be considered less useful in relation to learning needs: "The content that appears is quite random, so educational content doesn't always appear" (Interviewee 6). In addition, the duration of the videos provided by TikTok also reduces the usefulness value perceived by users: "If you watch a video that is divided into several parts, it's hard to find the next or another related part" (Interviewee 10) and "Videos on TikTok have a lot of parts because of their short duration" (Interviewee 8). Based on the interview results, it can be concluded that TikTok is only a trigger for learning.
Further, system quality was found to influence satisfaction (H6 is accepted). Based on the questionnaire results, the majority of users felt that the performance of TikTok is not slow (44.4%) and that TikTok is easy to use (51.2%). Moreover, TikTok has features and functionality that meet user needs (47.7%) related to higher educational content. Most users feel that the TikTok application has a good system quality and is fast when searching for educational content. Based on the interviews, several respondents felt that the performance of the TikTok application was quite good: "I think TikTok's performance is quite good unless there is indeed a connection problem" (Interviewee 17) and "So far, okay, it's never slow, and it provides content according to my preferences" (Interviewee 16). The performance and functionality of the TikTok application are essential aspects that affect its system quality value, which, ultimately, impacts the satisfaction felt by users.
Next, this study found that information quality has an effect on satisfaction (H7 is accepted). According to the questionnaire data, 89.35% of respondents were satisfied with regard to obtaining higher educational content using TikTok. In addition, 30.84% of the respondents accessed educational content through TikTok more than 6 days a week. Moreover, interactivity was found to have an influence on the perceived usefulness (H8 is accepted). The interview results revealed that most of the respondents felt that the features provided by TikTok were useful and easy to use. For example, the share content feature makes it easier for users to send educational content and obtain such material from their friends. As a result, users perceive that the features of the TikTok application are interactive. They feel they have control over the application and are provided with rapid responsiveness and appropriate two-way communication. This is also supported by the fact that 47.85% of the respondents agreed that they are happy with some of the features TikTok provides.
Finally, enjoyment was observed to have an effect on satisfaction (H9 is accepted). From the questionnaire data, we found that 62.16% of the respondents searched for higher educational content using TikTok because the content was easily accessible. In addition, 54.37% of the respondents agreed that they were happy to use TikTok to obtain higher educational content. Moreover, 47.85% of the respondents agreed they were delighted with some of the features provided by TikTok for acquiring and disseminating educational content.

Discussion
This study proves that enjoyment, system quality, information quality, and confirmation affect satisfaction, but perceived usefulness has not been found to affect satisfaction. This expands the research of Suzianti and Paramadini (2021), which was not able to prove that information quality affects satisfaction in e-learning. The results regarding the influence of system quality and information quality on satisfaction strengthen the research conducted by Pang et al. (2020) in the context of knowledge-sharing platforms and that of Halonen et al. (2009) in the context of a virtual learning environment. The results concerning the effect of confirmation on satisfaction strengthen the research performed by Daneji et al. (2019), and Zhang et al. (2020), and those relating to the effect of enjoyment on satisfaction support the findings of Akdim et al. (2022). However, this study has not been able to prove the effect of perceived usefulness on satisfaction, which is not in line with the research found by Suzianti and Paramadini (2021) but is in agreement with studies conducted by Alraimi et al. (2015), C.-C. Chang et al. (2015), Daneji et al. (2019), Hong et al. (2006), and Zhang et al. (2020). The results also show that confirmation and interactivity affect perceived usefulness. The effect of confirmation on perceived usefulness is in line with research conducted by Daneji et al. (2019) and Pang et al. (2020), and the influence of interactivity on perceived usefulness strengthens research conducted by Cheng (2020), Pituch and Lee (2006), Roca et al. (2006), Dağhan and Akkoyunlu (2016), and Cheng (2020) described how interactive features can make an e-learning system more useful, user-confirmed, and satisfying based on its functionality (Dağhan & Akkoyunlu, 2016;Roca et al., 2006). In addition, Lee and Lee (2019) observed how interactivity can affect users' perceived usefulness regarding a social media site. TikTok provides various features, such as those for sending content, commenting, liking, chatting, and more. The results concerning the effect of perceived usefulness and satisfaction on continuance intention are in accordance with the research of Daneji et al. (2019), Pang et al. (2020), and Suzianti and Paramadini (2021).
Finally, this research has contributed to the literature on the use of TikTok for education. This study also extends the ECM by adding system quality and information quality from the ISSM, and it can expand knowledge about the factors that encourage users to continue to use TikTok for education, especially in the context of the Indonesian population. To maintain and increase user satisfaction, TikTok must pay attention to confirmation, enjoyment, system quality, and information quality factors perceived by users. TikTok can expand its functionality and improve its features, such as subtitles, speed adjustments, length extensions, and playlists, to better support the learning process. In addition, TikTok needs to evaluate the algorithms that have been implemented to better suit the needs and preferences of users and to aid the learning process. Educational institutions should also investigate aspects that can support students in adopting technology for education. To improve the quality of distance learning, academic institutions can adopt TikTok as a learning tool. Based on the questionnaire and interview data, most users use TikTok for e-learning due to its attractive way of presenting content, the discussion of material that is easy to understand, and the inclusion of up-to-date information.

Conclusion
This study found that confirmation influences perceived usefulness and satisfaction. Consequently, the more users feel that TikTok provides interactive features, the more benefits they will associate with using it. Further, information quality, system quality, and the enjoyment perceived by users affect user satisfaction regarding TikTok. Finally, the continuation intention to use TikTok is influenced by the confirmation of the users' initial expectations and the level of the perceived usefulness of TikTok for accessing higher education content. This study has several limitations. For example, most of the respondents were domiciled in the Greater Jakarta region, and, thus, they do not necessarily represent TikTok application users throughout Indonesia. Future research could also examine other social media applications, such as YouTube, Instagram, and Twitter, that can also be used for accessing higher education content.