The Correlation Between Social Media Usage in Academic Context and Self-Efficacy Towards TPACK of Prospective Science Teachers in Indonesia

The purpose of this study is to investigate the relationship between the frequency of Social Media Usage (SMU) in an academic setting and Self-efficacy beliefs towards TK, TCK, TPK, and TPACK of Indonesian prospective science teachers. This research is quantitative based-research design using a self-administered survey. The research was conducted during the second semester of the academic year 2018/2019 from October to November 2018 in the Faculty of Mathematics and Natural Science of a State University located in Semarang City, Indonesia. The sample consists of 217 Indonesian prospective science teachers from the science and Biology Education Department. The result shows that the average Social Media Usage frequency has a statistically high correlation with TK Self-efficacy, TPK, and TPACK. However, in general, it does not correlate with TCK. Second, Social Media Usage for Download Media (DM), Searching Information (SI), and Entertainment and Motivation (EM) generally have a statistically medium correlation with TK, TPK, and TPACK for both male and female participants. Third, Social Media Usage for professional development (PD) has a medium correlation with TK and high correlation with TPK and TPACK. This study implies that training focuses on the application of social media in teaching, and learning should be integrated to improve Indonesian Prospective Science Teachers' TPACK.


INTRODUCTION
Professional Science teachers are crucial to improve the quality of education in Indonesia. They must have four competencies, including pedagogy, personality, professional, social, where they should not only master in their subject matter but also technology and pedagogy. The framework that can show a teacher's knowledge related to pedagogy, technology, and content is TPACK. TPACK's theoretical framework, which was developed by Mishra & Koehler (2006), stands for technological pedagogical content knowledge (TPACK) (Figure 1).
Teachers should have good TPACK to effectively integrate technology into their teaching. However, some research has shown that most of Indonesian Science teachers did not efficiently use technology in their teaching (Kazu & Erten, 2014;Semiz & Ince, 2012;Smeets, 2005). For example, Kamulung (2017) investigated the level of integration of information technology by Indonesian Biology, Chemistry, and Physics teachers in grade XI high school using the Technology Integration Matrix (TIM) method. He found that the ability of Indonesian Science teachers to Integrate technology in learning Science is only 55% in a medium category between the interval 41%-60%. Furthermore, another researcher found that although most of the Indonesian teachers had the technology to support students learning, their ability to use technology was still relatively low, where most teachers still used conventional methods based on the results of the questionnaire, observations, and interviews (Styaningrum, 2016). There are some complex factors affecting teachers' use of technology in K-12 classrooms such as computer proficiency, computer availability, and technical support such as electricity and hardware facilities (Eteokleous, 2008;Hew & Brush, 2007;Inan & Lowther, 2010). In Indonesia, some researchers show that the main factors are science teachers' lack of skills, knowledge for effective technology use, and management issues because they are common use of social media includes searching and studying in collaborative online activities, download media, sharing information, and entertainment (Wheeler, Yeomans, & Wheeler, 2008;Al-Daihani, Jumanah, & Sara, 2018). In teacher education, teacher educator used social media for motivating teacher candidate to collaborate with a cooperative teacher or mentor; providing a friendly atmosphere to encourage reflection to identify and address challenges of teaching practice; giving feedback and continue following up on their progress in teaching practice (Pitiporntapin & Lankford, 2015). In the actual setting of the teaching and learning process, in-service teachers are also reported integrating social media such as Facebook (Blonder & Rap, 2017) and YouTube (Blonder, Jonatan, Bar-Dov, Benny, Rap, & Sakhnini, 2013) in their science classrooms. Although flourishing research evidence shows the potential of social media, there is little research connect between social media use affect TPACK self-efficacy of prospective teachers.
This study has some potential significance, both theoretically and practically. By investigating the correlation between social media and TPACK, future researchers might consider some factors of social media usage that need attention to design the program or intervention that focuses on developing equitable TPACK for all PSTs. This is because prospective teachers who have low self-efficacy towards technology integration tend not to use it in their future teaching practice. Meanwhile, integrating technology into the Sciences classroom is the requirement of National Science Standard not only in Indonesia but also in many countries. Practically, the result of this study would be significant for policy and practice. Teacher educators can also consider the finding of designing the best practice of a professional development program (PDP) through ICT course evaluation. For example, social media can be integrated to improve their KTT Self-efficacy.
There are some operational definitions of the key terms of this research. TPACK framework stands for technological pedagogical content knowledge. It has three essential elements, including CK (Content Knowledge), PK (Pedagogical Knowledge), TK (Technological Knowledge). The intersection between them is PCK (Pedagogical Content Knowledge), TCK (Technological Content Knowledge), and TPK (Technological Pedagogical Knowledge) and TPACK. Since this study would focus on the element related to technology, CK, PK, and PCK were excluded from the analysis. Social Media Usage (SMU) in this project would only focus on five constructs of SMU in the academic life of university students including Searching and Studying (SS), Download Media (DM), Sharing Media (SM), Professionalism Development (PD), and Entertainment and Motivation (EM). SMU in an informal setting or leisure activity was excluded.

TPACK Framework
Technological Knowledge (TK) generally includes knowledge on a wide variety of different technologies from simple technology such as a pen to advance technology (Archambault & Crippen, 2009;Ozden, Mouza, & Shinas, 2016;Schmidt et al., 2009). However, TK in the TPACK framework only specifically includes technology associated with digital platforms and computers (Mishra & Koehler, 2006). It also consists of the knowledge of using computer software (basic office programs, operating system, and software related to instruction), using hardware (projection device, interactive whiteboard, etc.), web or the internet research devices (Chuang, 2013;Handal, Campbell, Cavanagh, Petocz, & Kelly, 2013) and using communication platform (social networks, email, forums, chat tools) (Chai, Koh, & Tsai, 2013). Teachers with a high level of technical knowledge tend to use technology more efficiently, which also has an implication in their classroom teaching and learning (Abbitt, 2011;Harris, Mishra, & Koehler, 2007).
Knowledge of Teaching with Technology (KTT) in TPACK Framework is the simplification of three elements in TPACK Frameworks, including Technological Content Knowledge (TCK), Technological Pedagogical Knowledge (TPK), and Technological Pedagogical Content Knowledge (TPACK). The three components are simplified into KTT because PSTs generally interpreted the items related to technology as being conceptually similar (Koh, Chai, & Tsai, 2010). Firstly, in the Science education context, TCK refers to the effectiveness of teaching science with technology integration (Schmidt et al., 2009). It implies that science teachers should master not only their knowledge of science concepts but also an understanding of how it can change with technology that more appropriate to be integrated (Mishra & Koehler, 2006;Schmidt et al., 2009). Secondly, TPK refers to an understanding of how technology can affect the teaching and learning process. For example, teachers should be able to consider the challenges, constraints, and affordance of technology by considering their teaching environment and curriculum (Harris, Mishra, & Koehler, 2007). This also includes what kind of technology more suitable for a particular teaching strategy, classroom management, and students' differences. Thirdly, TPACK is the presentation of science by being integrated with a pedagogical approach by using various technologies (Chai, Koh, & Tsai, 2013).

Social Media Usage (SMU) and TPACK
Some researchers argue that social media has a beneficial impact on the academic life of university students (Ahmed, Ahmad, Ahmad, & Zakaria, 2018;Al-Rahmi & Zeki, 2017;Ali, Yaacob, Endut, & Langove, 2016;Arquero & Esteban, 2013). For example, in social media, prospective teachers can contribute to express their idea or perspectives to the group debates to respond to the education issues (Mudaly, Morgan, van Laren, Singh, & Mitchell, 2015). Online chatting in a WhatsApp group can be used as a tool to support prospective teachers' professional development through sharing their experience in a group chatting and support emotionally each other (Cansoy, 2017). Kivunja (2013) supported that integrating social media is to improve their digital pedagogy as the preparation for them teaching using digital technologies. "Digital Pedagogy in Pre-Service Higher Education is crucial to Better Prepare Teachers in the new classroom because children in their classrooms will be Digital Natives, with skills for digital fluency" (Kivunja, 2013). Students have grown up in a world surrounded by and using computers. In the actual setting, science teachers also integrating social media such as YouTube (Blonder, Jonatan, Bar-Dov, Benny, Rap, & Sakhnini, 2013), Facebook forum (Blonder & Rap, 2017;Köseoglu & Mercan, 2016), Google docs, Skype, Social Networks and Wikis (Ahern, Feller & Nagle, 2016), to support students learning. Social media is also integrated into Moodle to support distance learning (Buus, 2012). Lau (2018) investigates the correlation between SMU in informal settings (outside of classroom context), including for Media Sharing, Internet Searching, and Video Gaming with TPACK level of pre-service teachers with Path Analysis. He found that Media Sharing is a useful predictor of Technological Knowledge (TK) and Knowledge of Teaching with Technology (KTT). In contrast, Internet Searching and Video Gaming do not correlate with any components of TPACK. However, the research is conducted in a broad context and only focuses on SMU in an informal setting. Meanwhile, some research shows that University students in teacher education using not only social media to media sharing, internet searching, and video gaming but also other activities such as to download media, entertainment, improve motivation, to improve their professionalism as prospective science teachers. Therefore, further research that investigates the relationship between SMU in an academic context and their TPACK Self-efficacy in Science education needs to be conducted in academic settings.

Gaps in the literature
Although previous research shows, TPACK correlates with the use of technology level (Kazu & Erten 2014), educational internet use (Sahin, Celik, Akturk, & Aydin, 2013), attitude towards computer (Baturay, Gökçearslan, & Sahin, 2017), attitude towards web-based instruction (Kavanoz, Yüksel, & Ozcan, 2015) and SMU in informal setting or leisure activities (Lau, 2018), no research connect it with SMU in an academic setting or school activities. Lau (2018) suggested that extended research of SMU in an academic setting and specific domain should be identified as an area for future research, especially in the context of Science Education. Previous studies (Kaya, Kaya, & Emre, 2013;Lin, Tsai, Chai, & Lee, 2013;Pamuk, Ergun, Cakir, Yilmaz, & Ayas, 2015;Schmidt et al. 2009) supported that TPACK could be domainspecific. Therefore, in this study, this research would analyze the association between prospective science teachers' demographic characteristics, SMU in their school activities, and their TPACK Self-efficacy in the Indonesian context.

Research Question (RQ)
Considering the importance of TPACK Self-efficacy, and the wide use of social media among prospective science teachers supported by given gaps of literature, this study proposes to examine issues related to TPACK surveys with a study of Indonesian pre-service teachers. The research question addressed in this study is: What is the relationship between the frequency of Social Media Usage (SMU) in an academic setting and Self-efficacy beliefs towards TK, TCK, TPK, and TPACK of prospective science teachers in Indonesia? 2. METHOD

Research Design
This research is quantitative based-research design. The research was conducted using a self-administered survey that provides a quantitative or numeric description of the opinions of a population (Creswell, 2008;Robson & McCartan, 2016).

Time, Location, and Participants
The research was conducted during the second semester of the academic year 2018/2019 from October to November 2018 in the Faculty of Mathematics and Natural Science of a State University located in Semarang City, Central Java Province, Indonesia. The sample of this research study consists of 217 Indonesian prospective science teachers from approximately 250 population invited to participate in this study. The population includes PSTs from the science education Department and Biology Education Department. The response rate of the survey is 86.8 %. The high response rate might because of the reward offered before the study. PSTs have been studying science to become secondary school teachers majoring in Science Education and Biology Education at various levels. The second year of 112 Bachelor degree students (51.61%) was not experience with ICT courses, which includes computer and technology integration in the Science classroom because it is offered in the fifth semester. In contrast, 105 Master's degree students (48.38%) were experienced with ICT courses that are expected to be better informed with technology application in their teaching. The majority of participants were female, with 124 students (58%), and a smaller proportion of male students responded to the survey with 93 students (42%). The unbalance ratio between females and males might because the research was conducted in the faculty where strong female-bias in Indonesia.

TPACK Self-Efficacy Scale (TPACK-SeS)
The first research instrument to evaluate TPACK Selfefficacy used in this research was explicitly developed for prospective science teachers (PSTs) by Bilici, Yamak, Kavak, & Guzey (2013). The item has been validated with 808 PSTs from teacher preparation institutions at 17 different universities in Turkey. The items are valid and reliable with the result of item-total correlation coefficients of Exploratory factor analysis (EFA), and Confirmatory factor analysis (CFA) showed a good fit. The item also reliable since Cronbach's alpha coefficient was found to be high. The format for measurement of this instrument was modified from a 100-point rating scale to a 5-point Likert scale (1=Very low, 2=low, 3=medium, 4=High, 5=Very high). Content Knowledge (CK) construct, as well as Pedagogical Knowledge (PK) construct and the intersection, was deleted from the survey because the primary focus is the technology constructs of TPACK. Technological Knowledge (TK) Self-efficacy has six constructs, and Knowledge of Teaching with Technology (KTT) Self-efficacy has 14 constructs. KTT is the composite of Technological Pedagogical Knowledge (TPK), Technological Content Knowledge (TCK), and Technological Pedagogical Content Knowledge (TPACK). The three components are simplified into KTT because prospective teachers generally interpreted the items related to technology as being conceptually similar (Koh, Chai, & Tsai, 2010). The questionnaire was translated by the author into Bahasa Indonesia (the national language of Indonesia).

Social Media Usage in Academic Context
The second research instrument, SMU in an academic context, is developed by looking at the common use of social media both theoretically and practically. The former, extensive literature review is conducted by reviewing related peer-reviewed articles to discuss Media and Technology usage. We found that the instrument to identify the frequency of social media use has been initially developed by Lau (2018). However, the instrument focus on developing an instrument to measure SMU in an informal setting which not fully reflect SMU in academic life. Therefore, an extended literature review is conducted to explore students' motivation and purpose to use social media in higher education (e.g., Ali, Yaacob, Endut, & Langove, 2016;Bal & Bicen, 2017;Balakrishnan, Teoh, Pourshafie, & Liew, 2017;Arquero & Esteban, 2013;Gupta, 2014;Rosen, Whaling, Carrier, Cheever, & Rokkum, 2013) and in teacher preparation institution (Carpenter & Krutka, 2015). The latter, the common use of social media, is conducted not only theoretically but also practically. Some graduate students were asked what kind of practical activity in Social media that they do during their study. This study was conducted by interviewing three pre-service science teachers enrolled in a teacher preparation program in Indonesia and two graduate students at Monash University Australia. They were asked the common use of Social media in their academic life. The result was rewritten to create the questions.
Based on the result of the literature review and elicitation study, the author them modify SMU in an informal setting, which is developed by Lau (2018). The instrument has five constructs includes five including Searching and Studying (SS), Download Media (DM), Sharing Media (SM), Professionalism Development (PD), and Entertainment and Motivation (EM). Media Sharing, which is developed by Lau (2018), initially has four subitems, including the use of Social media to watch TV shows, to watch video clips on a computer, to download media files, and to share media files on a computer. In the modified instrument, we remove SMU to watch TV because it is not related to the academic setting and teaching practice. Professional Development (PD) is added as an additional item because prospective teachers also use social media to enhance their professionalism. The scale of SMU frequency uses a five-point scale (never, rarely, sometimes, often, and always). The format of measurement slightly different from Lau (2018) that uses ten-point frequency scales of SMU in an informal setting.
This newly-modified instrument was presented to a group of graduate students studying at the Faculty of Education, Monash University (N=3), and doctoral students (N=1) to be reviewed for content validity. The results were reviewed, and slight alterations were made in language, and questions were added or dropped to strengthen the sub-items of the construct. Minor alterations were made, and the final instrument was used to collect data.

Data Collection
Ethical approval was obtained from the Monash University Human Research Ethics Committee (MUHREC) with approved number 2018-17330-24853 to ensure that the research follows research ethics principles such as informed consent, privacy, and confidentiality. Before the survey was conducted, the institution where the research was conducted was given a consent form to obtain a letter of permission. The prospective participants were also given information implicitly related to the purpose of research, what they need to do, how long they need to complete the survey, the potential benefits, and risk. The survey was anonymous, so they will not explicitly identify themselves to make sure the data remains confidential. However, they need to input their email address before completing the survey to avoid double response.
Data were collected through a web-based survey using google form because it is free and efficient to collect information (Djenno, Insua, & Pho, 2015). The researcher shares the link of a web-based survey to all of the selected students with the help of lecturers in the Faculty of Mathematics and Natural Sciences who teach the participants. After the link was shared using social networks (WhatsApp group), the participants completed three parts of the survey sequentially. First, they completed their demographic characteristics (gender, level of education, major, and age) in a multiple-choice format. Some alternative answers are provided for each question. Participants can only choose one answer for gender, level of education, major, and age. The second and third part of the survey is TPACK Self-efficacy and SMU in the academic context, respectively. They can only respond to one answer for each item. Fifteen minutes were granted for the students to complete the survey. Participants had a two-week timeframe in which to complete the survey.
After the survey was conducted, the data was downloaded from google form and transferred into Microsoft Excel 2016. All of the students' responses were entered as numerical values so that the computer can read and analyze the data. Therefore, the first step was coding, which aims to transform non-numerical information into numerical information. For example, the self-efficacy scale response, very high, was coded as 5, and 1 for very low. The data were organized into a cross-tabulation.

Data Analysis
The data analysis was Inferential Statistics. The examination of Pearson's correlation was used to analyze whether the PSTs' differences in perception of TPACK were related to their SMU. The correlational analysis was presented in a scatterplot. The multiple regression analysis was used to test for the significance of the relationships. The statistics of the multiple regressions of importance in determining the relationship will be the Pearson r, pvalue, and effect size (Gall, Gall, & Borg, 2003). Spearman's rank correlation coefficient (r) was counted to see the linear trend using Microsoft Excel. The direction of correlation can be positive, negative, and no correlation (Mertens, 2016). A positive correlation means that the two variables move together in the same direction. A negative correlation means that the two variables differ inversely in the opposite direction. If the correlation coefficient is near zero, no relationship exists. The interpretation to assess the strength of the correlation coefficient and the direction (Cohen, 1992) are presented in Table 1.

Figure 2
The correlation between the frequency of Social Media in an academic setting and TK Self-efficacy Figure 3 The correlation between the frequency of Social Media in an academic setting and TCK Self-efficacy Table 1 Interpretation of Spearman's rank correlation coefficient (r) r-value Interpretation r =.10 to .29 or r = −.10 to −.29 low r = .30 to .49 or r = −.30 to −.49 medium r = .50 to 1.0 or r = −.50 to −1.0 high The p-value indicates how confident each independent variable has some correlation with the dependent variable. In other words, the p-value indicates whether the relationship found in the r is statistically significant. When p<.05, the relationship is statistically significant.

RESULT AND DATA ANALYSIS
We found three important findings when the correlation between all construct of SMU and TPACK are analyzed based on gender (Table 2). First, SMU for Searching and Studying has a high statistically significant correlation with TK Self-efficacy, TPK, and TPACK (r>0.5) among male participants but has a medium correlation among female participants (0.3<r<0.5). Second, SMU for Download Media (DM), Searching Information (SI), and Entertainment and Motivation (EM) generally have a statistically medium correlation with TK, TPK, and TPACK for both male and female participants. Third, SMU for professional development (PD) has a medium correlation with TK and high correlation with TPK and TPACK. Interestingly, this research found that TCK only has a low correlation with SMU for PD for both male (r=.22, p<.05) and female (r=.13, p<.05), but it does not correlate with another construct of SMU including SS, DM, SI, and EM. Figure 2 shows that the average of Social Media Usage frequency has a statistically high correlation with Technological Knowledge Self-efficacy of Prospective Teachers (r=-.53, p<.01).
In contrast to TK, Figure 3 shows that the average of Social Media Usage frequency has no correlation with the Technological Knowledge Self-efficacy of Prospective Teachers (r<.1).

DISCUSSION
The purpose of this paper is to investigate the relationship between SMU in an academic setting and TPACK. This research found that the frequency of SMU, in general, has a statistically high correlation with TK (r=-.53, p<.01), TPK (r=-.58, p<.01), and TPACK (r=-.57, p<.01). However, the frequency of SMU does not correlate with TCK (r<.1). The result of this study aligns with the research of Lau (2018) investigating the relationship between SMU in an informal setting and TPACK. He found that although the relationship is not strong, TK correlates with Media Sharing (MS) (r=.237, p<.05), Internet Searching (IS) (r=.124, p>.05), and Video Gaming (VG) (r=.204, p<.01). However, the result of this study is slightly different from his findings. SMU in an informal setting is moderately correlated with all construct of Knowledge of teaching with technology (KTT), which is the composite of TCK, TPK, and TPACK. It might have the effect of the context of formal and informal setting usage of SMU. Some similar research also supports the result of this study. Who found positive relationships exist between the knowledge dimensions in the TPACK and the self-efficacy beliefs in educational Internet use (Sahin, Celik, Akturk, & Aydin, 2013;Kazu & Erten, 2014;Kavanoz, Yüksel, & Ozcan, 2015;Bingimlas, 2018). A study conducted by Kavanoz, Yüksel, & Ozcan (2015) also revealed that the level of general self-efficacy among participants regarding TPACK is associated with their attitudes towards web-based instructions. Some research reported that science teachers who integrate social media such as YouTube (Blonder, Jonatan, Bar-Dov, Benny, Rap, & Sakhnini, 2013), Facebook forum (Blonder & Rap, 2017;Köseoglu & Mercan, 2016), Google docs, Skype, Social Networks and Wikis (Ahern, Feller & Nagle, 2016), to support students learning is Figure 4 The correlation between the frequency of Social Media in academic setting and TPK Self-efficacy Figure 5. The correlation between the frequency of Social Media in an academic setting and self-efficacy towards Technological Pedagogical and Content Knowledge found an increased efficacy towards TPACK. Therefore, integrating social media in teacher education institutions would be beneficial for PSTs.
When further analyzed, the sub-items of SMU in academic settings according to gender, Searching, Download, and Sharing media have a positive correlation with TK, TPK, and TPACK for both males and females with the size of correlation among female PSTs is weaker than males. SMU for professional development also has a high correlation with TPACK Self-efficacy for both males and females. Professional development in this context refers to social media to learn new technological tools for teaching, to create learning media, to update information related to current new technology for teaching, to join distance learning workshop or online short course and to get information from professional teachers about how they apply ICT in their Science classroom. Then, SMU for entertainment and motivation (EM) also has a positive correlation with their TK, TPK, and TPACK for both males and females. EM in this research refers to the use of social media to read online news, listen to music, play Educational Games, watch online videos for entertainment and chat in Social media to improve learning motivation from others. The existence of gender difference as the moderator of the relationship between age and TPACK self-efficacy might be the effect of the ICT-user profile. ICT user profiles are defined as "a framework which is used to identify and categorize students based on how frequently they use ICT" (Tomte & Hatlevik, 2011). An increased level of self-efficacy in ICT is related to both an increased level of educational use and leisure use of ICT.

CONCLUSION
The average SMU frequency has a statistically high correlation with TK (r=-.53, p<.01), TPK (r=-.58, p<.01), and TPACK (r=-.57, p<.01). However, the frequency of SMU does not correlate with TCK (r<.1). When the correlation was analyzed based on gender, we found three important findings. First, among male participants, SMU for Searching and Studying has a high statistically significant correlation with TK Self-efficacy (r=.51), TPK (r=53), and TPACK (r=.50), but has a medium correlation among female participants with TK Selfefficacy (r=.37), TPK (r=.42), and TPACK (r=.42). Second, SMU for Download Media (DM), Searching Information (SI), and Entertainment and Motivation (EM) generally have a statistically medium correlation with TK, TPK, and TPACK for both male and female participants (r = .30 to .49). Third, SMU for professional development (PD) has a medium correlation with TK for both males (r=.47) and female (r=.36). However, PD has high correlation with TPK and TPACK (r = .50 to 1.0). Interestingly, this research found that TCK only has a low correlation with SMU for PD for both males and females (r=.14), but it has no correlation with another construct of SMU(r<.1). In this research, generally, Social Media Usage for all purposes has a high correlation with Technological Knowledge (r=.53), Technological Pedagogical Knowledge (r=.58), and TPACK (r=.57).

LIMITATION
We exclude some construct of TPACK that is not related to technology, including Content Knowledge (CK), Pedagogical Knowledge (PK), and the intersection, Pedagogical Content Knowledge (PCK). Thirdly, the items in the instrument to measure SMU in an academic context were the result of modification from SMU in an informal setting from previous researchers, but the modified items were not tested for validation due to time constraints to do the project.

RECOMMENDATION
1. Future research should investigate TPACK among in-service science teachers in Indonesia who enroll in Pendidikan Profesi Guru/PPG or Graduate Certificate for Teacher Profession. Some topics that can be explored are: (1) the effectiveness of professional development of technology and social media (Facebook, Twitter, YouTube) among in-service teachers in Indonesia to improve their TPACK competence (2) the effectiveness of Graduate Certificate for Teacher Profession in Indonesia to improve in-service teachers' competencies in technology integration.
2. The correlational research between SMU and TPACK in this research is only conducted in the context of Science Education, and the participants of this research are PSTs from the Department of Science Education and Biology education in one teacher education institution in Indonesia. Future investigation of TPACK should focus on other science domains such as physics and chemistry in more than one institution.
3. Future research can compare TPACK across crosscultural contexts between Indonesia and other countries in South East Asia, such as Indonesia and Singapore, Malaysia, or even with countries in the different continents to identify the possibility of cultural differences. This is because most of the research is conducted in one country and a lack of research comparing PSTs' TPACK across different cultural backgrounds (Alqurashi, Gokbel, & Carbonara, 2017).

Implication for Practice
Since social media usage has a positive effect on PSTs' TK Self-efficacy and KTT, training focuses on the application of social media in teaching and learning should be integrated to improve their TPACK. In this research, we especially found that SMU for Professional Development correlates with all construct of TPACK related to technology. Therefore, the lecturer can integrate professional development that is designed specifically for PSTs. For example, the introduction could be designed as an advanced program for guiding teachers to adopt adequate social media-based learning materials to learn new technological tools for teaching, to create learning media, to join distance learning workshops or online short courses. For example, advanced training programs could introduce and guide PSTs to utilize social media-related apps such as Facebook, YouTube, Twitter in their classroom. Students' motivation for learning Science might be further enhanced. Recently, science teachers who integrate social media such as YouTube (Blonder, Jonatan, Bar-Dov, Benny, Rap, & Sakhnini, 2013), Facebook forum (Blonder & Rap, 2017;Köseoglu & Mercan, 2016), Google docs, Skype, Social Networks and Wikis (Ahern, Feller & Nagle, 2016), to support students learning is found an increased efficacy towards TPACK. Therefore, integrating social media in teacher education institutions would be beneficial for PSTs.

ACKNOWLEDGMENT
The research in this paper is funded by the financial support of the Ministry of Finance Republic of Indonesia (Lembaga Pengelola Dana Pendidikan) with allocation for thesis or Research Project of master's degree program (Master of Education) according to the Scholarship contract No. PRJ-6893/LPDP.3/2016. We also acknowledge the lecturers of Unit EDF5614 (Research Project in Education) in the Faculty of Education at Monash University that provide guidance and support for the research project.