Tri Kaya Parisudha Based Time Token Learning Model Influences Students’ Speaking Skills

The problem that underlied in this research was that the students speaking skills were not optimal, the learning model used was less varied, the learning atmosphere in the classroom was less arousing the activeness of students to participate. This study aimed to analyze the influence of Tri Kaya Parisudha based Time Token learning model on the fourth-grade speaking skills of elementary school. The study was a quasi-experimental with Posttest-only No-Treatment Control Group Design. The population of the study was all grade IV elementary schools, namely 16 classes and 495 total of students. The research sample was determined by using cluster random sampling. In this study, the samples were obtained 66 students. Speaking skills data were obtained using structured observation. The obtained data were analyzed using polled variance t-test. Based on data analysis, obtained tcount = 7.042 and ttable (significant level 5%) = 2,000 with dk = 64, then H0 was rejected. So, it is concluded that there was a significant influence of Tri Kaya Parisudha based Time Token learning model on the speaking skill of the fourth grade elementary school. This model can be used as a reference by the teacher in the learning process to improve speaking skills.


Introduction
Now a days, globalization is entering a new era, namely the industrial revolution 4.0 which can be seen from the rapid advancement of the digital system. This revolution fundamentally changes the way people think, live, and affect various sectors of life (Prasetyo & Trisyanti, 2018), it is line with Sandu's opinion (Prasetyo & Trisyanti, 2018) that globalization refers to a system interconnection of economic and social condition. One of them is the education system in Indonesia. The changes of globalization cannot be avoided so the high-quality human resources are needed to compete globally. Improving the quality of human resources from primary and secondary education to college is the key to follow with the development of 4.0 industrial revolution, so it is necessary to develop human resources from the humanities side to suppress the negative impact of technology developments (Lase, 2019;Prasetyo & Trisyanti, 2018). The success of education field can be realized if the curriculum is aligned with global challenges, advances in science and technology, and the needs of society (Sagala, 2014). The curriculum in Indonesia is changed to meet the needs of society in the future. Now the current curriculum is the 2013 curriculum. Curriculum 2013 learning focuses more on three domains, namely cognitive, affective, and psychomotor (Astuti et al., 2018) but in reality, learning still focuses on the cognitive domain and pays less attention to students' speaking skills. Besides, the most visible implication of the 2013 curriculum implementation especially at the elementary school level is the use of integrated thematic learning (Sari, 2015). Integrated thematic learning is an approach with a distinctive characteristic by combining various competencies from all subject content into a theme. The content of Indonesian lessons is one of the lesson contents in an integrated thematic, it is in line with (Daryati, 2019) that argues Indonesian learning content requires the students to master four language skills, namely listening skills, speaking skills, writing skills, and reading skills. One of the language skills that must be mastered is speaking because communication and interaction at school and society are done a lot. Thus, speaking skill is a basic principle for initiating communication with other people. Speaking is the ability to pronounce an articulated sound to express or say opinions, feelings, and ideas to others (Tarigan, 2015;Khair, 2018). However, in reality, many people still underestimate speaking skills. It is because something in mind can not be clearly pronounced and conveyed to the listener. As the opinion expressed by (Rusli, 2019) states that most people think that speaking is easy but there are still many people who have not mastered speaking skills. Thus, paying attention to the speaking skill is important. Speaking skills can be acquired and mastered through practice and a lot of exercises.
The problem is on children's confidence because they are afraid to stand in front of many people (Kerta, 2019). Some children are just silent, sweating cold, and forgetting everything when they are speaking in front of people, as well as they are reluctant to express their opinions or even ask questions. As a result, children's speaking ability is still low. As stated by (Yenidar, 2017), low speaking skills of students are due to a lack of self-confidence and student's motivation in speaking activities that cause students afraid and embarrassed to express their opinions. If this problem is not covered, it will make speaking ability of children low, the children become less confident to express opinions and themselves. Based on the results of preliminary observations and interviews conducted to each class teacher of Public Elementary School (SD Negeri) Cluster III Kuta Utara, especially the content of Indonesian subject in the fourth grade, from seven elementary schools in this cluster, it was seen that the speaking skills of students were not able to develop optimally that appeared during the thematic learning of Indonesian subject content from 495 of sixth grade students, 54.95%, namely 272 students had not been able to express opinions and retold a reading. It was because students were lack on vocabulary, students felt afraid when they were asked to give opinion, and students felt insecure when they spoke in front of the class. The low mastery of students speaking skills was due to the use of conventional models, the lack of variation in the learning process which made it boring and tedious.
Conventional learning places the teacher as the center of learning that makes the student activeness low to participate in learning. The previously described problems illustrate the importance of implementing educational programs that focus on speaking skills. Therefore, teachers play an important role in class management and apply innovative learning models as well as familiarize students to be more active during learning so students' speaking skills can be developed. Joyce (Marliani, 2015) reveals a learning model, namely planning that is used as a guide when designing learning to achieve the desired goals. The learning model applied in this research was Tri Kaya Parisudha based time token model. The time token learning model is the implementation of democratic learning with students as subjects (Kurniasih & Sani, 2016). In the learning process, students are the main concern and always participate actively. This Time Token learning model can hone and optimize speaking skills so students do not dominate a conversation. (Wijayanti et al., 2019) in their study state that the time token learning model can train children's speaking and social skills at the same time. (Sahara, 2019) states that speaking skills will develop better because there is an active communication among friends, students, and teachers. Thus, students' confidence and vocabulary begin to appear and increase in school along with the problems or materials given teacher which means that the use of the time token learning model make students optimize and hone their speaking skills gradually. In this model, the teacher distributes several speaking opportunity vouchers with a limit of ± 30 seconds to each student (Huda, 2017). The student must give the coupon to the teacher before starting to speak. One coupon is valid for one speaking opportunity. If the obtained coupons run out, the students are not allowed to talk. On the other hand, students who still have a coupon are required to speak until the coupon runs out. There are several stages of implementing time token learning.
There are several steps of conducting time token learning model step developed by Riyanto in a study of (Hamidah et al., 2017), namely, conveying learning objectives, creating a discussion section, giving problems to students, and distributing several speaking opportunity coupons. Before speaking, students give coupons to the teacher. The advantages of Time Token learning model in improving speaking skills are revealed (Lestari et al., 2018) in their research that the time token learning model can optimize students' speaking skills. According to (Huda, 2017), the advantage of the time token learning model is that it prevents students from dominating the conversation and those who are silent, increasing student activity during the learning process. It develops students' skills in communicating and expressing ideas. It makes students accustomed to give input and have an open mind to criticism. It encourages students to solve problems. This is supported by (Sahara's study, 2019) which states that the time token learning model affects the speaking skills of fifth grade elementary school students. Learning using the time token model can create a relationship between stimulus and response to speaking skills and expressing opinions so that all students become more active, motivated, and create enjoyable learning. A study conducted by (Pramana & Suarjana, 2018) also reveals that there is an influence on science learning outcomes between groups of students who are taught using the time token learning model assisted by video media and groups of students who are taught using conventional learning. The implementation of learning using Time Token model focuses more on student initiative and activeness in learning. In learning, the teacher serves as a facilitator and motivator. Based on a similar study, it can be concluded that the time token learning model has a positive influence on students' speaking skills so it is good to use it for developing students' speaking skills.
In this study, to obtain maximum results, the Time Token learning model was modified based on the concept of Hinduism, namely Tri Kaya Parisudha. Tri Kaya Parisudha comes from the word Tri which means three, Kaya which means behavior and Parisudha which means clean, good, and holy. Thus, Tri Kaya Parisudha is three human behavior that is purified in the form of thoughts, words, and deeds (Mahardi et al., 2019). Tri Kaya Parisudha is a concept about the basic way of behaving well, namely good thoughts, making good words, which ultimately leads to good actions. The concept of Tri Kaya Parisudha is very important to be implemented in elementary schools to grow Manacika, Wacika, and Kayika and produce individuals with character . In instilling Tri Kaya Parisudha, the teacher directs students to see something positively, to themselves, others, and environment. Furthermore, students are directed to say honestly and politely as well as to say something well and clearly. Then, students are directed to behave orderly and politely, do something based on their obligations, not being selfish, do not disturb others, and help each other, especially in solving problems in groups. Growing Tri Kaya Parisudha is not only a direction from the teacher to students but the teacher must be a role model in implementing Tri Kaya Parisudha .
Based on a description of the Time Token and Tri Kaya Parisudha models and their advantages, the objectives of study were formulated, namely, to analyze the significant effect of Tri Kaya Parisudha based Time Token learning model on speaking skill in the fourth grade elementary school Cluster III North Kuta academic year 2019/2020. Tri Kaya Parisudha based Time Token learning model makes students discuss with groups to solve or find solutions from problems or questions given in a good way of thinking, then they can express opinions, retell a story, present with good words and can also do good such as being open to criticism and show respect to friends' opinion. Tri Kaya Parisudha based Time Token learning model is a system for delivering learning material based on Tri Kaya Parisudha to develop speaking skills so that students can speak skillfully in various situations. It can make learning more enjoyable and can foster student confidence, increase vocabulary, interest, motivation, and can affect speaking skills. This statement is also in line with (Yofangka, 2019) which states that the use of the time token learning model can optimize the speaking skills of class V students because the average value of students' speaking skills during pre-action is 55.44 which increases in cycle I to 64.08 and 74.42. in cycle II. In addition, it is also supported by a study conducted by (Paryanata et al., 2019) which states that the mean post-test score of the experimental group that is taught by Tri Kaya Parisudha based Think Pair Share learning model is more than the average of post-test score of control group. Thus, Tri Kaya Parisudha based Think Pair Share learning model has a positive effect on student science outcomes. (Artini et al., 2016) also state that Tri Kaya Parisudha based think talk write learning model has a positive effect on the science learning outcomes of the fifth grade elementary school students. It is because the t-test results show tcount = 22.16> ttable = 2,000. Considering the previous study, the present study has differences that can be seen from the material, the measured variables and the research subject. The choice of this model had been adjusted to the characteristics of students. There were several steps of conducting Tri Kaya Parisudha based Time Token learning modified from (Kurniasih & Sani, 2016), namely the teacher conveyed learning objectives, created classroom condition, formed study groups, and distributed speaking opportunity coupons. In conveying directions and conditions for the class, the teacher instilled Tri Kaya Parisudha (manacika, wacika, kayika) to students. Teachers instilled the concepts and conducted question and answer section by speaking well and correctly (wacika). Students were given problems and students solved problems with their groups based on the Tri Kaya Parisudha (manacika, wacika, kayika). Students conveyed the results in front of the class by speaking well and clearly (wacika) after giving the opportunity coupon to the teacher. Students can ask material that is not understood by the students and the students summarize the obtained learning material by speaking well and clearly (wacika).

Method
This study was conducted in the second semester of the academic year 2019/2020 in the fourthgrade elementary school Cluster III North Kuta. This study applied quantitative research with experimental design, namely a quasi experiment. This research design had a control group but it did not play a complete role in controlling external variables that influenced in conducting the experiments. In this study, Posttest-only No-treatment Control Group Design was applied. There were two groups in this design, namely the experimental group and the control group. The experimental group was given treatment (X), namely the Tri Kaya Parisudha based Time Token learning model, while the control group was given conventional learning. After 6 treatments, post-tests were carried out in both groups to see the differences in the control and experimental groups, especially speaking skills.
The population in this study was all the fourth grade elementary school Cluster III North Kuta in academic year 2019/2020 which was divided into 16 classes in 7 elementary schools with a total population of 495 students. Based on the results of interviews from the principal of Elementary School Cluster III North Kuta, namely the fourth grade of 7 elementary school with 16 academically equivalent classes whose mean scores are not much different. It is said to be equal since the formation of classes from 7 elementary schools, students are evenly distributed among students with low, medium, and high abilities. From this information, it was said that every school in Elementary School Cluster III North Kuta did not have superior classes and non-superior classes.
A part of the population that is used for data is called a sample. (Agung, 2014) states that a sample is a part of the population that is determined to use certain techniques which are have described the entire population. The sampling technique was Cluster Random Sampling technique by using a draw that randomized the class. Therefore, all classes have the opportunity to become research samples. The draw method was conducted by recording all the names of class 4 in all elementary school Cluster 3 in North Kuata on each paper of 16, then it was rolled up. Then, it was pulled into box and shaken to select the sample. The selected group was class 4A of Elementary School Number 2 Kerobokan Kaja and Elementary School Number 2 Kerobokan. Then, two sample groups were tested for equality using the students' final test scores and t-test formula. After knowing that two classes are equal, then a draw was conducted to determine the experimental group and the control group. Based on the draw, the first paper roll was class 4 of Elementary School Number 2 Kerobokan as experimental group as well as the second paper roll was class 4A of Elementary School Number 2 Kerobokan Kaja as the control group.
After all the data were obtained, then it was proceed with analyzing the data. The data analysis was statistical analysis with inferential statistics. Inferential statistics, namely how to process data by applying inferential statistical formulas in testing hypotheses, then it was concluded based on obtained results (Agung, 2014). The inferential analysis was the t-test with the polled variance formula but the prerequisite tests were conducted firstly including the normality test of the data distribution and the homogeneity test of variance.
The normality test was conducted to determine the distribution of the group speaking skills score whether the data were normally distributed or not. In this study, the data of normality test applied the Kolmogorov-Smirnov test. The criteria for testing data for normally distributed data were if the maximum value was ≤ the Kolmogorov-Smirnov table value. Meanwhile, the variance homogeneity test was conducted to find out the difference in the true hypothesis due to the differences in variance between groups. The calculation of variance homogeneous test was conducted using the F test with the test criteria if Fcount <Ftable means that the sample was homogeneous. The test was conducted at the 5% significance level. Hypothesis testing was t-test with the polled variance formula with testing criteria at significance level of 5% and degrees of freedom = n1 + n2-2. If tcount ≤ ttable, it means that Ho was accepted, on the other hand, if tcount> ttable means that Ho was rejected.

Result and Discussion
After obtaining the post-test value data, the data were analyzed and obtained the mean, standard deviation and variance of the post-test speaking skills which are described in more detail in Table 01. The data recapitulation of speaking skill results from the experimental group and the control group.
The data were analyzed using speaking skills data that were obtained from the post-test scores with the observation sheet. The obtained data were grouped into two, namely the speaking skill data of experimental group at Elementary School Number 2 Kerobokan and the speaking skill data of control group at Elementary School Number 2 Kerobokan Kaja. The study was conducted 6 times in the

Aspect of Speaking Skill
Indicator of Skill

Pronunciation
Retelling the text using own language with clear pronunciation orally

Intonation
Retelling the text using own language with the right intonation orally

Understanding
Retelling the text using own language with an understanding based on the topic, character, and message orally 4.

Structure of Sentence
Retelling the text own language with proper sentence structure orally
Based on the results of the post-test calculations, the experimental group obtained an average of 80.95, with a variance of 47.437, a standard deviation of 6.89, a minimum score of 70.00 and a maximum score of 93.33. On the other hand, the control group has an average of 69.36, with a variance of 39.563, a standard deviation of 6.29, a minimum score of 60.00 and a maximum score of 83.34. Furthermore, the average was converted to the five-scale PAP criteria which were described in more detail in table 03. Conversion with a scale 5 PAP.  After obtaining the results of the normality test and the homogeneity of the variance test, it was concluded that the data for the two sample groups were normally distributed and had homogeneous variances, therefore the hypothesis test used the t-test. The test criteria is if tcount ≤ ttable, it means that Ho is accepted, on the other hand, if tcount> ttable means that Ho is rejected. At the 5% significance level with degrees of freedom = n1 + n2-2. The hypothesis tested in this study is that there is no significant difference in the Time Token learning model based on Tri Kaya Parisudha on the fourth grade speaking skills of SD Negeri Gugus III North Kuta for the 2019/2020 academic year. The following presents more complete details in table 05. recapitulation of the results of the t-test analysis of the study sample group.
Based on the results of normality test calculation of the experimental group, the Kolmogorov-Smirnov value was 0,126 at the 5% significant level, then it was compared to the Kolmogorov-Smirnov table with N = 35, namely 0.224. It showed that the largest <the Kolmogorov-Smirnov table value was 0,126 <0,224, then the distribution data of the experimental group was normally distributed. Based on the calculation of the normality test for the control group, the Kolmogorov-Smirnov value was 0,137 at the 5% significant level, then it was compared with the Kolmogorov-Smirnov table with N = 35, namely 0,238. This shows that the largest <the value of the Kolmogorov-Smirnov table was 0,137 <0,238 so the distribution data of the control group was normally distributed. The following was presented in more complete details in table 03. the recapitulation of normality test results for the experimental group and the control group.
After the data was declared to be normally distributed, then the variance homogeneity test was conducted. Based on the analysis results, Fcount = 1,20 was obtained, in contrary, at the 5% Ftable significance level with db (34,30) was 1.84. Then, Fcount<Ftable 1.20 <1.84 means that the data has a homogeneous variance.
After obtaining the results of normality test and variance homogeneity, it was concluded that the data of two sample groups were normally distributed and had homogeneous variances. Therefore, the hypothesis test applied the t-test. The test criteria was if tcount ≤ ttable, it means that Ho was accepted, in contrary if tcount> ttable means that Ho was rejected. At the 5% significance level with degrees of freedom = n1+n2-2. In this study, the hypothesis test was there is no significant difference in Tri Kaya Parisudha based Time Token learning model in the fourth-grade speaking skills of Elementary School Cluster III North Kuta in academic year 2019/2020. The following shows more complete details in table 05. recapitulation of t-test analysis results of sample group.
Based on the analysis using the t-test, it was found that tcount = 7.042, in contrary, at the 5% significance level with dk = 64, the value of ttable = 2.000 was obtained, then tcount = 7.042> ttable = 2.000. Thus, the null hypothesis (H0) was rejected. It means that there is a significant difference in Tri Kaya Parisudha based Time Token learning model in the fourth-grade speaking skills of Elementary School Cluster III North Kuta in academic year 2019/2020. The average speaking skill of the experimental group was 80.95 while the average speaking skill of the control group was 69.36. So, it is known that the mean value of the experimental group is more than the control group, namely 80.95> 69.36, then the application Tri Kaya Parisudha based Time Token learning model affects the speaking skills of the fourth grade elementary school.
Based on the obtained findings, the school that applies the Tri Kaya Parisudha based time token model (Elementary School number 2 Kerobokan) has better speaking skills than Elementary School Kerobokan Kaja which applies conventional learning. It is because the experimental class that applies the Tri Kaya Parisudha based time token learning model gets a coupon for the opportunity to speak and discuss problems solving or problems based on Tri Kaya Parisudha. The following steps, namely the students were given a text regarding the importance of preserving the surrounding environment which was packaged in a story entitled Si Pendekar Kali Pesanggrahan (story about warrior) and distributed several opportunity coupons to talk to students, students were invited to ask questions about the content, characters, mandate contained in the story, and exemplary attitudes from story characters by thinking (answering given questions), speaking (conveying answers well and clearly), acting (respecting different opinion with friends and imitating the good attitudes of story characters). Before answering, students were required to allow coupons to speak to the teacher. Then, the students retold the story to the class using their own language by speaking well and clearly but before that, the students are allowed speaking to the teacher. The use of a speech opportunity coupon had a rule that one coupon was used to get one speaking opportunity. If the obtained coupons had run out, students were not allowed to talk. On the other hand, students who still had coupons were required to speak until the coupon run out. It can make all students participate actively and feel confident to express their opinions because they get equal opportunities to speak. Thus, students can hone their speaking skills optimally with the use of the Tri Kaya Parisudha based time token learning model. This statement is supported by a study conducted by (Nuruzzaman, 2019) which states that the results of the study show an increase after the provision of the time token model treatment, students who achieved a KKM (minimum completeness criteria) of 22% with a class average value of 67. After giving the treatment, students who complete the score get an increase to 90% with a mean value of 82. In addition, a study conducted by (Astina & Khair, 2020) states that the speaking skills of class 5A in Elementary School Karang Anyar has increased after applying time token learning. In cycle I, the mean value of students' speaking skills is 72.1 and in cycle II increase to 78.8. During the learning process, students become the main concern and actively participate. Students can increase their confidence in the learning process since the attention is given to them. By having selfconfidence, students can improve students' learning outcomes. It is in line with (Amri, 2018) which states that self-confidence is belief in oneself, so that there is no longer a feeling of anxiety in taking an action, feel free to do desired activity, and be responsible for its actions. Thus, self-confidence increases the result of learning.
Learning with Time Token model optimizes speaking skills so students do not dominate a conversation, each student gets a chance to speak with a limit of ± 30 seconds. The student must give the coupon to the teacher before starting to speak. One coupon is valid for one chance to speak. If the obtained coupons have run out, students are not allowed to talk. On the other hand, students who still have a coupon are required to speak until the coupon runs out. Thus, students are not afraid and embarrassed in expressing opinions, providing responses, suggestions, and retelling the reading as well as can create active and enjoyable learning conditions. In a nutshell, this model can improve the characteristics of students who follow the learning, increasing the character of students will improve learning outcomes. Student characteristics are one of the factors that determine success in achieving learning objectives (Nurfaidah, Suprapta, 2018).
Learning with the Time Token model makes learning more fun, active, and teaches students to be fair by giving each student the same opportunity to convey their aspirations. In addition, the advantages of this model according to (Huda, 2017), are 1) preventing students from dominating the conversation and those who are silent; 2) increasing student activity during the learning process; 3) developing students' skills when communicating and expressing their ideas; 4) making students accustomed to give input and have an open mind criticism; 5) encouraging students to solve the problems at hand. This statement is supported by a study conducted by (Tamara, 2019;Mariammas et al., 2018) which states that the time token learning model assisted by audio-visual media has a positive effect on social science learning outcomes of the fifth grade elementary school students. Apart from that, a study conducted by (Mariammas et al., 2018) states that communication skills of students in experimental class increase after the application of the time token learning model.
In the implementation of this study, students have been trained in developing speaking skills with the time token model, then to get maximum results it is combined with Tri Kaya Parisudha so that students can speak better and clearly. Tri Kaya Parisudha is three human acts that are purified in the form of thoughts, words, and deeds (Mahardi, 2019). Sukadi in a study conducted by (Mahardi et all., 2019) reveals that Tri Kaya Parisudha consists of three elements, including manacika (good thoughts), wacika (good words), and kayika (good deeds). The concept of Tri Kaya Parisudha is very important to instill from an early age because it can foster good character for students in thinking, speaking, and doing. The characters will color the behavior which is then judged by others based on prevailing social parameters (Tirtayani et al., 2019). In instilling Tri Kaya Parisudha, the teacher directs students to see something positively, to themselves, others, and the environment. Furthermore, it directs students to say honestly and politely, to say something well and clearly. Then, it directs the students to behave well, orderly,and polite, doing something based on their obligations, not being selfish, not disturbing others, help each other, especially in solving problems in groups. Instilling Tri Kaya Parisudha is not only a direction from the teacher to students but the teacher must a role model in implementing Tri Kaya Parisudha .
Based on the research results, there is a significant effect of Tri Kaya Parisudha based time token learning model on students' speaking skills. There is an influence because all students get the same opportunity when they are speaking so that students' speaking skills can be developed better and can foster the concepts of Manacika, Wacika, and Kayika. In line with this matter, (Ani et al., 2017) argue that the time token learning model is a learning model that aims to make every student get the same opportunity to express their opinion so students can listen to the thoughts of other students. When conveying information, there is a harmonious interaction with the surrounding environment, among people, and with God which is called Tri Hita Karana. The giving concept of the Tri Hita Karana will allow students to interact during the learning process while still emphasizing the relationship with God (Prahyangan), and nature/environment (Palemahan), among others (Pawongan). In addition, (Cahaya & Sudana, 2019) reveal that the application of Tri Kaya Parisudha concept encourages students to improve student discipline, especially in controlling their minds by thinking well, controlling words by saying good, and controlling actions by doing good.
Based on the previous description, in learning activities using the Tri Kaya Parisudha based time token learning model, it affects students' speaking skills. This present study is supported by a study conducted by (Erawati et al., 2017) which states that there is a significant effect of the Time Token learning model on speaking skills in Indonesian of the fifth grade students. (Muliawan et al., 2016) also state that Audio-visual learning model assisted time tokens can improve language skills in group B4 children in PAUD Kusuma II, West Denpasar. A study conducted by (Dewi et al., 2019) also state that the group of students who are taught using the Tri Kaya Parisudha based VCT learning model is better than the group of students who applied conventional learning model. It can be concluded that there was a significant effect of Tri Kaya Parisudha based VCT learning model on social attitudes of Civics in the fifthgrade elementary school. In addition, a study conducted by (Widiasih et al., 2019) state that the average score of the experimental group mathematics learning outcomes taught by Tri Kaya Parisudha based SFAE learning model is more than the average score of the control group mathematics learning outcomes taught by using conventional learning models and concludes there is a significant effect on learning outcomes between students who are taught using the SFAE learning model and groups of students who are taught using conventional learning in the third grade students. Based on the explanation of the research results and previous descriptions, it can be concluded that there is a significant effect of Tri Kaya Parisudha based Time in academic year 2019/2020. Tri Kaya Parisudha based time token learning model is suitable to be applied in the learning process as an effort to improve speaking skills and can be a reference for teachers to optimize students' speaking skills.

Conclusions and suggestions
Based on the analysis results of speaking skills data using the t-test, it is found that tcount = 7.042, in contrary, at the 5% significance level with dk = 64, the value of ttable = 2,000 is obtained because tcount = 7.042> ttable = 2,000 so H0 is rejected. It proves that there is a significant difference in Tri Kaya Parisudha based Time Token learning model in the fourth-grade speaking skills of elementary school Cluster III North Kuta in academic year 2019/2020. The existence of a significant difference indicates that the application of the time token learning model has a positive effect on students' speaking skills compared to the conventional model. It is in line with research objectives, namely there is significant influence of Tri Kaya Parisudha based time token learning model on speaking skills of fourth grade elementary school.