Predict Observe Explain Learning Model : Implementation and Its Influence on Students’ Critical Thinking Ability and Learning Outcomes (A Meta -Analysis Study)

: This study aims to examine the application of the POE (Prediction, Observation, and Explanation) learning model in Mathematics and Natural Sciences learning in the last 10 years in terms of advantages and constraints in Mathematics and Natural Sciences learning at elementary school, junior high school, and senior high school levels. The research used meta-analysis with a quantitative approach and data collection of Scopus/WoS and SINTA journals literature studies on POE in MIPA learning combined with MIPA materials for elementary school, junior high school, and senior high school. The data collection techniques of this research were (1) searching journal and repository databases, (2) data encoding & tabulation, (3) JASP software inputs, (4) interpretation & conclusion: JASP output, moderator variables. The research results showed that applying the POE learning model had a greater effect on student Learning Outcome than on student Critical Thinking with an effect of 60%. Then based on the education level, the POE model's influence on Critical Thinking and Student Learning Outcome was greater at the senior high school level, and based on the material field, the POE model had a greater influence on the mathematics material field. Then the moderator variable showed a very drastic decrease in the effect of the POE model on the Critical Thinking and Learning Outcome of students in 2018-2022, with a RE value of 47% smaller than the previous year. In addition, we recommend that the POE model is better applied at the high school level than at the elementary and junior high school levels.


Introduction
The learning model is a pattern that has been carefully planned and guides the implementation of learning, starting from the opening, core, and closing activities, as well as learning assessments that are arranged in such a way as to achieve learning objectives (Salsabila et al., 2022). The learning model describes the broadest level of learning practice and contains a learning philosophy orientation, which is used to select and arrange teaching strategies, methods, skills, and learning activities to emphasize one part of learning (Nasrun et al., 2018). So the learning model is very important to be applied during the learning process to achieve effectiveness and success in the learning process.
When the learning process is carried out, it is still found that many teachers or educators have not properly placed the learning model based on the conditions that occur. As a result, it causes problems that affect students' critical thinking skills and the achievement of learning objectives; this is because the learning activities carried out are still teachercentered, so students tend to be passive (Ratnawati et al., 2020), (Dari & Ahmad, 2020).
Critical thinking that students want is a way of thinking where the student can draw rational conclusions about something that is considered the truth (Lusiana & Tuti Zubaidah, 2020), as well as critical thinking skills that use a decision-making technique to solve problems for reasoned and reflective thinking In the sense that students not only look for answers, but more than that, such as asking questions about answers, facts, or existing information (Delita et al., 2022). To achieve this critical thinking ability, in general, there are 5 indicators of critical thinking skills that the teacher must consider, namely: (1) providing simple explanations (elementary clarification); (2) building basic skills (basic support); (3) making inferences (inferring); (4) making further explanations (advanced clarification); (5) organizing strategies and tactics (Islamiyah et al., 2019).
In addition, students are also expected to improve their learning outcomes, which requires a learning method that can improve cognitive abilities and affective and psychomotor abilities (Algiranto, 2018). However, the problem is the low learning outcomes of students; this is because students need to be fully involved in learning and be trained to explore and process information, make the right decisions, and solve problems. Students are only recipients of information, so they make students' thinking skills low. In other words, learning is felt to be less meaningful (Yupani et al., 2013).
Based on the above problems related to the lack of critical thinking skills and student learning outcomes, a learning model that combines students' cognitive, affective, and psychomotor abilities is needed, namely the POE (Prediction, Observation, and Explanation) learning model. The POE learning model was first developed by White and Gunstone (Rini et al., 2018), (Larasati et al., 2018) which aims to reveal students' ability to make predictions (Nur Hayati et al., n.d.), (Ikhfa et al., 2020). The POE model comes from the words Predict, Observe, and Explain (Khilyatin Ula et al., 2022), (Putra Erawan et al., 2022). POE is a learning model that involves students directly in the learning process by conducting experiments to make the knowledge gained more meaningful (Ulya Ulpa et al., 2019), (Nurfadilah et al., 2022). The POE learning model is a learning model that adheres to the theory of constructivism (Amaliyah & Nasrudin, 2019), (Abdillah & Surya Anggara, 2021) and involves students in thinking critically to understand the material to be learned (Ika Inayah et al., n.d.) (Lantik et al., n.d.). In the POE learning model, the first step is the ability to predict, known as the ability to formulate hypotheses (temporary answers). After that, the teacher writes down what the students predict. The teacher asks students, "Why is that?" to answer these questions, the teacher invites students to make observations, namely, making a series of observations through experiments. The teacher guides the students to conduct the experiment and use the data generated to conclude. The conclusions obtained are then matched with the predictions given by the students. If correct, students will be more confident in the concepts they master. However, if the student's prediction is incorrect, the teacher will help the student find an explanation. Thus students can correct conceptual errors themselves (Shofiah et al., 2017).
Applying the POE learning model to aspects of learning at the elementary, junior high, and high school levels is appropriate for improving students' critical thinking skills and learning outcomes. Such research (Linda Ayu Kusumaningsih et al., 2020) showed an increase in the critical thinking ability of grade V elementary school students through the predict observe explain learning model assisted by scrapbook media. It can be seen from comparing the classical completeness of the evaluation test results obtained in the pre-cycle 37%, cycle I 63%, and cycle II 85%. Then the research (Lusiana & Tuti Zubaidah, 2020) stated that by applying the POE learning strategy to statistics material in class VIII SMP, the results showed that students' critical thinking skills through the POE learning strategy reached a good category with a presentation of very good 12%, good 48%, sufficient 36%, and a category of less good 4%. Furthermore, research (Okta Nurfiyani et al., 2019) compared the success of POE learning in class X SMA through experimental and control classes, with an average score of learning outcomes and students' critical thinking skills in the experimental class of 26.41 and 46.64 while the control class was 19.92 and 35.33. Then (Elistiana Safitri*, Kosim, 2019) research shows that the POE learning model applied to the experimental class has a better effect than the conventional learning model applied to the control class. Furthermore, Hasrul's research shows that the application of the Predict-Observe-Explain (POE) learning model can improve the learning outcomes of VIII grade students at SMPN 1 Liukang Tangaya Pangkep based on data exposure and discussion of research results. The learning test results in evidence increased by 100%, with an average learning test score of 82.66, including in the good category. Some of the research results above show that the POE learning model affects learning outcomes and students' critical thinking skills.
Previous research was limited to discussing the effect of applying the POE model in improving critical thinking skills and student learning outcomes in Mathematics and Natural Sciences material limited to one level of education. Therefore, researchers conducted this study to see the comparative level of the influence of the POE learning model in improving critical thinking skills and student learning outcomes in Mathematics and Natural Sciences material fields at the elementary, junior high, and high school levels.

Research Method
This study is a research that summarizes and analyzes data from previous studies, known as a meta-analysis. Researchers collected data by searching for online articles in Scopus/WoS and SINTA journals. They focused on the "POE learning model" as the dependent variable and used "Critical thinking" and "learning outcome" as independent variables. The steps followed in this study were based on the opinion of (Hardina Aldira, siti et al., 2021) and (Wahyuni et al., 2022). First, researchers collected data from the "Google Scholar" indexer database based on inclusion and exclusion criteria. The inclusion criteria included (1) search keywords, namely "POE learning model" and "Critical thinking", and "Learning outcome"; (2) articles published in 2013-2022; and (3) articles using Indonesian or English. While the exclusion criteria include (1), there is the amount of data (N); (2) education level; (3) material field; (4) there is a correlation coefficient value (r), fisher value (F), t-test value (t) and classical value (k).
The second stage was analyzing data in the form of (1) labeling or numbering articles related to the topic, (2) writing the name of the researcher, the level of education, and the field of education, (3) writing the correlation test value (r), writing the proportional test value (class), and the number of research subjects (N), (4) calculating the effect size (ES) and standard error (SE), (5) conducting data analysis using JASP software (6) analyzing the results found from the articles that become data references (7) writing moderator variables, (8) drawing conclusions and suggestions.
The third stage was calculating the effect size (ES) and standard error (SE) values using the formula from (Mansurah et al., 2021) as follows: a) Correlation research results (r), fisher (F), and t-test In the fifth stage, the criteria for concluding according to the category of influence level was determined by the Effect Size (ES) and Standard Error (

Results and Discussion
The search results from the indexer database obtained as many as 110 publication data based on the last 10 years of research (2013)(2014)(2015)(2016)(2017)(2018)(2019)(2020)(2021)(2022). From this data, complete and incomplete data were obtained. Complete data consists of 70 data, and incomplete data consists of 40 data, with 28 data not having the number of participants (N) and 12 data not having test values (t, F, r, or k). From the 70 complete data, the results of the ES and SE values calculation are as follows.  Table 2 shows that applying the POE learning model in the teaching and learning process considerably influences students' critical thinking skills and learning outcomes, with the acquisition of a RE value of 60% in the moderate category. Furthermore, the researcher will calculate the effect by using the hypothesis and publication bias tests on the data obtained as follows.  Table 3, it can be seen that there is no publication bias as indicated by the p-value <001.The value of the funnel plot model according to the following Figure 2: Figure 2 shows that the shape of the funnel plot is symmetrical, which means that there is no publication bias and the points are above the curve, which means that the ES estimate is accurate, then the overall picture above shows that the points are around the curve, it can be concluded that there is no missing research data.  Table 4 shows the effect of the application of the POE learning model has a greater effect on student Learning Outcomes than the Critical Thinking ability of students. This can be seen from the RE value of the effect of the POE learning model on Learning Outcome of 66% versus 1% greater than its effect on student Critical Thinking of 51%. Table 2 shows the effect of applying the POE model on students' critical thinking based on their education level is obtained as follows:   Figure 3 shows the shape of the funnel plot is symmetrical, which means that there is no publication bias and the points are above the curve, which means that the ES estimate is accurate. Then, the overall picture above shows that the points are around the curve, it can be concluded that there is no missing research data. From Figure 4 shows the shape of the funnel plot is symmetrical, which means that there is no publication bias and the points are above the curve, which means that the ES estimate is accurate, then the overall picture above shows that the points are around the curve, it can be concluded that there is no missing research data. Figure 5 shows that the shape of the funnel plot is symmetrical, which means there is no publication bias, and the points are above the curve, which means that the ES estimate is accurate. The overall picture above shows that the points are around the curve; it can be concluded that all research data exists. Table 2 shows the effect of applying the POE model on students' critical thinking and learning outcome based on the material field; it is obtained as follows:  Table 6 shows the effect of POE learning model on Critical Thinking and Learning Outcome of students is greater in the field of mathematics material with RE value of 67%.

Based on Moderator Variable
Moderator variable analysis is needed to see the classification, year of publication, and amount of data. The results of data analysis using JASP are shown in Table 7.  Table 7 shows that a very drastic decrease in the effect of the POE model on Critical Thinking and Learning Outcome of students, namely in 2018-2022 with a RE value of 47% smaller than the previous year.

Discussion
Kosim, 2019), which states that by applying the POE model, the teacher can make students express their initial ideas, motivate students to want to explore concepts and arouse the desire to investigate. POE is also a learning model where the teacher plays a role in exploring students' understanding by asking them to carry out three main tasks, namely prediction (predict), observation (observe), and explanation (Explain) (Shafariani Fathonah, 2016).
The application of POE learning provides findings of an increase in students' thinking skills (Ai Linda Nurmalasari, Asep Kurnia Jayadinata, 2016) and can explore student knowledge (Ula et al., 2022), POE is also an efficient learning model to create student discussions about science concepts. This learning model involves students predicting a phenomenon, making observations through demonstrations, and explaining the results of their previous demonstrations and predictions.
Another finding of this study is that the use of this POE Method, according to (Ponna et al., 2022) has several weaknesses, namely the need for maximum preparation, especially those related to the problems presented and activities carried out in proving the predictions submitted by students, the need for special abilities and skills for teachers so that teachers are required to work professionally, the need for good teacher willingness and motivation for the success of the student learning process.

Conclusion
The results of the above research show that applying the POE learning model has a greater effect on student Learning Outcome than on student Critical Thinking with an effect of 60%. Then based on the education level, the POE model's influence on Critical Thinking and Student Learning Outcome is greater at the senior high school level, and based on the material field, the POE model has a greater influence on the mathematics material field. Then, the moderator variable shows a very drastic decrease in the effect of the POE model on the Critical Thinking and Learning Outcome of students in 2018-2022, with a RE value of 47% smaller than the previous year. It could be due to the emergence of covid-19 in 2020, where teaching and learning activities are online, so the POE model needs to be applied optimally.

Recommendation
We advise teachers to apply the POE learning model in the learning process. Moreover, this learning model goes through three stages: predict, observe, and explain, which can help students improve critical thinking skills and student learning outcomes. Then we suggest that the principal provide a policy to teachers to implement a strategy/model/learning tool to support students' thinking skills and success.  Khilyatin ula, f., khoiri, n., akmalia, h. A., biologi, j. P., sains, f., & teknologi, d. (2022).