FACTORS AFFECTING SCHOOL PERFORMANCE: DOES A MIXED CURRICULUM MAKE A DIFFERENCE?

This article reports on a study that attempts to fill the gap of research focusing on school performance measurements, especially those which involve curriculum types. It established the factors that measured school performance using Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA) and analyzed any differences in the performance of schools adopting national curriculum and those adopting mixed curriculum in Greater Jakarta. From 29 items, eight invalid items were dropped, and the EFA identified eight factors, which were categorized into social-emotional learning, school participation, relationship, physical-mental health, physical safety, emotional safety, academic growth, and discipline. Then, the scale was validated using data collected from 684 secondary students using CFA. Chi-square goodness-of-fit test was not statistically significant. However, other indices such as Incremental Fit Index, Comparative Fit Index, Tucker-Lewis Index, Root Mean Squared Error of Approximation, and Standardized Root Mean Square Residual were within acceptable limits, indicating that the eight EFA factors had been validated. Moreover, this study found that schools with mixed curricula had higher performance than those with a national curriculum. Nevertheless, it cannot be generalized because the linear regression shows that the p-value was higher than 0.05 (0.164).


INTRODUCTION
The curriculum quality has an essential part in the academic achievement of students (Andrietti & Su, 2019;Dewi, 2021;Krupa & Confrey, 2017). Therefore, the curriculum, a series of lesson plans and subjects of learning materials, which are implemented at school should be able to address and improve students' learning outcomes. Still, it can be used as the guideline for policymakers and society to set some agreement regarding education which can answer the future needs in ontology, epistemology, and axiology (Tedesco et al., 2014). Thus, the curriculum becomes a policy representative in the education system that is presented in a class.
Policymakers in Indonesia also continue to develop their national curriculum. The curriculum in Indonesia can be categorized based on dynamic, contextual, and relative policy products (Andrian et al., 2018). In the latest development, Indonesia uses a national curriculum known as Curriculum 2013. The curriculum is considered to emphasize modern learning by using an evidence-based approach. The thematic learning process in the Curriculum 2013, according to the Ministry of Education and Culture (2014), purposely uses a scientific approach to provide a space for students to master learning materials by not just relying on teachers alone. Students are educated to become long-life learners. Therefore, the learning process must be based on three main factors namely, skills, attitude, and knowledge.
In addition to the national curriculum, the government also allows private schools to use international curricula such as Cambridge International, International Baccalaureate, and others. Initially, the government allows these schools to use the name International school until finally the government issued the Regulation of the Minister of Education and Culture No. 31 of 2014. Since the publication of this regulation, schools in Indonesia no longer use international labels but changed their status to SPK -Satuan Pendidikan Kerjasama (Joint Education Unit).
Parents in urban areas are interested in sending their children to SPK schools (Rinaldi & Saroh, 2017), mostly because their children desire to master a foreign language early on. The increasing use of English as a communication medium in Indonesia is in line with the increasing number of SPK schools. In non-anglophone countries, there tends to be a rapid change from English as a foreign language (EFL) to English as the medium of instruction (EMI) for academic subjects such as science, mathematics, geography, and medicine (Dearden, 2016). Many families in non-English-speaking countries consider learning English from an early age to be a critical factor in achieving global higher education success, international job opportunities, and prosperity (Muslim et al., 2020;Santos, 2019). Policymakers, educators, and business organizations in some countries regard English-medium teaching as a "golden ticket to a global world" (Galloway et al., 2017;Shimauchi, 2018;Zhang, 2018).
In the research on why parents particularly preferred an international school, around 83 percent of school admission staffs agreed that the curriculum seemed to be "very significant" to most parents, while 17 percent believed that it is 'somewhat relevant,' and none of them were sure whether parents were concerned about the curriculum (ISC Research, 2021). Today's SPK schools are more likely to adopt an international curriculum or provide an adapted version of a national curriculum that pay attention to what country a school belongs. Mixed curriculum models are also developing, incorporating various national as well as international curricula strands, including the host country's national curriculum elements (Buchanan et al., 2018;Hameed, 2020;Sihotang & Datrix, 2018). Today, many international schools use a flexible curriculum tailored to meet the needs of the school's venue, student population, and strengths. However, does the mixed curriculum make schools perform better than schools that use the national curriculum? This question then became a reference for this study.
The average standardized test results have many limitations to measure school performance but are still often used (López, 2019;Schneider et al., 2017). Generally, in the search for the schools with good performance in Jakarta, we will be given a list of schools with the highest average in the entrance test of public universities or other standardized test results. Meanwhile, parents and people of the community tend to focus on other indications of school excellence. They mostly rely on reputation, word-of-mouth, and what they see with their own eyes, such as facility facilities or student demographics (Mayer et al., 2020;Schneider et al., 2017).
Shreds of evidence show that a wide variety of academic and life outcomes are predicted by non-cognitive skills or student skills other than academic achievement (Duckworth & Yeager, 2015;West, 2016). These non-cognitive attributes are varied, but they all work together to support the goal-directed effort, for example, social-emotional skill (Brackett, 2016;Durlak et al., 2011;Goldberg et al., 2019) and emotional intelligence (Astatke, 2019;Molla, 2018;Rai & Khanal, 2017). Besides, longitudinal studies have shown that these characteristics are significant predictors of academic, economic, social, psychological, and physical wellbeing (Duckworth & Yeager, 2015;Jackson et al., 2015).
In public education, the value of reliably evaluating academic success is critical. To ensure all students get high-quality education, schools must provide acceptable academic proficiency and advancement metrics. However, considering school performance, concentrating solely on the standards-based academic testing may lead to misrepresentation of student/school performance, streamlining of course work around restricted subjects, and an overemphasis on student outcomes over the multitude of personal, contextual, and environmental factors contributing to student's academic knowledge, wellbeing, and individual development (Babineau, 2017;Saputra et al., 2020).
This study examined the impact of school form on academic proficiency, school climate, and social/emotional learning (SEL) rather than academic achievement. This paper was based on the data obtained from Greater Jakarta secondary students in grade 7-12. The performance of mixed curriculum schools (SPK schools) and national schools were compared based on students' perceptions. The aim of this study was to establish the factors that measured school performance using Exploratory Factor Analysis (EFA) and Confirmatory Factor Analysis (CFA). In addition to establishing the factors, this study, by using linear regression, was also aimed at observing whether any school-related performance could give some benefits to students who considered national and mixed curriculum schools.

METHODS
This study used a quantitative method by using questionnaires to measure students' perception of schools' performance in Greater Jakarta. This study employed school performance as a latent variable consisting of some indicators as the sub-constructs, such as academic proficiency, school climate, facilities, safety, and social/ emotional learning. The sampling techniques applied were purposive sampling and snowball sampling using an individual unit analysis. Six hundred eighty-four secondary students from 132 public and private schools in Greater Jakarta officially participated as research respondents. The description of the participants can be seen in Table 1.
The questionnaire was given to secondary students, and the results were gathered. Then, a confirmatory factor analysis with Lisrel was used to look at the contribution of each aspect as mentioned above to overall school performance. In addition, SPSS was used to see a different performance between national curriculum schools and mixed curriculum schools with simple linear regression. The following steps were used to collect the data. First, students were given a set of questions asking about their school performance. The questionnaire was uploaded to Google Form, so the results could be easily drawn right after the link was spread out. Next, the students answered the questionnaire independently at their own places, without any distraction from other parties, such as school staffs or researchers. The designed measurement was divided into two different parts: (1) a set of questions about the profile of participants (gender, degree of study, school's location, and school's curriculum); (2) a set of questions about the school performance. The answer type was ranged from 1 to 5 on a 5-point Likert scale, with 1 indicating strong disagreement and 5 indicating strong agreement. For the factor analysis, Mundfrom et al. (2005) suggested a minimum sample size of t to 20 observations per scale item or 100 to over 1000 participants.
The designed instrument was initially given to 30 students as the part of the pilot study. The pre-test allowed researchers to see if various response groups accurately phrased, arranged, and understood the questions. Before the full-scale study, some errors found during the implementation were fixed. The content validity of the research instrument was confirmed by confirming the study's goal aligned with the items in the research instruments, and the items were assessed by three experts in education prior to the implementation of data collection. Cronbach's alpha of 29 items' reliability ranging from 0.772 to 0.792, indicating that the instrument was reliable, as shown in Table 2. For Exploratory Factor Analysis (EFA), SPSS Version 25 was used, while Lisrel was employed for Confirmatory Factor Analysis (CFA). The first stage was to determine the standard normal distribution by ensuring that the kurtosis (<7) and skewness (<2) were within acceptable limits (George & Mallery, 2010). Next, the mean, standard deviation, skewness, and kurtosis were used to assess twenty-nine variables (Table 2)the mean with a range of 3.25 to 4.01 and a standard deviation of 0.825 to 1.373. The skewness was between -0.879 and -0.084, while the kurtosis was between -1.207 and 0.037.  The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy (> 0.5) and Bartlett's Test of Sphericity (< 0.05) were used to filter data in the initial factor analysis (EFA) phase (Williams et al., 2010). A parallel analysis using Principal Component Analysis was done to determine the number of factors kept in the model (Iantovics et al., 2019;Ledesma & Valero-Mora, 2007). There were 29 items calculated with a sample size of 684, and factors with Eigenvalues > 1 were evaluated (Hayton et al., 2004). According to Hayton et al. (2004), eight factors that gave the researchers the confidence to proceed to the EFA were identified, categorized as social-emotional learning, school participation, relationship, physical mental health, physical safety, emotional safety, academic growth, and discipline.
The CFA was done to validate the identified factors, and a variety of indices were used to determine the model's fit. However, Jöreskog et al. (2016) said choosing fit indices indicating the best fit is critical, and providing everything in the program output would be too much for the reader, so, in this study, the model was evaluated using the Chi-square goodness of fit statistics (p> 0.05), the Comparative Fit Index (good fit: CFI> 0.97; mediocre fit: CFI>0.90), the Incremental Fit Index (IFI>0.90), the Turker-Lewis Index (good fit: TL>0.95; mediocre fit: 0.80<TLI< 0.95), the Root Mean Square Error of Approximation (RMSEA<0.08), the Goodness of Fit Index (GFI>0.90), the Adjusted Goodness of Fit Index (AGFI>0.90) and the Standardized Root Mean Square Residual (good fit: SRMR<0.05; mediocre fit SRMR<0.08) (Hooper et al., 2008).
Thus, linear regression model with dummy variables was estimated for separate school groups (national and mixed curriculums), followed by a significance test for the set of dummy variables. By fitting a linear equation to the observed data, linear regression describes the connection between one dependent and more independent variables (Gordon, 2015). The linear regression could answer whether the type of curriculum affect the school performance or not.

FINDINGS AND DISCUSSION Findings
Exploratory Factor Analysis EFA was used to validate the instrument after experts' judgment. This research was based on the responses of 684 secondary school students who completed the 29-item questionnaire.
The Kaiser-Meyer-Olkin (KMO) value in this study is 0.825, which indicates that the sample size is sufficient for factor analysis, and the Bartlett Chi-Square approximation is 3089.58 with p = 0.000, as in Table 3. KMO value close to 1 means the correlation pattern is compact enough to yield different and dependable factors. Therefore, the EFA approach is found to be appropriate for use in this study, based on the findings of the Kaiser-Meyer-Olkin and Bartlett sphericity tests (Auerswald & Moshagen, 2019). There were eight factors with Eigenvalues > 1 that were obtained using Parallel Analysis (Horn, 1965), one of the preferred ways to determine the number of factors. In order to presume that the measurement has good validity, the commonalities must be more than 0.30 while doing the extraction. In all items, the analyzed instrument meets this condition, as presented in Table 4.
Furthermore, the analysis of the main components revealed that the eight factors account for 49.918 percent of the overall variance. The following are the names of the eight factors: Factor 1 is Social-Emotional Learning (SEL1-SEL6) which has six items, each of which has a factor loading ranging from 0.438 to 0.553. Factor 2 is the School Participation (SPA1-SPA5) which has five items with factor loadings from 0.406-0.619. Factor 3 is the Relationship (REL1-REL3) which has three items with factor loadings from 0.488-0.687. Factor 4 is the Physical and Mental Health (PMH1-PMH3) which has three items with factor loadings from 0.459-0.633. Factor 5 is the Physical Safety (PSA1-PSA3) which has three items with factor loadings from 0.439-0.705. Factor 6 is the Academic Growth (ACG1-ACG3) which has three items with factor loadings from 0.401-0.626. Factor 7 is the Emotional Safety (ESA1-ESA3) which has three items with factor loadings 0.481-0.513. Factor 8 is Discipline (DIS1-DIS3) which has three items with factor loadings 0.465-0.627. Only those items with a factor loading greater than 0.40 are shown in Table 5. Confirmatory Factor Analysis They identified eight components of the school performance Confirmatory Factor Analysis (CFA) subjected cross-validation, and multicollinearity was discovered among the factors as independent variables. Figure 1 shows the CFA diagram of school performance factors. It is represented in route diagrams with latent variables represented by circles and observable variables represented by squares. The single-headed arrows reflect the covariance between the six latent variables, whereas the two-headed arrows show the expected direction of impact (Costa et al., 2016).
Eight items were dropped because they were not valid since the estimate loading factor was less than 0.5. Their codes were SEL1, SEL2, SPA1, SPA2, SPA3, PSA2, ACG1 and DIS2. The Chi-Square of the school performance instrument was value 528.69. It is not achieved the threshold with df=349 and p-value = 0.00. Nevertheless, other indices were passed at least the threshold of mediocre fit. RMSEA was 0.058 (achieved the threshold of 0.08 -good fit), GFI was 0.93 (achieved the threshold of 0.90mediocre fit), AGFI was 0.93 (is achieved the threshold of 0.90mediocre fit), CFI was 0.91 (is not achieved the threshold of 0.90mediocre fit), IFI was 0.92 (achieved the threshold of 0.90mediocre fit) and TLI was 0.89 (achieved the threshold of 0.90mediocre fit). These results were within acceptable limits, indicating that the EFA's eight factors had been validated.

Linear Regression with Dummy Variables
A linear regression analysis with the curriculum as a binary variable (0=National Curriculum and 1=Mixed-Curriculum) is required to see the difference in performance between schools with a national curriculum and a mixed curriculum. The calculations with SPSS found that schools with mixed curriculum did have higher performance on average than schools with a national curriculum. However, the type of curriculum usage has minimal effect on school performance. It is showed by the R-square value that reached at 0.003, as shown in Table 6. Regression equation was formed with variable de-pendent school performance (Y) and X as an independent variable binary curriculum with the national curriculum as the basis of calculation is Y = a + bX, while a and b are the coefficients listed in Table 7, so the equation becomes Y = 3.702 + (0.041 * curriculum). For national cur-riculum Y = 3.703 + (0.041 * 0) while for mixed curriculum Y = 3.703 + (0.041 * 1). However, the result is not clearly significant because p > 0.05 (0.164). Thus, it could not be generalized that using a mixed curriculum can affect school performance

Factors of School Performance
This study confirmed that SEL is one of the school performances factors. The finding is the meta-analysis of the 213 school-based SEL programs by Durlak et al. (2011). That analysis found that SEL program participants' social and emotional skills, attitudes, behavior, and academic achievement all increased considerably. Social and Emotional Learning (SEL) is a conceptual framework for supporting students' social, emotional, and academic competence (Edwards et al., 2019). It entails fostering social and emotional competencies through explicit instruction and student-centered learning approaches that encourage students to participate in the learning process and develop analytical, communication, as well as collaborative skills (Weissberg et al., 2015).
This finding is also in line with Dowling et al. (2019) research. They recorded significant gains in SEL intervention students' social and emotional abilities, including the reduced emotion suppression, improvements in mental health and wellbeing, and significant lower stress levels. Therefore, school-based social and emotional learning intervention aims to equip children with skills they need to deal with life's obstacles, improve their social and emotional wellbeing, accelerate their academic performance, and lower their risk of mental health problems.
The results of this study also indicate that students' participation played an essential role in their school performance. The finding has been strengthened by the result of Banatao (2011), Dogan (2015, and Pan and Zaff (2019) studies: students' level of participation in school activities can predict their academic achievement. In addition, involvement in extracurricular is connected to improvement in academic performance and school attachment (Badura et al., 2016). Although, participation of students in the educational process is not yet a standard educational practice (Cervantes & Galván, 2019).
Schools need to transform their environment to boost student participation in class activities or events. Cervantes and Galván (2019) said that improving school environment cannot be accomplished by creating rules and regulations, or discussing topics connected to the topic, but rather through strategies for changing scholar practices and management policies. Students offered numerous opportunities to participate in their school's activities perceive their school's environment as harmonious and pleas-ant. The students can also feel good about themselves at school and are driven (Mithans & Grmek, 2020).
The study results indicate that good relationships among students and adults in school also signified high performance. This finding is in line with the result of Scales et al. (2020) research. They found that students' academic motivation was substantially predicted by teachersstudent relationships at the start and end of the school year and students' sense of belonging and school climate. Moreover, the school climate may influence the success of the teaching and learning process in schools. The school climate is generated by interactions between the principal and teachers, teachers and peers, teachers and staff, teachers and students, or relationships among students (Syahril & Hadiyanto, 2018). Graham et al. (2016) state that relationships play a vital role in improving wellbeing in schools. Thus, the starting point for school leaders to increase student and teacher relationships is to establish a positive, open, and collaborative academic ambiance (Dorina, 2013). Moreover, the relationship between students and school staffs (non-teaching staff) has also affected students' health and wellbeing (Littlecott et al., 2018), so school management should carefully hire school staff in order to form mutually trusted relationships between faculty and students.
The findings of this study also show that physical and mental health is a part of school performance dimensions, which are in line with Pojednic et al. (2016), that school activities focused on physical and mental health could positively affect student's cognitive capacity and academic performance. Likewise, it has a positive correlation with other factors. Furthermore, poor school environmental conditions have been linked to the decline in both physical and mental health and promoting cognitive function failures (Hameen et al., 2020;Temprano et al., 2020).
School counselors can improve students' social and academic performance, but in some cases, job ambiguity and conflict limit them to acting as school administrators rather than master's-level professional educators with a mental health background (Blake, 2020). Nevertheless, good mental health services in schools could treat students and provide chances for collaboration and innovation among teachers, counselors, and psychologists (Hubbard et al., 2018). In addition, the counselors can help students to receive early detection of anxiety symptoms and access to therapy.
This study also explicitly implied that physical and emotional safety was attributed as the dimensions of school performance. This finding reinforces Kim et al. (2020) that students' school safety problems and anxiety are substantially connected with fighting, threats, and bullying, resulting in worse academic performance for both girls and boys. Nevertheless, students' safety tends to be neglected in school performance evaluation (Casey et al., 2018). For example, 42 percent of all students in a global survey of 5,805 children aged 10-12 years in both emerging and developing countries claimed they did not feel safe at school (Shean & Mander, 2020).
School-wide bullying was found to be highly associated with emotional and physical safety in both male and female students, with physical bullying being more strongly related (Fredrick et al., 2021). Additionally, cyber bullying is a new type of bullying that has been rising over the last decade within the society, particularly in schools. It has negative social, physical, and emotional consequences for victims, offenders, and bystanders among K-12 students. Therefore, school psychologists and counselors must advocate for prevention, intervention, and more effective policies (Elbedour et al., 2020).
According to Shean and Mander (2020), emotionally unsafe surroundings are linked to stress, decreased school attendance, and worse learning engagement whereas emotionally safe conditions are linked to more positive identity development, better learning experiences, and more worthy emotions. Furthermore, feeling safe at school, connected to school, and peer support are protective variables for both mental and emotional wellbeing throughout the transition phase whereas connectivity to teachers is protective of emotional wellbeing (Lester & Cross, 2015). Hence, it is a solid suggestion to improve emotional safety in schools to have a favorable impact on learners' academic, behavioral, emotional, physical, and mental wellbeing outcomes.
This study also shows that academic growth was one of the school performance dimensions. This finding has also been supported by a study conducted by Giersch et al. (2021) from which it can be seen that the academic growth has a more significant relationship with outcomes than either of the school performance measures. Students' scores can be compared to their previous performance rather than a benchmark, allowing for improved school perfor-mance scores. Policymakers can consider the progress achieved by all students and instructors, not just those who achieve competency in a particular subject, including this component in school performance measurement (Babineau, 2017).
Academic growth is a more objective than academic proficiency. Proficiency assessments track students' progress over time and are linked to their personal histories, favoring students from higher socioeconomic backgrounds. When public schools highly appreciate students with a high level of proficiency, or the number of students can reach the threshold of standardized exams, the other unfavorable schools can only receive students with lower level of proficiency (Giersch et al., 2021). Perhaps the most irritating policies that stress proficiency appear to harm students in the most underprivileged schools the most (Jennings & Sohn, 2014;Lauen & Gaddis, 2016). The impact of a particular school on its pupils can be better captured by judging schools based on their growth on standardized examinations (Guarino et al., 2015). As a result, researchers frequently suggest policies that prioritise measurements that indicate growth in student mastery over a school year when evaluating students (Giersch et al., 2021).
The study results indicate that controlling self-discipline is also a factor in school performance. This finding is also previously written in a research conducted by Baumann and Krskova (2016) which mentions that students with good discipline can work efficiently and leads to improved academic performance. Ehiane (2014) research demonstrated that effective school discipline should be promoted in controlling students' behavior, impacting the environment. According to Lumadi (2019) findings, a lack of disciplinary management abilities can lead to disruptive behavior, non-compliance with school rules, and low student performance.
The Effect of Curriculum Type on School Performance The study implies that the curriculum types (national curriculum or mixed curriculum) do not affect school performance. The mixed curriculum schools, indeed, have higher performance on average, but it cannot be generalized in the population. As expected, this result does not have any solid reinforcements since lack of comparative studies between those two school types. The closest one is found by Kortelainen and Manninen (2019) that private schools out-perform public schools by a slight margin, but the difference is so minor that it is statistically insignificant. Public schools in Indonesia officially belong to national curriculum schools, but private ones have national curriculum schools and mixed curriculum schools (SPK). Thus, those findings have a low correlation.
A mixed curriculum school is a unique entity because of the restrictive Indonesian education policies (Sakhiyya, 2011). For this reason, it is very unlikely to find any previous study about mixed curriculum school. Education researchers tend to examine 'International School' as most other countries can apply only the International curriculum without mixing it with the local curriculum. Researches with 'bilingual school' as object analysis could be the most correlated study to reinforce mixed curriculum findings.
The eight factors from the established model of school performance cannot assist the comparison as well. For example, a lack of shreds of evidence that shows Social-Emotional Learning had been applied entirely in the Jakarta schools (Rahmawati, 2019). Moreover, according to a study of 452 eighth grade students under 15 in Jakarta junior high schools, the students receive insufficient social and mental health support from their schools (Triana et al., 2019). In line with a study involving 723 Senior High School students from five Jakarta administrative districts as samples, many programs have been adopted to combat bullying; however, they are ineffective at the moment (Rahmawati, 2020).

CONCLUSION
The results confirmed that socialemotional learning, school participation, relationship, physical-mental health, physicalemotional safety, academic growth, and discipline are the essential factors to improve in order to enhance school performance. A model with those validated factors could help policymakers or parents to examine the school's quality rather than using the average of standardized exams. When discussing the academic quality of an educational institution, the genuine level of a good education cannot be expressed in a single score or measured by a single factor. This study could possibly avail a chance for both policymakers and students' parents to redefine the schools accountability. Educators might embrace the more extensive and more nuanced definition of the students and school performance by strengthening the academic measurements and moving toward considering the nonacademic factors that help the students to succeed.
The results also confirmed that schools with mixed curricula had higher performance than schools with curriculum national, but the type of curriculums did not affect school performance. English Medium Instruction (EMI) could be the most significant contributing factor in favoring SPK (mixed curriculum) schools more than national ones. Besides, the indirect effect of students' higher socioeconomic background could significantly put SPK schools in advantage. However, if the measurement only focused on the factors controlled by the school administrators, the results stated that the difference was not significant.