Contributing factors to perceived educational quality in Mongolia: Developing instruments using principal component analysis

Abstract Educational quality is perceived differently as it is embedded in its social, cultural, political, and economic circumstances. The current study identified the factors contributing to perceived educational quality in Mongolia and attempted to develop an instrument to measure educational quality using principal component analysis. Using qualitative and quantitative approaches, the analysis yielded 71 items contributing to educational quality in six themes: school environment, school administration, students, parents, curriculum, and teachers. Each theme yielded one to four components as measurable instruments. The study suggests that school conditions should be prioritized to improve educational quality in Mongolia.


Introduction
The quality of education is one of the most debated issues in Mongolia.The term educational quality is a vague concept that begs a clear definition.Stakeholders, such as educators, government officials, parents, and students, view educational quality from different perspectives.Many social issues are related to the perceived quality of education.When we observe discussions on education on social media or elsewhere, stakeholders relate problems such as an oversupply of the labor force or lack of skilled labor force, standardized tests, different kinds of curriculums, parental involvement, student-centered instruction, teacher development, private schooling, and adoption of educational approaches from well-developed countries with "good quality of education."However, what we mean by educational quality and what properties it should contain is unclear.Is it a good school?Is it a good teacher?Is it a good student?In an attempt to answer this broad question, we were interested in the properties of our "perceived educational quality," as our understanding of educational quality is multifaceted.Based on previous literature, we started by questioning the possible contributing factors that can create a proxy for a differently perceived quality of education in Mongolia.
In light of exploring contributing factors to educational quality in Mongolia, we began surveying earlier research to understand educational quality elsewhere.The issue of defining educational quality is not a recent one, as scholars have attempted to define it.Previous research has stated that the quality of education depends on environmental factors, such as the culture and values of the country, sociopolitical circumstances, and economic situation (Mortimore & Stone, 1991;Hanushek & Raymond, 2002;Hickey & Hossain, 2019;Scheerens et al., 2011a).Our environments, such as a country's economic situation, cultural values of education, socio-political dimensions, and social demands, affect our understanding of education and its quality.Perceptions of educational quality can differ among cultures, economies, countries, and stakeholders and evolve depending on environmental factors and paradigms.Therefore, it can be challenging to precisely define educational quality and understand it in a broader sense.
Despite the challenges in defining the term "educational quality," scholars have attempted to provide definitions from different perspectives.For instance, Tikly & Barrett (2011) consider educational quality in terms of social justice and capabilities perspective; Cheng (1999) approaches educational quality in terms of school effectiveness, whereas Barrett et al. (2006) define educational quality from an integrated approach of economist and humanist perspectives.However, the literature review showed that there is little attempt to understand and measure educational quality in the context of Mongolia.
This study was conducted for the following two reasons.First, few empirical studies have defined or measured educational quality in Mongolia.Second, developing proxy measures based on factors contributing to the perceived quality of education in Mongolia could be significant for developing indicators to measure and improve certain aspects of educational quality, especially in public schools.Thus, this study attempted to develop factors associated with the perceived quality of education in Mongolia by identifying important issues conducive to the educational quality and analyzing them using principal component analysis through a dimension reduction of important items.Here, we focused our study on secondary education in Ulaanbaatar, the capital city of Mongolia.We sought to answer the following research questions: "What factors contribute to perceived educational quality in Mongolia?" and "What proxy measures can be developed by identifying factors contributing to educational quality in Mongolia?"The significance of this study is to develop scales for measuring educational quality in Mongolia and utilize the survey results to facilitate policymaking in the field of education.The authors also intend to use these measures to build the capacity of Mongolian public schools to provide better educational services.This study has some implications for improving overall educational quality, especially for publicly funded schools in Ulaanbaatar, Mongolia.Finally, it can be applied to understand similar educational situations outside of Mongolia, that is, in other countries with similar educational contexts.

Literature review
The term quality is related to the context and shifting paradigms.Many scholars have discussed the quality of education and educational quality from different paradigm approaches.For instance, Tikly (2010) stated that the issues of educational quality could vary and can be prioritized differently depending on economic development, such as in developed versus developing countries.Barrett et al. (2006) identified the relative priorities of the educational system depending on national development: post-conflict or newly founded states emphasize curriculum; low-income countries emphasize primary schools; middle-income countries emphasize continuation, such as secondary schools and disadvantaged groups; and OECD countries focus on competencies, responsibility, lifelong learning, and sustainability.Similarly, Hanushek and Wößmann (2007) found more significant educational deficits in developing countries than in developed countries when examining educational quality as a cognitive skill.Many studies refer to educational quality in developing countries in the context of African and Latin American countries, possibly owing to the availability of data and studies.In this study, we attempt to fill this gap concerning educational quality issues in one of the Central Asian countries: Mongolia, a developing country with low-middle income (as classified in the World Bank Data Common, 2021) and a post-socialist, democratic country, where the educational system is in transition to adapt to the rapidly changing environment.
Previous studies have provided insights into understanding educational quality from various philosophical perspectives.For instance, Barrett and Tikly (2011) discussed two perspectives to understand educational quality: human rights and social justice perspectives.The human rights perspective emphasizes meeting individual students' needs and acknowledging their characteristics and experiences.The social justice perspective focuses on three dimensions: the inclusion, relevance, and democratic characteristics of education, which include participatory elements of democracy in the classroom and engagement in debates about educational values.Based on these premises, Madani (2019) analyzed the "Education for All" policies and argued that education policies should be flexible given the different educational contexts and sociocultural backgrounds.On the other hand, Széll (2013) views educational quality as related to efficiency, effectiveness, and equity, with teachers playing a crucial role in providing high-quality education.Széll (2013) identified factors affecting educational quality, including student socioeconomic backgrounds, attitudes, and motivations, school infrastructure, atmosphere, teacher professional training, attitudes toward teaching, motivation, and cooperation.Hence, educational quality can be regarded as pertaining to human rights, social justice, effectiveness, efficiency, and equality.
Another perspective for understanding educational quality is the human capital perspective.Scholars maintaining this perspective posit that well-educated individuals significantly contribute to a country's economy.Previous studies have illustrated that the quality of education, as identified by the higher cognitive skills of students, greatly impacts the economic growth of both developing and developed countries (Hanushek & Woessmann, 2008;Hanushek and Woessmann (2020).Castelló-Climent and Hidalgo-Cabrillana (2012) also found that educational quality significantly affects economic prospects, as determined by the choice to study further, such as enrolling in higher education or increasing investment in developing human capital.Hanushek and Raymond (2002) emphasized school quality in terms of economic growth and human capital to provide quality education.Recently, Fomba et al. (2022) corroborated that a deterioration in institutional quality or effective public spending and governance of the school system can affect educational quality.Thus, the human capital perspective is viewed to be important in understanding educational quality.
Moreover, educational quality can be understood as an observable and measurable concept through outcomes reflected in performance indicators.For instance, Mortimore and Stone (1991) suggested measurable components such as buildings and resources, pedagogy as teaching quality, achievement outcomes as tests and exams, students' attitudes and behavior, and long-term achievements such as employment rate.Performance measurement was, however, seen to hold schools accountable for "good" quality of schooling (Ogawa & Collom, 2000) and is usually assessed based on the outcomes in relation to purposes and processes (Banta & Borden, 1994).Such an outcome-based approach fell short of fully understanding educational quality, as Bagnall (1994) argued that performance indicators are "dehumanizing and educationally trivializing" from the lifelong education perspective.Garira (2020) emphasized the interconnectedness of many layers of education systems and the role of stakeholders in educational quality.Thus, outcomebased approaches such as performance indicators may not necessarily account for educational quality.
There are some factors that account for the educational quality.For instance, student achievement is central to understanding the quality of education.At the international level, tests such as Programme for International Student Assessment (PISA) or Trends in International Mathematics and Science Study (TIMSS) tend to serve as thresholds for measuring student achievement, thereby implying good-quality education.Developing countries may easily adopt these tests, as they may indicate what we perceive to be the "good educational quality" of the developed societies (Scheerens et al., 2011a).Teachers and academic resources have been found to contribute to student achievement.It was found that student achievement measured by the PISA and TIMSS tests has a strong positive correlation with teachers with strong mathematical backgrounds and students with a prior background in formal mathematics education and high family academic resources (Carnoy et al., 2016).The number of books and teacher-centered methods was significant for student success, as measured by mathematical skills (Yayan & Berberoglu, 2004).Family structure also correlates with educational achievements (Cameron & Heckman, 2001;Patrinos & Psacharopoulos, 1995).Other studies documented student characteristics such as personality (De Fruyt & Mervielde, 1996), gender (Legewie & DiPrete, 2012), prior educational experiences, class participation (Oghuvbu, 2010;Partington & Gray, 2003;Gray & Beresford, 2008), extracurricular activities (Cooper et al., 1999), motivation (Tokan & Imakulata, 2019), attitudes to learning and school (Kpolovie et al., 2014), creativity (Gajda, 2016), parental involvement, and student socioeconomic background (Benner et al., 2016;Liu et al., 2020).These documented characteristics were all found to be conducive to student achievement.As shown in the literature reviews, educational quality is the interplay of many nested factors characterized by student achievement and factors contributing to student achievement.Hattie (2008) contributed significantly with his/ her meta-analysis of student achievement by identifying factors at five levels: student, home, teacher, curriculum, and school.The present study draws on these factors to explore educational quality, particularly as guides for interviews and documents for qualitative surveys.
In addition, school effectiveness is also viewed as one of the important indicators of educational quality.Scheerens (2000) argued that school effectiveness is a causal concept in which schools are compared, and its characteristics are considered conducive to better school performance, with the implications of quality education (Scheerens, 2000) and school accountability (Hanushek & Raymond, 2002).Fuchs et al. (2022) found that the student year is an important factor in perceived service quality in higher education.A review of service quality and higher education performance indicators found that resources, student-centered education, impactful research, and stakeholder engagement measure service quality (Camilleri, 2021).Similarly, several factors have been identified to characterize a "good" school as an indication of good performance.Galiani et al. (2002) estimated that school decentralization significantly enhanced student performance in public schools.In a literature review of studies conducted in developing countries studies between 1990-2010, Glewwe et al. (2011) synthesized several factors associated positively with the quality of education: school infrastructure-related factors such as libraries, desks, chairs, walls, etc.; teacher characteristics such as educational level, knowledge, and experience; and school organization factors such as student-teacher ratio, group work, and meals at school.School effectiveness studies examine educational quality and the provision of quality education in relation to students' achievements and educational attainment.Schools foster students' educational achievement, mainly in terms of measurable academic performance.However, as educational quality is embedded in values, needs, and societal expectations, stakeholders can define it differently.Hence, educational quality is questioned by many stakeholders as to which factor is important or should be prioritized as conducive to "good" educational quality and to our perceived educational quality.The purpose of this study was to explore the important factors conducive to the perception of educational quality by stakeholders in Mongolia and analyze these factors using principal component analysis.

Overview of education in Mongolia
Mongolia is a landlocked country between Russia and China with a population of over three million and a 1.5 million square km landscape.The capital city is Ulaanbaatar, where approximately 1.5 million people reside (approximately half of the population).There are 21 aimags or provinces, with over 300 soums or counties in the country and sparsely populated residents.According to the World Bank Data Common (2021), Mongolia is a lower-middle income developing country.The country is usually referred to as a post-socialist country following the collapse of the Soviet Union and as a former satellite country under the socialist regime.Being a formerly socialist country, many spheres of society have been adopted and followed similar structures that of socialist countries in sectors such as education, health, government, and the economy.As of 2020, the GDP of Mongolia was around USD 4,000 per capita, with a labor force participation of over 60 percent and a Gini index of 32.7 as of 2018 (World Bank Data Common, 2022).
Mongolia's educational system was adopted in the 20 th century, introducing a formal schooling system similar to that of the Soviet Union.Children started school at the age of 8 and completed eight years of compulsory education.After this, some students continued their education at a vocational school, while others chose to further their studies through 10 th grade.To continue their higher education, students were required to complete ten years of study at secondary schools.There were several Russian schools where students studied all courses in Russian.There were also specialized schools for disabled students and some for talented students.Both primary and compulsory education were offered for free to all children, as all schools during that time were publicly or government funded.The government played a central role in school management, administration, funding, and curriculum.Teachers were trained at a university designated specifically to train teachers and were assigned to schools in urban areas and the countryside.Generally, all decisions regarding teacher training and teachers' working assignments through school-based issues relied on the guidance of the central authority.Schools were centrally managed, and education in Mongolia was made accessible to everyone, with few dropouts and a literacy rate of 99 percent as of 1990 (Smith, 2008).
Since 2000, the educational system has undergone incremental changes.The ten-year schooling shifted first to 11 years and subsequently to 12 years in 2008 (UNESCO, 2009).The administration of the educational system was semi-centralized compared with the administration in the socialist period.The Ministry of Education and Science provides guidance for policy implementation and regulations.Schools have authority over teacher hiring, firing, and other regulations, whereas the government directs teacher remuneration.Schools in aimags are mostly managed by administrations at the provincial level.According to the Statistical Office of Mongolia ( 2022), as of 2021, there were 848 secondary schools with 35, 110 full-time teachers providing educational services to 712 000 students.Schools are divided into public schools (682 public schools), fully funded by the government, and private schools (166 private schools), subsidized by students' tuition and fees.General education consists of elementary schools lasting from 1 to 5 th grade, middle schools from the 6 th to 9 th grade, and high schools from the 10 th to 12 th grade.The class size was estimated to be 30 pupils per class, with an average of 31 in public and 20 in private schools.The teacher-student ratio is estimated to be 22 students per teacher in Ulaanbaatar, a minimum of 14-16 in some provinces and a maximum of 24 in others (Statistical Office of Mongolia, 2021).
Recently, parents and educators have begun to condemn the large gap in educational services provided in public and private schools.Jargalsaikhan (2019), a journalist and economist, documented that education is becoming a privilege in the light of private schools and kindergartens, causing societal inequality between public and private school students.The dire situation of education in public schools can be seen in students' lack of necessary facilities, good schooling conditions, and "good" teachers.Many pupils are studying in three shifts per day in a single classroom.Moreover, he lamented insufficient funding for public schools, while private schools received some government funding in addition to student tuition fees.This raises the question of systematic inequality in education and impediments to opportunities for Mongolia's children and youth.On these grounds, we also intend to develop proxies for public schools to strengthen their capabilities to improve educational conditions and provide students in public schools with better learning and higher-quality education.

Research methods
This study employed two methods.First, we employed a qualitative method to explore issues related to the quality of education in Mongolia and developed a questionnaire for principal component analysis.The rationale for choosing a qualitative method before a quantitative one is based on the fact that the idea of the quality of education is embedded in its cultural, economic, and sociopolitical environment.This helped us identify the important factors likely to contribute to the educational quality in Mongolia.For this phase, we collected online sources for one month using keywords such as educational quality, quality of education and students, educational quality and teachers, educational quality and students, educational quality and evaluation, teacher quality, student achievement and quality, learning and teaching quality, teacher education, school effectiveness, and learning environment.The qualitative data collection yielded 68 online sources.
Upon completing the qualitative data, we conducted a content analysis to identify the most pronounced issues pertaining to the quality of education in Mongolia.Additionally, we interviewed three experts in the field of education regarding their opinions on educational quality in Mongolia.The interviews were conducted during the 2021 pandemic.Interviews were used to clarify the issues identified in the online sources.Based on qualitative data from interviews and online sources, we developed 123 questionnaire items for the quantitative analysis.
Second, we employed principal component analysis or exploratory factor analysis to examine and identify the factors indicating the quality of education in Mongolia.A principal component analysis aims to identify several different constructs based on given measures and their correlation coefficients (Fabrigar & Wegener, 2011).In other words, it is used to reduce relative dimensions based on the correlation analysis and identify factors or components to measure, in our case, the quality of education in Mongolia.

Questionnaire administration
The questionnaire was administered from late September to December 2021 using the survey collection portal www.surveymonkey.com.The total sample size was 412, yielding 323 valid samples with missing data for each item below .05.Schafer (1999) suggested that data with a missing value of less than .05or completion of a 95 percent minimum should be safe to proceed with relevant analyses.
The questionnaire consisted of 5 demographic questions and 123 items demonstrating possible factors that may contribute to the perceived quality of education in Mongolia.The items were initially structured into six broad themes of educational quality: teachers, students, school environment, school administration, curriculum, and parents.Each theme was assigned a different number of items based on the content analysis and issues identified by the qualitative study.The "teacher" theme consisted of 35 items; the "student" theme consisted of 29 items; the "school environment" theme consisted of 26 items; the "school administration" theme consisted of 16 items; the "curriculum" theme consisted of 11 items; and the "parents" theme consisted of 6 items.The items assigned to each theme also relied on previous literature, as discussed in the literature review section.
The questionnaire consisted of 123 items, which was relatively long for the participants.Here, we expected the respondents to drop some of the insignificant items.It was also mentioned that "The longer the questionnaire, the less likely potential participants will be to volunteer for the study or to complete all the items" (Converse & Presser, 1986, as cited in Worthington & Whittaker, 2006, p.814).Hence, we expected that some unimportant factors for the respondents may not be evaluated in the survey.
We collected data through SurveyMonkey's website due to the pandemic situation and concerns about participants' convenience.In the heading section of our survey, we notified the respondents that the data will be used only for research purposes and that personal information will not be shared with any third party.We asked the participants to rate each item in terms of how it can contribute to the perceived quality of education in Mongolia.Respondents were asked to rate each item according to its importance, ranging from 1 (not important) to 6 (extremely important)."

Data descriptions
A sample size is considered to be important for the exploratory factor analysis.To determine the sample size in our study, we followed the recommendations of the principal component analysis.Rouquette and Falissard (2011) recommend a minimum sample size of 300, whereas Goretzko et al. (2021) recommend a minimum sample size of 400 when running principal component analysis.Based on the recommended sample size, we collected data using online survey platforms and shared the link for voluntary participation until the recommended sample size was reached.
Although we managed to collect 412 responses to ensure the robustness of the analysis, valid responses were estimated to be 323, which still satisfied the requirement for the analysis.
Our data comprised 323 valid samples out of the 412 responses.There were 76 male respondents, 246 female respondents, and one unanswered respondent.There were 194 respondents aged between 12 and 20 years, 70 respondents aged 21-30; 34 respondents aged 31-40; 11 respondents aged 41-50; 5 respondents aged 51-60 years, and five respondents aged 60 or over.Four respondents did not report their ages.Regarding employment status, 33 were teaching staff, 212 were students, two were unemployed, 11 were self-employed, 14 worked in government organizations, 22 were employed in the private sector, one was in the field of farming/herding, 3 were pensioners, 4 were housekeepers, and 21 answered "other."As for the educational level, 28 reported having a master's degree or above; 105 reported having a diploma or bachelor's degree; 4 had vocational education; 176 had secondary education (12 years of education), 3 had compulsory education (8 years of education); 2 had primary education; and 1 answered "other."Four respondents did not provide their educational levels.We also asked for their email contact addresses to answer further questions.

Data analysis
We used the principal component analysis or exploratory factor analysis in this study as the purpose was to examine and identify the factors underlying the perceived quality of education in Mongolia.The tool used for data analysis was SPSS 25.
Initially, we prepared our data to determine a suitable rotation for extracting components for each theme.Rotation is defined as "performing arithmetic to obtain a new set of factor loadings from a given set" (McDonald, 1985, p.40).Tabachnick and Fidell (2007) suggested that oblique rotation be employed first to decide between orthogonal and oblique rotations.If the correlation matrix yields items that are more than .32,then it should be assumed to have an overlap of 10 percent more to run an oblique rotation.Hence, we proceeded with oblique rotation analysis based on the above suggestion and our assumption of correlation among items in each theme.
Subsequently, we performed our analysis by exploring the themes individually.Our survey asked questions on six themes to rate each item in terms of its significance to the perceived quality of education in Mongolia.The themes drawn from the qualitative study were the school environment, teachers, students, curriculum, school administration, and parents.
We report the results by examining the Kaiser-Meyer-Olkin (KMO) and Bartlett's test, communalities, total variance, scree plot, and reporting on the pattern matrix.Data analysis procedures are reported in the results section.First, we examined the KMO measure of sampling adequacy, which measures whether the data are appropriate for factor analysis.The most values for each theme indicated "marvelous" sampling adequacy, which satisfies the criteria for further analysis (Kaiser, 1974).Each value is reported in the corresponding theme.Next, we examined the communalities coefficients to decide whether to retain or eliminate items.The corresponding analyses for retaining and eliminating items were reported for each theme.Here, in the result section, we begin reporting the results in order of the most-rated theme or theme with the greatest sample size.

Results
The first theme was the school environment, with the largest sample size (N = 323).We extracted 17 items from the initial 26 items with two components: "Facilities for a better learning environment," yielding ten items, and "School interior and exterior design," yielding seven items.The loading criterion was .50, in which we assumed significant items rated from four (fairly important) to six (extremely important) as a clear indication of contributing factors to the perceived quality of education.The total explained variance was 56.389, with a variance of 47.765 for the first component and 8.623 for the second component.The KMO and Bartlett's tests were estimated to be .950,which indicates marvelous sampling adequacy (Kaiser, 1974).The Cronbach's alpha score indicated a reliability of .931(Table 1).The coefficients with the highest loading, .806"friendly learning environment" and .794"safe school environment," are the most important items for respondents to provide better learning environments for students.As recommended by Worthington and Whittaker (2006), communalities with a sample size of less than 200 should contain a set of factors loadings higher than .50.In our case, with a sampling of over 300, the communalities of the items estimated to be around .51 to .64 meet the above-suggested criteria for the analysis.
The second theme was school administration.We extracted seven items from an initial 16 items, with one component named "school administration activities."The loading criterion was .50,assuming that the contributing factors were rated four or more.The total explained variance was 58.852.The KMO and Bartlett's tests were estimated to be .901,indicating marvelous sampling adequacy (Kaiser, 1974).Cronbach's alpha score indicated a reliability of .883.(Table 2).As the principal component analysis was employed for dimension reduction and the extraction of important items, we applied both criteria of communalities of the items above .50and factor loadings above .50.First, we examined the coefficients of the communalities and eliminated them based on the criteria above.We also examined component matrix loadings above .50.This analysis did not yield a pattern matrix but a component matrix, as it yielded one component.The items extracted indicate the roles that school administrations should perform to improve perceived educational quality.As we can see, the items for school service or effectiveness are viewed as important factors to contribute to the quality of education in Mongolia.
The third theme (Table 3) for extraction was on students.Under the "students" theme, we asked respondents to rate 29 items based on the importance of the items that contribute to the perceived quality of education.This theme contained the second highest number of items as compared to the other themes.We first attempted to extract items with the same criteria communalities at .50 till all items reached the desired saturation level.We ran the analysis multiple times with the above assumptions.However, the desired level was reached with only a few items, eliminating 21 out of 29 items and leaving eight items for the theme relevant to the students, which is one of the most important factors contributing to the perceived quality of education.This led us to run the extraction with less loss of information in our data and more reasonable items, both theoretically and practically.As we were looking at the yielded components that lost much information from our data, we decided to select items with the same loading at .50 but accepting coefficients for communalities at .40.Moreover, the principal component analysis primarily aims to decrease the number of items, while maintaining as much of the initial item variance as possible (Park et al., 2002;Worthington & Whittaker, 2006).Therefore, we repeated the extraction.
The extraction of the items reached saturation levels after several attempts.The items with loadings of .50 and communalities coefficients of .40 were extracted for the components.We eliminated items with a factor loading of .50.The first extraction yielded 11 items within the acceptable range, which were eliminated by examining the communalities coefficients.We ran the second extraction with the remaining 18 items and eliminated six items with the same criteria.In our third attempt, 16 items loaded above .50,and two items were eliminated.Again, we repeated the same procedure with 16 items for extraction, which yielded four components with items above .50and a communalities coefficient of .40.The results are shown in Table 3.The total explained variance is 60.652.The KMO and Bartlett's tests were estimated to be .813,which is "meritorious" (Kaiser, 1974).We extracted four components named "student's characteristics" with eigenvalue 4.378, "preference for gender equal treatment" with eigenvalue 1.867, "student's family condition" with eigenvalue 1.155, and "student's capacity" with eigenvalue 1.091.The component "student's characteristics" consists of 7 items, whereas the other three components consist of 2 and 3 items.We opted to report and retain this loading even with two items.Worthington and Whittaker (2006) argued that "it is possible to retain a factor with only two items if the items are highly correlated (i.e., r > .70)."Moreover, we retained the loadings while considering conceptual understanding, particularly for perceived educational quality in Mongolia.The loadings for these factors are estimated to be significant with coefficients .920and .900for the component "preference for gender equal treatment," .856,.839for the component "student's family condition," and .786,.720,and .700for the component "student's capacity." Table 4 shows the results for the theme of parents and guardians.The analysis yielded one component with no extraction based on the loading criteria .50,eigenvalue .3621above .1,and the KMO and Bartlett's test with an estimation of .847for sampling adequacy.The theme, "parents and guardians," is loaded with six items.The Cronbach's alpha score indicated a high reliability of .865.We applied the same criteria while extracting the items loading at .50 and communalities coefficients at .50.The analysis yielded a component matrix because the pattern matrix was not extracted for one component.This component shows the importance of parental and guardian support in many aspects of student life for the perceived quality of education.This component emphasizes the role of supportive parents and guardians in academic achievement, which indicates the cognitive aspects of accomplishment.On the other hand, a similar loading for "parents and guardians should work in collaboration with teachers to become a well-rounded, self-reliant individual" shows the importance of parental and guardian involvement not only in cognitive aspects but also in building character in their children.
Table 5 shows the results for the theme of the curriculum.The analysis yielded two components with the initial 11 items.The first component is named "general curriculum characteristics," with seven items indicating what can generally be contributing factors to the perceived quality of education.Textbook-related items loaded the highest at over .80,and communalities coefficients at .60.The second component is named "curriculum reflecting national interests" and loaded with four items, indicating parental and guardian voices in designing the curriculum.The criterion for extraction is .50,and the communalities coefficient is also at or above .50.All the items satisfied these criteria.Other criteria were satisfied, including eigenvalues of 5.425 and 1.315 for the first and second components, respectively.The sampling adequacy of the KMO and Bartlett's test was marvelous-.917 with a sample size of N = 304 (Kaiser, 1974).Cronbach's alpha was estimated to be .893.The total explained variance is 61.278.
Table 6 presents the results of the teachers' theme.This theme contained the highest number of items (35 items of our questionnaire) as the topic of teachers was an extensively discussed issue in the qualitative sources and literature.We applied the criteria for loadings at or above .50and communalities coefficients at or above .50for this theme.First, we looked at the communalities coefficients and eliminated three items, retaining 32.The sample size was 296.We repeated the analysis for the second round, and one item was eliminated based on our initial criteria for communalities, retaining 31 items.The sample size was increased to 298.In the third round, we again looked at the communalities coefficients.While the criteria for communalities were met, those for loadings in the components were not.At this point, we eliminated items with loadings below .50 and eliminated ten items, retaining 21 items.During the analyses, we examined the items in terms of their meaning to the respondents and how they rated them.We ran the fourth round of analysis using the 21 retained items.Again, this time, the communalities coefficient fell below .50 for the four items.We again looked at the meanings of the items and the possibility of them being retained or eliminated.For this analysis, the sample size increased to 301.We retained 17 items for the analysis.At this point, we eliminated two items by looking at the communalities coefficients, which fell below the criteria.The sample size was 302.Here, we ran the analysis again with 15 items and reached saturation with 13 items, eliminating two items that did not meet the criteria for both the communalities coefficients and loadings.The sample size was increased to 303.The initial 35 items were then reduced to 13 significant items, with three components named "teacher's general characteristics," "using evaluation," and "teacher's learning."The final extraction results are illustrated in Table 6.Although we extracted 13 out of 35 items, the sample size increased slightly from 296 to 303, indicating small changes in the rates of items that contribute to the quality of education.All remaining criteria for the extracted components were met.Eigenvalue for the first component, "teacher's general characteristics," was 5.626; the second component, "using evaluation," was 1.451; and "teacher's learning" was 1.007.The total variance was 62.189.The variance for the component "teacher's general characteristics" was 43.280, the second component "using evaluation" was 11.163, and "teacher's learning" was 7.746.KMO and Bartlett's test was estimated to be .912,indicating marvelous sampling adequacy (Kaiser, 1974).Cronbach's alpha score was .887,indicating good reliability.The first component extracted six items reflecting general characteristics with positive loadings, the second extracted three items with positive loadings, and the third extracted four items with negative loadings, indicating negative correlations with the two components.The results of the negative loadings may be due to the length of the questionnaire, as the teacher's theme included 35 items with abstract and concrete terms with overlapping meanings.Also, it can be attributed to the fact that items for teachers' general characteristics included more abstract terms, which, in turn, reflect the ideas of teacher learning and teacher development.Moreover, it may be that the respondents take the idea of teacher learning and training as part of what they should do.
In summary, we extracted 68 items from the 123 items in our questionnaire.The analysis yielded 13 components meeting all criteria, such as loading coefficients .50,communalities coefficients .50,KMO and Bartlett's test for sampling adequacy being significant at p = .000above .80,and eigenvalues above 1.We also looked at the sample size, as it reported the ratings for factors contributing to the quality of education.The theme with the largest sample size (N = 323) was school environment, which yielded two components with 17 items.The next largest component was school administration (N = 312), which yielded one component with seven items.The next theme was students (N = 308), yielding four components with 14 items.The fourth theme, parents and guardians (N = 308), yielded one component with six items.The fifth theme was curriculum (N = 304), which yielded two components with 11 items.The last and least rated theme was teachers (N = 303), yielding three components with 13 items.The principal component analysis was used to reduce the dimensions of our data to better understand the variables contributing to the perceived educational quality in Mongolia, thereby developing proxy indicators that could possibly contribute to perceived educational quality in the country.

Discussions
This empirical study explored contributing factors to the perceived quality of education in Mongolia.The purpose was achieved first by examining and identifying crucial contributing factors to perceived educational quality in Mongolia and, second by developing a proxy-measuring instrument for assessing educational quality through its significant factors in Mongolia.Moreover, we believe that the current study can help policymakers prioritize issues of concern regarding educational quality, specifically to improve the conditions of public schools in Mongolia.We also believe that the issues of educational quality, in terms of its contributing factors in the context of developing countries, have filled the gap in previous studies conducted mainly in African and Latin American countries.In line with other scholars (Tikly, 2010;Barrett et al., 2006), this study demonstrates important issues and factors that should be prioritized in the educational context of Mongolia.In our study, we can see that the most important issue is the provision of good schooling conditions which are reflected in the themes of the school environment and school administration, extending the categorization of priorities suggested by Barrett et al. (2006), thereby emphasizing primary schools as well as school conditions for secondary education.Although the missions, worldviews, and policies for education in Mongolia may appear to align with the current mainstream educational views and goals, the infrastructure of education-schools and school services -may not meet the desired standards or expectations among stakeholders perceived to be conducive to a better quality of education.
The theme for students suggests general characteristics, such as student motivation (Tokan & Imakulata, 2019), attitude toward studying (Kpolovie et al., 2014), self-expression, openness, and personality (De Fruyt & Mervielde, 1996), which are important for the perceived quality of education.The results suggest equal treatment for students, especially with regard to gender, with no emphasis on students being male or female.Items loaded in the student's capacity component indicates the student's maximum potential and ability that can be espoused through all sorts of competitions and exams at domestic and international level.It is consistent with previous studies on student achievement, such as PISA and TIMSS (Scheerens et al., 2011b).Moreover, the role of the family was also found to play an important role in the perceived quality of education, which we can also see in the confirming results in the parent and guardian theme.Parents and guardians, especially when supporting their children emotionally and working closely with teachers for student achievement, can contribute significantly to the quality of education.For example, Benner et al. (2016) found similar factors, such as parental involvement and students' socioeconomic background, to correlate positively with student achievement.Our results are consistent with the characteristics suggested by Széll (2013).
Our analysis of the theme of teachers did not yield significant positive coefficients for professional training, which was puzzling.The items loaded in the component "Teacher's learning" suggest the presence of negative aspects or potentially unnecessary aspects of teacher's development that may not necessarily contribute to the educational quality in Mongolia.Similar results have been documented in previous literature, as in-service teacher training decreases students' time in school, just as new buildings increase their time at school (Glewwe et al., 2011).Another interpretation of our results can be that the item "teacher's knowledge and skills" loaded in the component named "Teacher's general characteristics" may have suggested teacher training or learning to respondents due to the general nature of the terms "knowledge and skills."Moreover, the result of this loading may also imply better training for teachers before they start their careers than after entering the job market.Sahlberg (2011) emphasizes "professional educators" comprehensive and rigorous preparation with well-balanced knowledge and skills.Lastly, the component "Teacher's general characteristics" suggests a close engagement with students and an unbiased approach to their relationship with students.
The extracted items for the curriculum suggest that textbook availability and quality for students should be prioritized for better educational provisions in Mongolia.It was also observed that issues around textbook availability and quality had been raised occasionally among stakeholders in Mongolia, especially in public school contexts.As part of the curricular implementation, textbooks are extracted to play an important role in conveying knowledge and "conditioning" students' learning.Benavot (2011) argues that quality education requires context-specific knowledge of textbooks and curricula.Such context-relevant attitudes toward curricula were also loaded as important, as items such as reflecting the opinions and initiatives of parents and guardians in curriculum development and their involvement in designing curricula indicate the context sensitivity of the curriculum.Moreover, we also see that the curriculum contributes to the perceived quality of education by fostering students' autonomy and experiential learning and encouraging well-rounded individuals with morally humane and cognitively capable characters.

Conclusion
In conclusion, our study identified crucial factors affecting the perceived education quality in Mongolia.Quality of education or educational quality is a lucid term perceived differently by stakeholders depending on social, cultural, political, and economic circumstances.Understanding educational quality through possible contributing factors in Mongolia, a developing country, can help improve the educational situation, especially in public schools.Further studies should confirm the measurements obtained in this study and develop more comprehensive instruments for use as guidelines.Policymakers should pay attention to strengthening public schools by providing better schooling conditions, such as improving the school environment, building necessary facilities, facilitating more parental and guardian involvement in school activities and curriculum development, and providing equal opportunities for students, such as access to good quality textbooks, fostering teacher accessibility to students and parents, nurturing an unbiased approach in relationships between students and teachers, and focusing more on teacher preparation than teacher training in-service.
This study had some limitations.First, it can be limited to country-specific situations but not limited to similar conditions.Second, there was a tendency for the components to be loaded more in the first component, yielding higher eigenvalues.