Gruodžio 1-3 dienomis vykusioje tarptautinėje konferencijoje „Data Analysis Methods for Software Systems“ pristatytas stendinis pranešimas „Principal Component versus Data Envelopment analysis in Construction a Composite Indicator for Education Monitoring“.
Abstract. We can monitor various indicators of education system over time and rank periods of each indicator, but we can’t rank the countries or the periods of all indicators. It is not possible to rank if we do not aggregate the indicators. Composite indicators are very useful for this purpose. This research focuses on the construction of a composite indicator for the education monitoring. We went through following five stages for construction of composite indices: data treatment, data normalisation, weighting, aggregation and comparing the indices. At the first stage we used single imputation for missing data. More over all indicators were treated as the profit type – “the larger the better“. At the second stage we standardized data by subtracting the mean of the data and dividing by the standard deviation. At the data weighting stage we used two different methodologies in order to compare how the different approaches affect the results. The first one is principal components analysis, which groups individual indicators which are collinear to form a composite indicator that captures as much as possible of the information common to individual indicators. The second one is the application of data envelopment analysis known as the “benefit of the doubt” approach, which selects weighs so as to maximise the composite indicator for each country. Composite indicator for education monitoring was calculated for Lithuania, Latvia, Estonia, United Kingdom, Finland and Germany over time using data from EUROSTAT and OECD databases. The index was constructed following structural CIPO framework, which describes relationships between Input, Process and Output in education within a certain Context. This model includes comprehensive information of education system.
Spalio 21-22 dienomis habil. dr. Rimantas Želvys skaitė plenarinį pranešimą „PISA phenomenon: the many faces of the international student assessment“ tarptautinėje konferencijoje „Education Policy and Culture: Consistent and Radical Transformations“, kuri vyko Vilniuje.
Abstract. The Programme for International Student Assessment (PISA) has become an extraordinary phenomenon of the contemporary educational world. Initiated in 2000 by OECD, now it includes 65 countries and teritories from all over the world. PISA is the only international education survey to measure the knowledge and skills of 15-year-olds, an age at which students in most countries are nearing the end of their compulsory time in school. Subject to severe criticism as well as the source of inspiration for educationalists, policymakers, journalists and other interest groups, PISA plays a significant role in contemporary educational landscape. No matter whether we are supporters or critics of PISA, it‘s difficult to deny the scope of the influence this international project. However, when discussing about PISA, opponents often consider various aspects of this phenomena, therefore the arguments for and agains PISA are directed towards different dimensions of the survey. The aim of the presentation is to identify multiple aspects, or, speaking metaphorically, faces of PISA, which carry different messages and are subject to different value judgements by various interest groups.
1. PISA as a symbol of globalization. The global education reform movement is gaining momentum and in this respect PISA symbolizes the global trend of seeking for universal standards and common ways of development iof education systems. Global educational space enables making global measurements of national education systems and PISA provides an opportunity for participating countries to become comparable on a global scale.
2. PISA as manifestation of neoliberal ideology in education. Neoliberal ideology, which directs education towards greater efficiency and accountability, competition and market orientation, is reflected in many educational initiatives of the last decades, and PISA is a typical example of trying to make national educational systems compete. Proponents of neoliberal ideology often refer to PISA when they urge to assess the effectiveness of provision of educational servises.
3. PISA as a methodological controversy. On one hand, PISA uses sophisticated methods of sampling and statistical analysis, and on the other it often makes you wonder how valid are conclusions which are made as a result of a two-hour paper-and-pencil testing of students.
4. PISA as research data base. PISA provides a vast research data base, where researchers can follow a progress of participating nations in literacy, science and maths achieved during the period from one survey to another, compare different countries, regions or educational models. No matter whether we accept the ideological rationale of PISA or not, the data base is a valuable resource for researchers in the field of comparative education.
5. PISA as benchmarking. PISA has been acknowledged as an official benchmarking tool by the European Commission and member states are urged to follow the target of 15 year-old students achieving a certain level of basic skills in reading, maths and science by 2020. PISA results are also considered as one of the educational indicators for the OECD member states.
6. PISA as a league table. Perhaps the most familiar face of PISA for the wide public is that of a league table. Without getting much into details, politicians, journalists and authors of popular publications provide information and commentaries about the place the country takes in a list of participating countries in the fields of reading, maths and science. Often we can observe public discussions whether the country is leading or laging behind when compared with other nations of similar level of socio-economic development.
7. PISA as promotion. Education is not among the most popular topics for mass media, therefore one of the occassions when education is in the focus of everybody‘s attention is the day when PISA results are announced. PISA results help to maintain public interest in education at keep education on the agenda of national politicians at least for some time.
8. PISA as punishment. PISA results are often use as an argument in order to prove the ineffectiveness of those in charge of national education. Critique is not allways adequate as PISA results are announced several years after the survey was conducted and unsatisfactory results could be the result of wrong policy decisions made by previous political leaders. However, oponents usually use national failure in PISA (and usually it‘s considered a failure, even if the country shows average results) as a strong argument against those who are currently in charge of education.
9. PISA as business. PISA requires a wide array of different human and material resources and attracts a large number of temporary or permanent employees. For each PISA survey, international contractors (usually made up of testing and assessment agencies) are responsible for the design and implementation of the surveys. There are also many consultancy firms which claim to know how to improve PISA results and offer their services for potential customers.
10. PISA as policymaking. PISA data is used as a basis for further development of national educational policies. Policy papers often refer to PISA when defining possible trajectories of education reforms. International bodies urge national states to use PISA data for policymaking and making efforts to improve education. For example, the European Commission in its Education and Training Monitor for Lithuania refers to PISA results and states that „so far there have been no concrete government initiatives to address either the relatively poor performance in basic skills or gender differences in educational performance“.
Different actors in educational domain should use in a more skillfull and selective way the various aspects of PISA for achieving their goals and securing their interests. Educational researchers should be more involved in using PISA as a research data base and use all the opportunities it can provide. Those involved in educational policy should analyse the impact of PISA on educational policy decisions and provide guidlines for future policy development. Representatives of mass media should use PISA results as a means of raising public awareness of the importance of education for national development.
D. Stumbrienė skaitė pranešimą „Constructing a Composite Indicator for Education Monitoring“ rugpjūčio 23-26 dienomis vykusioje tarptautinėje konferencijoje COMPSTAT 2016 (The 22nd International Conference on Computational Statistics) Ovjedo mieste Ispanijoje.
Abstract. This paper focuses on the construction of a composite indicator for the education monitoring. It is common awareness that socio-economic phenomena are complex and cannot be measured by a single descriptive indicator – it should be represented with multiple dimensions. Phenomen such as education can be measured and evaluated by applying methodologies known as composite indicators. This paper reviews methods to create the education monitoring index that apply data treatment and normalisation procedures, weighting and aggregation strategy, which assigns weights to the components when combining them and chooses a synthetic function. At the data weighting stage we used factor analysis and data envelopment analysis in order to discuss how the different approaches affect the results. In order to compare the different methodologies, the education monitoring index was calculated for Lithuania, Latvia, Estonia, United Kingdom, Finland and Germany over time using EUROSTAT data. The index was constructed following structural CIPO framework, which describes relationships between Input, Process and Output in education within a certain Context.
Birželio 20-21 dienomis vykusioje Lietuvos matematikų draugijos LVII konferencijoje Dovilė Stumbrienė pristatė pranešimą „Švietimo stratifikacijos identifikavimas taikant daugiamatę regresiją: PISA 2012 duomenų analizė“.
Santrauka. Švietimo stratifikacija analizuojama remiantis akademine nelygybe (mokinio pasiekimai) ir socialine nelygybe (ekonominis, socialinis ir kultūrinis statusas). Šiame darbe pristatomas daugiamatės regresinės analizės taikymas švietimo stratifikacijai nustatyti. Skirtingų švietimo sistemų modelių palyginimui pasirinktos šešios šalys: Lietuva, Latvija, Estija, Suomija, Vokietija ir Didžioji Britanija. Tarptautinio tyrimo PISA 2012 (Programme for International Student Assessment) duomenų analizės metu nagrinėta, kaip socialinė stratifikacija įtakoja nevienodus matematikos mokymosi rezultatus bei kaip nelygybė pasireiškia per mokinio, mokyklos ir bendramokslių charakteristikas. Nustatyta, kad didžiausi mokinių matematikos pasiekimų skirtumai tarp mokyklų egzistuoja Vokietijoje, tuo tarpu Suomijoje ir Estijoje mokyklos yra labiau homogeninės, o mokinių pasiekimų skirtumai pasireiškia mokyklų viduje.
VU Matematikos ir informatikos instituto Sistemų analizės skyriaus (SAS) seminare birželio 6 dieną Dovilė Stumbrienė pristatė pranešimą „Education data analysis: Multilevel Regression and Constructing a Composite Index“.
Habil. dr. Rimantas Želvys skaitė pranešimą „Moving towards different models of the welfare state: education in the baltic countries“ konferencijoje CESE XXVII (Comparative Education Society in Europe), vykusioje Glazge, Škotijoje gegužės 31 – birželio 3 dienomis. Konferencijos santraukų knygą galima rasti čia.
Abstract. The education systems of European countries can be analysed using the distinction along different types of welfare states. The three ideal-types identified by Esping-Andersen (1990) served as a theoretical framework for the current study. In his classical work „The Three Worlds of Welfare Capitalism“ (1990) Esping-Andersen describes the Scandinavian universalistic, Continental corporatist, and Anglo-Saxon liberal models. Some authors, in particular, Fenger (2007), claim that post-socialist countries do not quite fit into the three-type model and point out Central and Eastern European countries as a separate subgroup. However, our assumption is that the process of convergence is taking place as a result of the European integration and the new EU member states are gradually moving towards one or another model identified by Esping-Andersen (1990). The purpose of our study is to trace the directions of development of the educational systems in the Baltic states. The three Baltic states – Estonia, Latvia and Lithuania – are the only three former republics of the Soviet Union which joined the EU in 2004.
We applied multilevel linear regression and descriptive analysis for the comparison of countries approaches and used the OECD PISA 2012 survey data in the study. For our comparison with the Baltic countries we selected three „old“ EU member states, namely, UK as representing the Anglo-Saxon liberal model, Germany as representing the Continental corporatist model and Finland as perhaps the most successful example of the Scandinavian model. We considered seven important aspects of the organization of the national school systems – ability grouping in schools, the level of school autonomy, the level of responsibility for resource allocation, the level of responsibility for curriculum and assessment, the index of quality of physical infrastructures, the index of quality of educational infrastructures, and the student-teacher ratio. Results indicate that Lithuania and to a lesser extent Latvia in most of the aspects of school organization described above tend to move closer to the United Kingdom, thus developing their systems of education in line with the liberal Anglo-Saxon model. Seemingly Estonia has chosen another – Scandinavian – model and Estonian school organization comes closer to that of Finland. Our general conclusion is that in some aspects all three Baltic states still demonstrate similar approaches to school organization but the process of differentiation among the three of them is gaining momentum.
Projekto vadovė dr. Audronė Jakaitienė skaitė pranešimą „Towards an education monitoring index“ Londone gruodžio 12-14 dienomis vykusioje tarptautinėje konferencijoje CMStatistics 2015 (The 8th International Conference of the ERCIM WG on Computational and Methodological Statistics).
Abstract. Information about education monitoring indices is very diverse and limited, especially over time. Given that education is one milestones, which guarantees the well-being of a country, the understanding about the education system status might be crucial. However there are plenty of education variables and a single index, which incorporates big data and gives notion about the system status, might be essential. The European Commission within the Education and Training Monitor initiative calculates indices towards Europe 2020 strategy. Each country’s indices are calculated normalizing data with respect to EU28 average. We develop a methodology to construct an Education Monitoring Index for a country using country or region data. For methodologically consistent data for a country, we use Europe 2020 strategy main indicators as starting dataset. The objective is to calculate an Education monitoring index for Lithuania, Latvia, Estonia, United Kingdom, Finland and Germany available over time.
Projekto vadovė dr. Audronė Jakaitienė skaitė pranešimą „Construction of Education Monitoring Index“ Druskininkuose vykusioje 7-oji tarptautinėje konferencijoje „Duomenų analizės metodai programų sistemoms“ (Data Analysis Methtods for Software Systems), gruodžio 3–5 dienomis.
Abstract. Given education is one of milestones, which guarantees long-run well-being of a country, the assessment of the situation in education system status might be vital. Researches calculated many indices those monitor diverse activity. However information about Education monitoring index is very variant and scarce, especially over time. A single index, which incorporates big data and gives notion about status of education system, might be relevant. European Commission within the Education and Training Monitor initiative calculates indices towards Europe 2020 strategy. Each country indices are calculated normalizing data with respect to EU28 average. Though calculated indices are not aggregated to a single index for a country. The objective of the research is to calculate Education monitoring index, as integral indicator, for Lithuania, Latvia, Estonia, United Kingdom, Finland and Germany available over time. We construct Education Monitoring Index for a country following structural CIPO framework, which describes relationships between Input, Process and Output in education within a certain Context. Data reduction methods are applied for an index construction. For methodologically consistent data for a country, we use EUROSTAT as data source.
Atspausdintas straipsnis „Švietimo duomenų tyryba: apžvalga ir tyrimų kryptys“ leidinyje Lietuvos matematikos rinkinys LMD darbai.
Santrauka. Straipsnyje pateikiama sisteminė literatūros apžvalga, kurios tikslas dažniausiai naudojami duomenų šaltiniai ir taikomi duomenų analizės metodai švietimo duomenų tyryboje. Apžvalgai pasirinkta tarptautinė duomenų bazė „Web of science“. Atmetus trumpus konferencijų pranešimus bei straipsnius, kuriuose nepateikti empiriniai duomenys, išsamiai nagrinėjami 14 straipsnių. Nustatyta, kad dažniausiai tiriamos besimokančiųjų duomenų bazės, kuriuose su mokomųjų dalykų įvertinimais pateikiama ir kontekstinė informacija. Švietimo duomenų tyryboje dažniausiai taikomi klasiﬁkavimo metodai ir kiek rečiau regresijos analizė bei klasterizavimas. Straipsnis baigiamas švietimo duomenų tyrybos apžvalga Lietuvoje.
D. Stumbrienė, A. Jakaitienė. Švietimo duomenu tyryba: apžvalga ir tyrimų kryptys. LMD darbai.
Projekto vadovė dr. Audronė Jakaitienė Lietuvos matematikų draugijos LVI konferencijoje, kuri vyko birželio 16-17 dienomis Kaune, skaitė pranešimą „Lietuvos švietimo sistemos indekso modeliavimas“, o Dovilė Stumbrienė pristatė parnešimą „Švietimo duomenų tyryba: apžvalga ir tyrimų kryptys“.