Conference Agenda

National educational databases play an increasingly important role in educational research, policy analysis, and system monitoring. Across many countries, Education Management Information Systems (EMIS) now contain detailed longitudinal information on students, teachers, schools, and educational trajectories. While these systems were primarily developed for administrative and policy purposes, they are increasingly used for research. At the same time, substantial differences exist across countries with respect to data structure, accessibility, legal frameworks, and possibilities for data linkage. This session brings together contributions from several national contexts (Denmark, Finland, Italy, Poland, Portugal, Sweden) to examine how education data infrastructures are used for research, and under which conditions they enable robust and policy-relevant analyses. The papers cover a range of educational systems, including national education registers, assessment databases, and links to international large-scale assessments, and illustrate both opportunities and constraints in working with large-scale education data.

A central theme of the session concerns data access and governance. Several contributions describe national arrangements for researcher access, including secure remote access environments, licensing procedures, and legal requirements related to data protection and privacy. Differences between countries highlight how institutional and regulatory frameworks shape what kinds of research can be conducted, and by whom. The papers also illustrate how anonymisation and pseudonymisation procedures affect analytical possibilities and raise methodological trade-offs between data protection and research quality.

A second theme is data linkage and methodological challenges. Many contributions demonstrate the analytical value of linking education data with other sources, such as national examinations, higher education registers, labor market data, or international large-scale assessments. These linkages enable the study of educational trajectories, persistence and dropout, inequalities, and policy effects over time. At the same time, the papers discuss challenges related to longitudinal consistency, missing data, hierarchical data structures, and the validity of administrative indicators, as well as the need for appropriate statistical models.

A third theme concerns the use of national education data for policy evaluation and system improvement. Several papers illustrate how administrative and assessment data are used to study educational inequalities, school effectiveness, student transitions, and dropout risk. The use of rich register data allows complex conceptual processes to be refined and examined empirically. Methodologically, the contributions range from descriptive statistical analyses to longitudinal modelling and quasi-experimental approaches addressing policy-relevant research questions.

Taken together, the session highlights both the research potential and the limitations of national education data systems. While these data offer unique opportunities for large-scale and longitudinal research, their use requires careful attention to data quality, governance, and methodological choices. By presenting experiences from different national contexts, the session contributes to a more informed discussion on how educational databases can be used responsibly and effectively for research and policymaking.

The Trends in International Mathematics and Science Study (TIMSS) is a large-scale international assessment designed to monitor student achievement in mathematics and science across countries worldwide. Its primary objective is to provide valid and comparable estimates of student achievement that can be generalized to clearly defined target populations, such as all Grade 4 students within a country. To achieve this, TIMSS employs probability-based sampling designs that allow for inferential statistics, provided that sampling assumptions are met and participation rates are sufficiently high. TIMSS employs a two-stage sampling design (Martin, Mullis, & von Davier, 2020). In the first stage, a random sample of approximately 150 schools offering the target grade is drawn from a national school sampling frame. In the second stage, one intact class at Grade 4 is randomly selected within each sampled school, and all students in that class are assessed. If all sampled schools and students were to participate, this design would yield unbiased estimates of population parameters. In practice, however, non-participation at the school level poses a serious threat to representativeness and may introduce bias into achievement estimates. To address this risk, TIMSS sampling procedures include the selection of two so-called replacement schools for each initially sampled school. If a sampled school declines participation, the first replacement school is invited, followed—if necessary—by the second replacement school. Replacement schools are selected to be similar to the original school with respect to observable characteristics such as school size, and public or private status. Despite the widespread use of replacement schools, it remains an open empirical question whether this procedure effectively mitigates bias. The fundamental challenge is that the key outcome of interest—achievement—is typically unobserved for non-participating schools. The assumption that similarity on administrative characteristics is sufficient to recover population parameters is rarely tested. This study addresses this gap by leveraging rich Danish register data. We combine TIMSS sampling information with administrative data from the Danish Ministry of Children and Education, including school-level achievement indicators based on grades and detailed measures of schools’ socioeconomic composition and well-being measures. First, we examine differences between participating and non-participating schools. Second, we assess whether replacement schools resemble non-participating schools in terms of achievement and social composition and thus whether they plausibly recover population parameters. By empirically evaluating the effectiveness of replacement schools, this study contributes to methodological debates on survey non-response, sampling validity, and bias mitigation in international large-scale assessments.

Martin, M. O., von Davier, M., & Mullis, I. V. S. (Eds.). (2020). Methods and procedures: TIMSS 2019 technical report. TIMSS & PIRLS International Study Center, Boston College; International Association for the Evaluation of Educational Achievement (IEA). Retrieved from https://timssandpirls.bc.edu/timss2019/methods/pdf/TIMSS-2019-MP-Technical-Report.pdf

In Finland, register data on education is comprehensively collected. The KOSKI database, maintained by the Finnish National Agency of Education, contains information on learners' study rights, study attainments and degrees completed at different levels of education. The organizers of education have an obligation to store information on teaching in the database that includes, among other things, final grades for basic education, grades for matriculation examinations and grade information for upper secondary school leaving certificates. It provides also information on support for learning and schooling. Each learner has their own ID, which can be used to merge data from the KOSKI database to Statistics Finland’s population-level longitudinal register data, which contain extensive information on, for example, the education, socioeconomic status and labor market outcomes of individuals and their parents. The register data provided by Statistics Finland can be accessed using the FIONA remote access system. It is a secure environment that allows for detailed, individual-level analysis of education-related research questions. There is a detailed catalog of available microdata, and researchers can also request merging external datasets with this microdata. Using the FIONA remote access system requires that researchers prepare a user license based on a research plan. In earlier studies, Statistics Finland’s register data have been used, for example, to study students’ academic success, educational choices and progression in further education (e.g., Helin et al., 2023; Helin et al., 2024) or the effect of special education and changes in special education system (e.g., Kirjavainen et al., 2016; Pulkkinen et al., 2020). In Finland, there are also examples of the linking register data on students’ school grades with PISA data to study correspondence between PISA performance and school achievement (Pulkkinen & Rautopuro, 2022). In this presentation, we will focus on the use of national register data and the FIONA remote access system for educational research. We will present how these were used in studying the correspondence of Finnish upper secondary school grades and high stakes Finnish Matriculation examination results. In our research, we used data covering the years between 2022 and 2024. We studied the extent to which subject specific grades given by the teachers correspond to high stakes matriculation examination results in Finnish as a mother tongue, mathematics, English as long foreign language, and history.

Helin, J., Jokinen, J., Koerselman, K., Nokkala, T., & Räikkönen, E. (2023). It runs in the family. Using sibling similarities to uncover the hidden influence of family background in doctoral education and academic careers. Higher Education, 86, 1–20. Helin, J., Koerselman, K., Nokkala, T., & Räikkönen, E. (2024). The socioeconomic background of PhDs across Europe. Työpaperi (vertaisarvioinnissa). Kirjavainen, T., Pulkkinen, J., & Jahnukainen, M. (2016). Special education students in transition to further education: A four-year register-based follow-up study in Finland. Learning and Individual Differences, 45, 33-42. https://doi.org/10.1016/j.lindif.2015.12.001 Pulkkinen, J., Räikkönen, E., Jahnukainen, M., & Pirttimaa, R. (2020). How do educational reforms change the share of students in special education? Trends in special education in Finland. European Educational Research Journal, 19(4), 364-384. https://doi.org/10.1177/1474904119892734 Pulkkinen, J., & Rautopuro, J. (2022). The correspondence between PISA performance and school achievement in Finland. International Journal of Educational Research, 114, 102000. https://doi.org/10.1016/j.ijer.2022.102000

This study examines the methodological and practical challenges of using Education Management Information System (EMIS) data for higher education research within the framework of the StdPerson - Student Persistence and Degree Completion in Higher Education: Theoretical and Empirical Models project. The study draws on large-scale microdata, notably the RAIDES database provided by the Portuguese Directorate-General for Education and Science Statistics (DGEEC), to move beyond traditional descriptive analyses based on aggregated indicators (Ferrão, 2023; Ferrão & Almeida, 2018, 2019a). A central methodological challenge lies in the hierarchical nature of educational data, in which students are nested within courses, departments, and institutions. This structure violates the independence assumption of standard regression models and requires the application of multilevel logistic and random coefficient models to properly capture variability across programs and institutional contexts. Longitudinal data integration constitutes another major challenge. Analysing student persistence and degree completion demands the linkage of pseudonymized enrollment records with graduation files across multiple academic years using unique identifiers. This process is technically demanding and must comply with strict data protection and GDPR requirements (Fazendeiro et al., 2025). Moreover, defining “dropout” is conceptually complex, as researchers must differentiate between students who permanently exit higher education and those who temporarily interrupt their studies, transfer between programs, or move across institutions—each reflecting distinct trajectories and determinants (Ferrão & Almeida, 2019b, 2021). Additional challenges include handling missing data, which often arise from alternative access pathways or incomplete academic records, potentially undermining model robustness. Finally, the StdPerson project underscores the limitations of single-institution analyses and highlights the value of integrating nationwide EMIS microdata to produce findings that are generalizable at the national level. By addressing these methodological constraints, the project establishes a robust empirical framework for identifying long-term predictors of student success—such as prior school failure and first-choice admission—thereby informing institutional policies aimed at promoting equity and the democratization of success in higher education.

Fazendeiro, P., Prata, P., & Ferrão, M. E. (2025). Subtle biases introduced in equity studies through data anonymization. PLoS ONE, October 8,. https://doi.org/https://doi.org/10.1371/journal.pone.0332441 Ferrão, M. E. (2023). Differential effect of university entrance scores on graduates’ performance: the case of degree completion on time in Portugal. Assessment & Evaluation in Higher Education, 48(1), 95–106. https://doi.org/10.1080/02602938.2022.2052799 Ferrão, M. E., & Almeida, L. S. (2018). Multilevel modeling of persistence in higher education. Ensaio: Avaliação e Políticas Públicas Em Educação, 26(100), 664–683. https://doi.org/10.1590/s0104-40362018002601610 Ferrão, M. E., & Almeida, L. S. (2019a). Differential effect of university entrance score on first-year students’ academic performance in Portugal. Assessment & Evaluation in Higher Education, 44(4), 610–622. https://doi.org/10.1080/02602938.2018.1525602 Ferrão, M. E., & Almeida, L. S. (2019b). Student’s access and performance in the Portuguese higher education: Issues of gender, age, socio-cultural background, expectations, and program choice. Avaliação: Revista Da Avaliação Da Educação Superior (Campinas), 24(2), 434–450. https://doi.org/10.1590/s1414-40772019000200006 Ferrão, M. E., & Almeida, L. S. (2021). Persistence and academic expectations in higher education students. Psicothema, 33(4), 587–594. https://doi.org/https://doi.org/10.7334/psicothema2020.68

We describe the education information system in Poland from the perspective of using it for research purposes. The system collects detailed information from school and other education providers, including data on students and teachers. While the main purpose of the system is to monitor the use of resources and collect statistical data, it can be also used to conduct research. We discuss existing and potential research possibilities created by linking the system data to examination results, higher education systems, and social security database. We also discuss how representative surveys can be linked to the system and how the system creates possibilities for conducting surveys and assessments online lowering costs, increasing reliability but opening ways for potential uses of research results.

N/A