Computational typologies of multidimensional end-of-primary-school performance profiles from an educational perspective of large-scale TIMSS and PIRLS surveys

Oct 15, 2026

Computational methods are used to derive and compare multidimensional performance profiles for pupils at the end of primary school, drawing on large-scale TIMSS and PIRLS survey data. The study applies multidimensional item response theory and latent profile analysis to classify children into reading, mathematics and science performance types. Clustering solutions from classic model-based and model-free approaches are benchmarked for consistency, and results are further contrasted with a qualitative competency-level typology grounded in TIMSS guidelines, including an evaluation of goodness-of-fit.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages20

Reading Time50 min

AccessFree to download

Introduction: TIMSS and PIRLS Studies

Performance scales and international benchmarks

Competency levels and interpretation

Computational typologies of multidimensional end-of-primary-school performance profiles from an educational perspective of large-scale TIMSS and PIRLS surveys Ali Ünlü1 Technical University of Munich (TUM) Centre for International Student Assessment (ZIB) TUM School of Education Michael Schurig IFS, TU Dortmund University Recently, performance profiles in reading, mathematics and science were created using the data collectively available in the Trends in International Mathematics and Science Study (TIMSS) and the Progress in International Reading Literacy Study (PIRLS) 2011. In addition, a classification of children to the end of their primary school years was conducted in accordance with these performance types. To create performance profiles and classifications, multidimensional item response theory and latent profile analysis were used. The focus in this study is on the comparison and usability of clustering methods in their application in large-scale assessments. In a first step, the cluster solutions of classic approaches such as kmeans, fuzzy c-means and hierarchical procedures are compared to one another and assessed in terms of their proximity to the reference typology of latent profile analysis. In the second step, the results of the model-supported latent profile analysis are compared directly with the findings of the classic model-free cluster analyses by means of appropriate measured values. The result is a high consistency in the classification of invariant ranked profiles. In the last step, the calculated“quantitative” cluster solutions are compared to the “qualitative” typology of the pupils derived from the content-based benchmarking of the competency levels in mathematics using the TIMSS guidelines. It is evident that, as a cluster solution, the benchmarking breakdown of the sample in the five competency levels does not show a high goodness offit with the available data. Introduction: TIMSS and PIRLS Studies Every five years, fourth grade students from around the world are compared on their performance in reading comprehension through the Progress in International Reading Literacy Study (PIRLS; e.g., Bos, Tarelli, Bremerich-Vos, & Schwippert, 2012a) . Another international test that evaluates student academic performance is the Trends in International Mathematics and Science Study (TIMSS; e.g., Bos, Wendt, Köller, & Selter, 2012b), which assesses fourth and eighth grade students on their performance in mathematics and science on a four-year cycle. The International Association for the Evaluation of Educational Achievement (IEA) is responsible for conducting both studies ([www.iea.nl](www.iea.nl)) . Data collection for both studies coincided for the first time in 2011, making it possible to connect the two studies. Thus, the focused subject areas of reading, mathematics and science could be mutually investigated while considering their interactions. As was the 1 Corresponding author. Prof. Dr. Ali Ünlü ([ali.uenlue@tum.de](ali.uenlue@tum.de)), Technical University of Munich (TUM), Arcisstr. 21, 80333 Munich, Germany. © 2015, Current Issues in Comparative Education, Teachers College, Columbia University. ALL RIGHTS RESERVED. Current Issues in Comparative Education 18(1), 6-25. Computational typologies of multidimensional end-of-primary-school performance profiles from an educational perspective of large-scale TIMSS and PIRLS surveys case in 37 other countries and regions, a common random sample of pupils for both PIRLS and TIMSS was surveyed in Germany (Martin & Mullis, 2013) . This special data set opens up new possibilities for analysis and allows for an overarching view of pupils’performances at the end of the fourth grade in reading, mathematics and science (e.g., Bos et al., 2012c) . From a methodical viewpoint, these data have particular demands in terms of the appropriate handling of the multidimensional structure (Foy & O’Dywer, 2013) . Multidimensional procedures have thu

FAQs

What data sources are used to build the performance profiles?

The analysis uses data from the TIMSS and PIRLS 2011 large-scale assessments, focusing on reading, mathematics, and science performance at the end of fourth grade.

Which methods are used to create the performance profiles and classifications?

Multidimensional item response theory and latent profile analysis are used, complemented by comparisons to classic clustering methods such as k-means, fuzzy c-means, and hierarchical procedures.

How are competency levels defined in the study?

TIMSS and PIRLS benchmark scales are divided into five performance ranges; in Germany these are reported as competency levels I to V using specified score intervals.

Oct 15, 2026

Upload by PakDamar76