Research and Development Fund

Methodological advances to enhance the extraction of policy-relevant information in TIMSS and TIMSS-L

Henry Braun, Damian Betebenner, Kiara Sanchez

TIMSS and TIMSS-L, collect enormous amounts of data. This project will substantially enhance the amount of ILSA policy-relevant information by implementing statistical methods that heretofore haven’t been applied in international contexts. For example, it will demonstrate how TIMSS data can be used to obtain estimates of comparative growth from grade 4 to grade 8 within and between countries.

Outcome Forthcoming

IEABank: Documentation standards and item bank for past and future IEA studies

Andrés Strello, Andrés Christiansen, Maximiliano Romero, Pablo Torres Irribarra, Rolf Strietholt

IEABank creates a centralized item bank template for IEA studies. It will set common documentation standards, improve collaboration across teams, support consistency across studies and cycles, and make IEA data easier to use for trend analyses – both internally and for external researchers.

Outcome Forthcoming

Beyond the Score: A multi-level and cross-grade analysis of process data in TIMSS 2023 for equity, instruction, and item design

Philipp Doebler, Purya Baghaei Moghadam, Hamdollah Ravand, Susanne Frick

Beyond the achievement scores in TIMSS 2023, what does the rich log-data reveal about test-taking behavior? Log-data comprises timings, click events, navigation paths, and action sequences, an unprecedented objective record of the micro-level behaviors students exhibit while solving problems. Specifically, the project studies persistence, rapid guessing, and strategic reviewing.

Outcome Forthcoming

ILSAreport: Accessible tables using IEA data with R

Andrés Christiansen, Andrés Strello, Katie Zuber

The ILSAreport project aims to create an R package that will streamline the production of tables for IEA international reports, ensuring their compliance with accessibility guidelines.

Outcome Forthcoming

Bridging Time and Context: Longitudinal comparability of background scales in international assessments (COMPASS)

Diego Carrasco, Andrés Sandoval-Hernandez, Nurullah Eryilmaz

COMPASS examines longitudinal and cross-national comparability of ICCS background scales using ordinal CFA and alignment methods. The initiative will generate comparable scale scores when feasible, develop methodological tools for invariance diagnosis, and aid trend analyses in international large-scale assessments.

Outcome Forthcoming

Modeling problem-solving trajectories in ICILS 2023 using multi-state survival analysis: A study with computational thinking (CT) assessment data

Jennifer Viñas Forcade, Mariana González Burgstaller, Valentina Cancela, Déborah Zak Godoy, Gimena Castelao, Leonardo Moreno

This project examines how students solve computational thinking tasks in ICILS 2023 by modeling problem-solving states and transitions. Using multi-state survival models, it examines transitions between incorrect, partial, and correct states, response times and strategies, comparing students and countries to provide evidence for teaching and to improve computer-based assessments design.

Outcome Forthcoming

Real-time dashboard for monitoring data collection procedures and outcomes in international large-scale assessments

Sebastian Meyer, Diego Cortes, Marius Kaltenbach, Ali Raza

This project develops a reusable real-time dashboard to monitor data collection activities in large-scale assessments. By providing timely indicators, visual summaries, and early warnings, it enables proactive interventions to address emerging risks, supports efficient study management, and strengthens transparency and collaboration between IEA and participating countries during ongoing fieldwork.

Outcome Forthcoming

Exploring the Application of AI/Machine-Learning Based Options for IEA Translation and Translation Verification Procedures

Lauren Musu, Miguel Zea, David Ebbs, Anja Waschk, Aisha Elsayed

The project explores integrating AI-assisted translation into IEA workflows by developing an IEA-specific translation and quality estimation model. Using verified past translations, its performance will be compared with other translation outputs through automated metrics and blind human reviews to assess feasibility, quality, and scalability.

Outcome Forthcoming

Linking grade 4 and grade 8: structural differences and comparability of TIMSS student context indices via MGCFA and network analysis

Yulia Kuzmina

Student context indices such as interest or confidence in mathematics are widely used in TIMSS analyses, yet their comparability across grades remains understudied. This project investigates similarities and differences between Grade 4 and Grade 8 contextual indices to support valid cross-grade comparisons of students’ motivation and attitudes toward school.

Outcome Forthcoming