Department’s Research Grants
2024 – 2025 Using Large Language Models for Automatic Item Generation: Development and Validation for TIMSS Fourth Grade.
With this grant, we will explore the possibilities of LLMs for automatically generating items for cognitive assessments. This research aims to automatically generate fourth grade TIMSS items using large language models. The items will be validated through subject matter evaluations and field tests simulating TIMSS settings. The generated items will be compared to human made TIMSS items with respect to their measurement properties (IEA Research and Development Fund, 2023).
2024 – 2025 Approximate Areas of Interest for Enhanced Understanding of Student Motivation and Task Interaction in IEA Assessments.
We will use deep learning models to investigate how cursor movements predict test-takers’ motivation. The approach will also differentiate high-performing students from their peers, find flawed assessment items, and provide insights into improving educational assessment practices (IEA Research and Development Fund, 2023).
2021 – 2025 Research Lab for the Digital Social Sciences. About the use of digital footprints in studying social problems.
This project aims to advance methods to use joint digital footprints and survey to solve social problems (i.e., trust, attitudes towards migration, environment, and governments, misinformation, and polarization of attitudes). To do that, we propose a large methodologically oriented study that will utilize five sources of data: representative survey data, online trace data, data from a panel study (mixing survey and online trace data), experimental studies, and qualitative data. (The National Science Centre grant “Sonata Bis” 2020/38/E/HS6/00302).
2020 – 2024 Understanding response styles in self-report data: consequences, remedies and sources.
Self-report data, essential for research, politics, and commerce, face issues like non-comparability and biased responses due to different response styles and systematic tendencies to answer beyond the item content. Despite ample research, the effects, control methods, and reasons for response styles remain unclear, raising questions about their impact on data quality and individual and contextual causes. This project explores these questions using simulation studies, international database analyses, machine learning, and experimental studies to understand predictors like personality, motivation, and cognitive abilities (The National Science Centre grant “Opus”, 2019/17/HS6/00937)
Other Research and Activities
2022 – 2025 CSS is collaborating with The Centre for Advanced Study (CAS) at The Norwegian Academy of Science on a research grant: Disadvantaged Students Who Beat the Odds. Toward a New Generation of Research in Academic Resilience (#BeResilient project)
Ongoing The goal of the department is to develop and provide methods and tools for the social sciences to analyze digital behavioural data (produced in interactions with computers, phones and other devices) and to combine digital with traditional types of survey data to improve the analysis of a wide range of sociocultural phenomena. The department addresses these tasks using machine learning, natural language processing, network analysis, psychometrics and statistical modelling, combining various research and statistical techniques with social science expertise. The department engages in purely sociological research and multidisciplinary projects, applying the developed methods to various data types. It pays particular attention to the implementation nature of research, which can positively influence the socio-economic development of our country.
Recent Publications
Pokropek, A., Żółtak, T., & Muszyński, M. (2026). Do You Agree? Do You Strongly Agree? The Effect of the Number of Response Categories on Response Processes and Verification of Substantive Hypotheses. International Journal of Public Opinion Research 38(1), edaf070. https://doi.org/10.1093/ijpor/edaf070
Pokropek, A. (2026). Detecting careless response patterns in Likert scales using supervised machine learning and Monte Carlo simulations. Current Psychol 45, 518. https://doi.org/10.1007/s12144-026-09186-3
Żółtak, T., Pokropek, A., & Muszyński, M. (2026). Response Styles and Omitted Variable Bias. Detection and Mitigation. Structural Equation Modeling: A Multidisciplinary Journal, 33(1), 9–18. https://doi.org/10.1080/10705511.2025.2555203
Muszyński, M., Banasik-Jemielniak, N., Żółtak, T. et al. (2026). Moving intelligence measurement online: adaptation and validation of the Polish version of the Pathfinder general cognitive ability test. Quality & Quantity 60, 329–354. https://doi.org/10.1007/s11135-025-02254-z
Pokropek, A., Żółtak, T., Muszyński, M. (2024). Identifying Careless Responding in Web-Based Surveys. Zeitschrift für Psychologie 232(2), 95-108. https://doi.org/10.1027/2151-2604/a000555
Cachia, R., Pokropek, A., & Giannoutsou, N. (2024). Supporting the monitoring of the digital capacity of schools through optimal shortening of the SELFIE tool. Computers & Education, 208, 104938. https://doi.org/10.1016/j.compedu.2023.104938
Rogoza, R., Krammer, G., Jauk, E., Flakus, M., Baran, L., Di Sarno, M., … & Fatfouta, R. (2024). The peaks and valleys of narcissism: The factor structure of narcissistic states and their relations to trait measures. Psychological Assessment, 36(2), 147. https://doi.org/10.1037/pas0001295
Pokropek, A., Żółtak, T., & Muszyński, M. (2023). Mouse chase: Detecting careless and unmotivated responders using cursor movements in web-based surveys. European Journal of Psychological Assessment, 39(4), 299–306. https://doi.org/10.1027/1015-5759/a000758
Meuleman, B., Żółtak, T., Pokropek, A., Davidov, E., Muthén, B., Oberski, D. L., … & Schmidt, P. (2023). Why measurement invariance is important in comparative research. A response to Welzel et al.(2021). Sociological methods & research, 52(3), 1401-1419. https://doi.org/10.1177/00491241221091755
Zając, T. Z., Żółtak, T., Bożykowski, M., & Jasiński, M. (2023). All that glitters is not gold—Mixed early labour market outcomes of STEM graduates in Poland. European Journal of Education, 58(3), 477-497. https://doi.org/10.1111/ejed.12564
Pokropek, A., Khorramdel, L., & von Davier, M. (2023). Detecting and Differentiating Extreme and Midpoint Response Styles in Rating Scales using Tree-Based Item Response Models: Simulation Study and Empirical Evidence. Psychological Test and Assessment Modeling, 65(2), 220-258.
Kajdy, A., Sys, D., Pokropek, … Mind-COVID Collaborative Team (2023). Risk factors for anxiety and depression among pregnant women during the COVID-19 pandemic: Results of a web-based multinational cross-sectional study. International journal of gynaecology and obstetrics: the official organ of the International Federation of Gynaecology and Obstetrics, 160(1), 167–186. https://doi.org/10.1002/ijgo.14388
Muszyński, M., Pokropek, A., Castaño-Muñoz, J., & Vuorikari, R. (2023). Can Overclaiming Technique Improve Self-Assessment Tools for Digital Competence? The Case of DigCompSat. Social Science Computer Review, 41 (6), 2318-2341. https://doi.org/10.1177/08944393221117269