Random Item Effects Modeling for Cross-National Survey Data

Abstract

Item response theory (IRT) methods are standard tools for the analysis of large-scale assessments of student’s performance. In educational survey research, the National Assessment of Educational Progress (NAEP) is primarily focused on scaling the performances of a sample of students in a subject area (e.g., mathematics, reading, science) on a single common scale, and measuring change in educational performance over time. The Organization for Economic Cooperation and Development (OECD) organizes the Program for International Student Assessment (PISA). The program is focused on measuring and comparing abilities in reading, mathematics, and science of 15-year-old pupils over 30 membercountries and various partner countries every three years and started in 2000. Another example is the Trends in International Mathematics and Science Study (TIMSS) conducted by the International Association for the Evaluation of Educational Achievement (IEA) to measure trends in students’ mathematics and science performance. Large-scale (educational) survey studies can be characterized by: (1) the ordinal character of the observations, (2) the complex sampling designs with individuals responding to different sets (booklets) of questions,