Center for Comparative Analysis of Educational Achievement, COMPEAT

COMPEAT

Center for Comparative Analyses of Educational Achievement

COMPEAT is an infrastructure project with the general aim to build databases of international large-scale studies in educational achievement conducted by IEA and OECD before the year 2000 and to support secondary analyses of these data.

During the last 50 years, a large number of studies of student achievement in different school subjects have been conducted by the IEA (the International Association for the Evaluation of Educational Achievement) and the OECD (the Organisation for Economic Co-operation and Development). And the number of participating countries continues to increase. The results from these studies have a great impact on educational policy and practice. Data collected in these studies offered a valuable resource for different research purposes. However, there are certain difficulties in these data that make data management and analyses more complicated and technically demanding. 

• Complex and stratified cluster sampling design needs advanced statistic modeling techniques to take care of the non-independency in the data;
• Achievement measures from matrix designs, e.g. students have only partly been administered the same test tasks, needs missing data techniques to take care of the non-responses;
• Measurement errors in the observed variables need to be handled by latent variable analyses;
• Variables observed at a low scale level, usually categorical or ordinal level, need rescaling techniques such as IRT and factor score approach.
• Data from the studies before 1990 are more difficult to access, and were processed in a different manner compared to data from studies conducted later;
• Country-level information is limited;
• Causal inferences are difficult to draw with cross-sectional data.

The potential of these databases has therefore not fully been taken charge of. The purpose of COMPEAT is to create better conditions for secondary analysis by the means of:

• Building databases
  • Make data and documentation from studies before 1990 available in updated formats
  • Add information to the databases with variables at the national level
• Prepare data for analysis
  • Create aggregated (country level) and disaggregated (individual level) databases
  • Create new variables from combinations of several variables
• Build analytical competences
  • Courses
  • Workshops
  • Seminars