Objective

The first objective of this workshop is to provide knowledge of the theory and applications of multilevel linear modeling (MLM), focusing especially on those features that are particular to large-scale assessment data (such as complex data structures, weighting and using Plausible Values). The second objective is to gain basic practical experience on the applications of two-level models.

Procedures

The first part of the workshop consists of a methodological introduction to MLM and its underlying assumptions. In the second part, a hands-on training offers opportunity to practice multilevel analysis with MPlus (Muthén & Muthén, 2012). IEA data is used for fitting and interpreting results of relevant models. Further, methodological concepts related to complex study and sampling design in large-scale assessments are presented and recommendations on suitable implementation of multilevel models are provided.

Expected outcomes for participants

After the workshop participants will be able:

• Understand the theoretical principles and assumptions associated to MLM;
• Specify MLM models considering the complex design of IEA studies;
• Conduct two-level model analyses using the MPlus software;
• Interpret and present results of analyses software with focus on policy research.

Target audience and requirements

The workshop builds on prior knowledge of inferential statistics (such as regression, correlation, and variance analysis). Familiarity with SPSS or other statistical software as well as with large scale data is expected. MPlus knowledge is not required. Familiarity with syntax based analysis is an advantage. Participants have to bring their own computers with a demo version of MPlus installed.

Recommended reading

Raudenbush, S. W. & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods. Newbury Park: Sage Publications.

Hox, J. J. (2008). Multilevel analysis. Technics and applications. New York: Psychology Press.

.