11-08 Multilevel Analysis with SPSS

Researchers often face situations where data are generated by different observational units. A classical example from educational research involves the investigation of the impact of school characteristics on the performance of the students. In this example, students are nested within schools. This means that we cannot consider students as independent observations as students from the same school are more aloke than students from different schools.

In another field of study, the same subjects are measured on several occasions. Hence, observations are nested within subjects. And, of course, two observations of the same individual are more alike than two observations from two different individuals. Multilevel analysis takes these dependencies into account. That is, statistical models are formulated in order to estimate both the differences between schools and within schools (first example), or between and within individuals (second example). Failure to do results in a too easy rejection of the null-hypothesis . Furtheremore,  Multilevel analysis, or more generally mixed modelling, provides a general framework for both the analysis and the conceptualisation of data from a hierarchical sample.

In this course you will learn to perform multilevel analyses with SPSS. We will start from a very basic multilevel model and elaborate this model to accommodate research designs with longitudinal data, multiple dependent variables and categorical dependent variables. For this course we will use the software package SPSS. In the course theory and hands-on exercises are mixed-up in such a manner that theoretical explanations can be applied directly in a realistic research situation. On the final day we promote attendees to work with their own datasets.

If you do not have data yet, we will provide you with a data set which more or less resembles the data you are gathering.

Requirements

A basic knowledge of SPSS as well as a basic statistical knowledge is a must. That is, students should be familiar with SPSS interface and can use statistical concepts such as mean, variance or testing statistics (Chi2, student-t, z-score). Some knowledge of regression analysis and/or analysis of variance is a necessity.

Lecturers

Meetings

Students should bring their own laptop with SPSS 16.0 (or higher) installed. The course will take three full days (10.00 - 17.00).

Assessment

There wil be no formal assessment. We assume that active participation willl lead to understanding of the statistical procedures and the logic behind multilevel modeling.

Literature

Some relevant literature (copies of parts will be provided):

  • Heck, R.H, Thomas, S.L. & Tabata, L.N. (2010). Multilevel and Longitudinal Modeling with IBM SPSS (Quantitative Methodology Series) Routledge.
  • Snijders, T., & Bosker, R. (1999). Multilevel analysis: An introduction to basic and advanced multilevel modeling. London: Sage.
  • Quené, H. & Bergh, H. van den (2004). On multi-level modelling of data from repeated measures designs: A Tutorial. Speech Communication, 43, 103-121.

Dates and location

November 16, 17 and 18, 2011
Zalencentrum Vredenburg, Utrecht

Registration

PhD students can register for the course by completing the registration form. A maximum of 24 participants will be admitted to the master class.
Deadline for application: October 5, 2011.