Multilevel Analysis

Posted on

Multilevel Analysis

  • Instructor(s): TBA TBA
  • Available also as advanced: Yes

Many datasets a kind of natural hierarchy exists. For instance, repeated measurements are in fact measurements within subjects, or, to give another example, students are nested within classes which themselves are nested within schools. This nesting of observations have been an enormous problem for data analysis; in the first example both the variance within subjects (between measurements) and between subjects have to be estimated simultaneously. In the second example the differences within classes (between students) and between classes (within schools) as well as the variance between schools have to be estimated. Failing to do so always results in an underestimation of the total variance, and therefore the null hypothesis will be rejected too easily.

Today the magic words are ‘multilevel analysis’. Multilevel analysis can be carried out in SPSS. In this hands-on tutorial the principles of multilevel analysis will be dealt with and participants should be able to carry out (simple) multilevel analysis afterwards. However, the tutorial is only open to participants who have some knowledge of statistical testing (i.e. they should know concepts like variance, null hypothesis and testing statistics (t, F, χ2)).

Return to the list of tutorials

Sidebar