The application deadline has passed, applying is no longer possible
The course introduces multilevel regression analysis for researchers featuring models for nested or hierarchical social science data. Models and examples are discussed in as non-technical a way as possible; the emphasis is on understanding the methodological and statistical issues involved in application of the models. Every course day starts with a lecture followed by a computer exercise in which you complete an assignment.
During the computer exercises various aspects of multilevel modeling will be trained using the R software. In completing the assignments you will work with cross-national and longitudinal datasets.
The day-to-day content includes, on day 1, discussion of the features of multilevel data and their consequences for statistical analysis, including number of effective observations, intra-class correlation, null and random intercept model, and the R software.
Topics covered on day 2 are fixed effects, random slopes and significance testing, and you discuss level-1 (X) and level-2 (Z) predictors variables, types of regression effects (fixed, random), and null hypothesis tests.
On day 3 the issues covered include cross-level interaction and proportion explained variance. Here you will discuss interaction of X and Z variables, R2 measures, and within and between regression.
On day 4 the attention shifts to longitudinal data. The topics discussed are wide vs long data files, fixed and random parts of multilevel longitudinal models, time-constant and time-varying predictor variables, within and between effects, and fixed effects models.
On day 5 the multilevel longitudinal model is juxtaposed with Generalized Least Squares followed by a discussion of multilevel logistic regression.
