Many of the questions that matter most in educational and social sciences involve understanding relationships between multiple factors. For example, whether a student's social background predicts their performance at university, which factors predict progress at school or completion of vocational education, whether teachers respond differently to students from different backgrounds and whether this in turn shapes how those students do, in each case several factors act together, and each of them has an effect and relationship with the others. Multivariate analysis refers to the techniques that allow several variables to be examined simultaneously, so that the pattern among them becomes visible and the effect of one is not mistaken for the effect of another.
This is especially relevant in educational science, because education is inherently complex. Students, teachers, classrooms, schools, workplaces and communities all contribute multiple interacting variables. Examining these one at a time often conceals the pattern that gives the better account of what is happening, and it can suggest relationships that disappear once something else is taken into account.
This is an applied course. The course focuses on the logic of choosing among multivariate statistical techniques, which includes recognising what kind of problem you are facing, selecting an analysis appropriate to that problem and to the data available, understanding the assumptions a technique makes understanding the consequences where assumptions are not met, and interpreting output in plain language. The aim is to produce researchers who can identify and use multivariate analyses in their own work, whether in professional practice or as preparation for their own research. Analyses are carried out in SPSS and used throughout on real datasets.
- Docente: Tobia Fattore