Semester 2, 2021-2022Type of course
Methodological and Practical CoursesDate
April 4, April 5 and April 6, 2022Location
3 daysMaximum number of participants
1 EC will be appointed for participation in the complete courseStaff
Marcello Gallucci (University of Milano-Bicocca)
Seminar 1: A Moderated View of Regression
In this class we review the general linear model as applied to various research designs and problems, with particular emphasis on mediation and moderation and their combination. We discuss a general approach to obtain effect sizes and hypothesis testing for regression, ANOVA and ANCOVA, path analysis, and models with continuous and nominal mediator and moderator variables.
Seminar 2: A mix within
In this class we review theory and applications of the general mixed model, with fixed and random effects, applied to different research designs. Student will get familiar with the classical toolbox associated with the regression/ANOVA approach, but generalized to mixed models applied to clustered data.
Seminar 3: Practice makes perfect
Seminar on how to select, estimate and interpret GLM (regression and ANOVA), mixed models, and hierarchical linear models with the help of popular statistical software, such as SPSS, SAS or R (depending on students' preference). Many practical aspects of the mentioned models are discussed with the help of practical examples. The aim of the seminar is to illustrate some of the major theoretical issue of important statistical models from a very practical point of view. Students will work with the help of the teacher on datasets provided by the teacher in order to concretely solve some statistical challenges, and acquire confidence in every-day data analysis.
Seminar 4: Count on me
In this class we review theory and applications of the generalized linear model, useful to analyze data when the dependent variable cannot follow a normal distribution, such as nominal, categorical, counts, percentages, frequencies and response time kind of variables. Techniques such as logistic regression, Poison regression, loglinear models are also discussed.
Seminar 5: Let your data be seen!
Seminar on how to select, estimates and interpret results useful for real research problems provided by the students. Students will work with the help of the teacher on their own dataset (or colleagues' datasets). The aim of the seminar is to illustrate how to approach and analyze data from very up-to-date and interesting research designs. Within a certain range, all major topics may be discussed depending on the statistical problems raised by data provided by the students.