Course Syllabus
Obiettivi
Introduction to the construction and interpretation of Directed Acyclic Graphs (DAGs) Causal Diagrams as tools for:
- represent assumptions about the relationships between variables in observational and randomized studies
- obtain from this representation a guide in pragmatic statistical analysis
- thanks to this representation, avoid common errors and biases in statistical analysis
Contenuti sintetici
- Causality and Correlation between variables.
- Graphical representation of causality between variables.
- Simpson's paradox
- Causal path between variables and characteristics: open, closed, backdoor, D-separation.
- Characteristics of variables within a causal path: confounder, mediator, common cause, other.
- Impact of study design on the possible distortion of causal relationships: the observational case-control study and eligibility criteria in randomized studies.
The concepts introduced from a theoretical point of view will always be accompanied by an example on simulated data analyzed with the STATA software.
The student will be asked to reproduce/integrate some calculations to be discussed together using a software chosen from: R, SPSS, STATA, SAS.
Prerequisiti
- Basic elements of probability calculation
- Construction of frequency tables and regression models with a statistical software chosen from: R, Stata, SAS, SPSS
Modalità didattica
Standard class
Materiale didattico
Slides
Periodo di erogazione dell'insegnamento
3 April 11-13 - Kythos Building in Monza
7 April 11-13 - Kythos Building in Monza
10 April 11-13 - Kythos Building in Monza
14 April 11-13 - Kythos Building in Monza
Aims
Introduction to the construction and interpretation of Directed Acyclic Graphs (DAGs) Causal Diagrams as tools for:
- represent assumptions about the relationships between variables in observational and randomized studies
- obtain from this representation a guide in pragmatic statistical analysis
- thanks to this representation, avoid common errors and biases in statistical analysis
Contents
- Causality and Correlation between variables.
- Graphical representation of causality between variables.
- Simpson's paradox
- Causal path between variables and characteristics: open, closed, backdoor, D-separation.
- Characteristics of variables within a causal path: confounder, mediator, common cause, other.
- Impact of study design on the possible distortion of causal relationships: the observational case-control study and eligibility criteria in randomized studies.
The concepts introduced from a theoretical point of view will always be accompanied by an example on simulated data analyzed with the STATA software.
The student will be asked to reproduce/integrate some calculations to be discussed together using a software chosen from: R, SPSS, STATA, SAS.
Prerequisites
- Basic elements of probability calculation
- Construction of frequency tables and regression models with a statistical software chosen from: R, Stata, SAS, SPSS
Teaching form
Standard class
Semester
3 April 11-13 - Kythos Building in Monza
7 April 11-13 - Kythos Building in Monza
10 April 11-13 - Kythos Building in Monza
14 April 11-13 - Kythos Building in Monza
Assessment method
Quiz