The course is the union of two 3 ECTS courses: Statistical Data Analysis II and Statistical methods for policy evaluation and causal inference. Students will attend these to courses in sequence for a total of 3 ECTS.
The aim of the course 'Statistical models for program evaluation' is to introduce students who have already studied the basic concepts of statistical inference and of the linear regression model to the analysis of categorical data, with a special focus on generalized linear models, and to inferential statistical methods for program evaluation.
The main means of assessment will be a written test with exercises. The not compulsory oral test will be used to better specify the mark of the written test.
After the course students will be able to apply the learned statistical tools of inference to analyse data and to interpret the output of the analysis.
The written test will contain numerical exercises as well as theoretical questions.
Students will learn how to analyse data with the appropriate statistical method.
During the lessons some exercises similar to those of the final written exam will be solved, so that it will be possible to evaluate the acquired skills.
Basic knowledge of statistical inference (probability, random variables, confidence intervals, hypothesis testing for one population).
None.
During the lessons a pen tablet (digital blackboard) will be used. After each lesson the professor's notes will be made available to the students in pdf format on the Moodle of the course.
Other teaching material (e.g. exercises with solution) will be available on the Moodle of the course.
The course will first introduce the distributions and inference for categorical data and for contingency tables. Then, the course will introduce generalized linear models, with a special focus on logistic regression.
This course is an introduction to the inferential statistical methods for program evaluation. The statistical concepts are illustrated using data and real examples, focusing on the methods used for causal inference in public policy contexts. 
Main topics: Randomized Trials. Matching. Regression. Instrumental Variables. Regression Discontinuity Designs, Differences-in-Differences.
During the course exercises and case studies will be solved using the R and Stata software.
Agresti A. (2002) Categorical Data Analysis - Second Edition
McCullagh P., Nelder J.A. (1989) Generalized Linear Models – Second Edition, Chapman & Hall
The syllabus and assessment method of the course also apply to non-attending students.
The written test will be formed by three/four exercises.
Exercises may consist in questions to be solve by applying formulas, questions on the theory of the methods and questions on the interpretation of the output if the methods.
None.
https://moodle.ec.unipi.it/course/view.php?id=497