The targeted audience includes government officials with a technical profile who work in social areas (e.g. ministries of labor, economics, social development, statistical institutes), social analysts in the region, and advanced students and graduates in Economics and other social sciences. Prior experience or knowledge of the Stata software is not required.

The content of the course is developed over 5 days, with 6 daily lesson hours (3 in the morning and 3 in the afternoon). Each day includes a theoretical presentation for each topic followed by the discussion and implementation of empirical examples. In addition, an optative intensive one-day training on the basic use of Stata software is offered – at no additional cost.

• Unit 1: Why evaluate? Evidence-based policy making What is impact assessment? How to formulate evaluation questions?
• Unit 2: How to evaluate? Causal Inference and Counterfactual Construction
• Unit 3: Random selection methods. Random allocation of treatment. Variations of random assignment
• Unit 4: Non-experimental Methods (I) Regression Discontinuity Design. Limitations. Example and Application
• Unit 5: Non-experimental methods (II). Difference in differences. Limitations. Example and Application
• Unit 6: Non-experimental methods (III). Matching estimators . Limitations. Example and Application
• Unit 7: How to implement an evaluation. Ethical considerations. How to choose an impact evaluation method. Sampling, sample size and data collection

The course is led by María Laura Alzúa, Leonardo Gasparini y Martín Cicowiez. Moreover, during the practical sessions there is one instructor every five students.

2016, April 18-22: In the context of the project “Enhancing Women’s Economic Empowerment through Better Policies in Latin America”, the Center for Distributive, Labor and Social Studies, CEDLAS , in collaboration with CIEDUR, and with the support of the International Development Research Centre (IDRC-Canada)

2015, April 20-24: In the context of the project “Enhancing Women’s Economic Empowerment through Better Policies in Latin America”, the Center for Distributive, Labor and Social Studies, CEDLAS , in collaboration with CIEDUR, and with the support of the International Development Research Centre (IDRC-Canada)

2014, September 1-5: In the context of the project “Enhancing Women’s Economic Empowerment through Better Policies in Latin America”, the Center for Distributive, Labor and Social Studies, CEDLAS , in collaboration with CIEDUR, and with the support of the International Development Research Centre (IDRC-Canada)

2014, April 21-25: in collaboration with GRADE and with the support of Iniciativa Think Tank (ITT)

2013, August 26-30:  in collaboration with GRADE and with the support of Iniciativa Think Tank (ITT)

The targeted audience includes government officials with a technical profile who work in social areas (e.g. ministries of labor, economics, social development, statistical institutes), social analysts in the region, and advanced students and graduates in Economics and other social sciences.

The development of the course assumes a level of basic knowledge, corresponding to the following concepts: the fundamental problem of evaluation; Causal inference and counterfactual construction; Random selection methods and non-experimental methods: difference in differences and matching estimators. For more detail, see the basic course program in the previous section. See Gertler, Paul J., Sebastian Martinez, Patrick Premand, Laura B. Rawlings, and Christel MJ Vermeersch. Impact evaluation in practice. World Bank Publications, 2011 as reference material.

The content of the course is developed over 5 days, with 6 daily lesson hours (3 in the morning and 3 in the afternoon). Each day includes a theoretical presentation for each topic followed by the discussion and implementation of empirical examples.

A.    The  fundamental problem of Evaluation

a.    Potencial results and selection bias

B.    Experimental Methods

a.    Randomization methods and levels.

b.    Power calculations

c.    Multiple treatments

d.    Balance

e.    Multiple testing

f.    Sampling and surveys

C.    Non experimental Methods: regression discontinuity design

a.   Background

b.   The evaluation problem and RD

i.    The problem
ii.    Sharp RD and Fuzzy RD
iii.    Identification
iv.    Treatment effects in RD and interpretation
v.    External and internal validity

c.    Implementation

i.  Graphic analysis
ii.   Regression model selection
iii.  Test of validity for the RD model
iv.  Robustness

d.    Empirical examples

e.    Extentions to the basic model

i.    Longitudinal data in a RD approach
ii.   Multiple allocation variables

D.    Non experimental methods: differences in differences and event study

a.   Difference in differences:

i.   The evaluation problem and DD
ii.    Identification
iii.   Inference
iv.   External and internal validity

b.    Event Studies:

i.   Graphic analysis
ii.   Test of validity
iii.    Robustness

E.    Mechanisms and multiple treatments experiments

a.   Mechanisms and multiple treatments experiments

i.   The evaluation problem: from impacts to mechanisms
ii.    Identification in multiple treatments: beyond treatment and control
iii.    Experiments and structural models

The course is led by María Laura Alzúa, Guillermo Cruces, Santiago Garganta, and Marcelo Bergolo. Moreover, during the practical sessions there is one instructor every five students.

2016, October 17-21: In the context of the project “Fostering capacities in Impact Evaluation in Latin America” with the support of the International Development Research Centre (IDRC-Canada)

2015, October 19-23: In the context of the project “Fostering capacities in Impact Evaluation in Latin America” with the support of the International Development Research Centre (IDRC-Canada)

2014, October 20-24: In the context of the project “Fostering capacities in Impact Evaluation in Latin America” with the support of the International Development Research Centre (IDRC-Canada)