Advanced data analysis in public policy evaluation

[4 half days] - [English]

Would you like to be able to assess the causal effect of a public policy on a variable of interest?

Treatment-effect estimators quantify the causal effect of a treatment/policy on an outcome, based on observational or quasi-experimental data. For example, a treatment could be a new drug and the outcome blood pressure or it could be a job training program with employment or wages as outcome, etc.

Causal inference requires the estimation of the outcomes for each treatment level. But one only observes the outcome of each subject (individual, firm, region…) conditional on the received treatment. Experiments may help but can be expensive and sometimes unethical. Fortunately, many things can be done with observational data. But one needs some statistical machinery.

Training aims
Equip the participants with the most commonly used methods available in Stata to estimate treatment effects from observational data.

Prerequisites

By registering for this training, you commit to a level of knowledge equivalent to a course of intermediate econometrics or biostatistics and to the following training(s):

Content

Regression Adjustment (RA) Inverse probability Weighting without (IPW) and with Regression Adjustment (IPWRA)
Propensity score matching (PSM)
- Matching over a single (propensity) score
- Matching algorithms (nearest neighbourhood, caliper, Kernel)
Difference-in-Difference (DiD) estimation:
- Treatment effects estimation using (repeated) cross-sectional data
- The canonical DiD estimator
- Assessing the common trend assumption
- Relaxing the common trend assumption: from DiD to DiDiD
- Combining DiD and inverse probability weighting

Methods and method families discussed

Please log in to register as interested in this training.

It is necessary to create an account for a first use of our services.


	UCLouvain > SMCS Platform EN \| FR LOGIN NEWSLETTER