Monday, 20 April 2015: Getting Started

Morning: Methods presentations on causal inference

  • Research questions: what are X and Y
  • The potential outcomes framework and the fundamental problem of causal inference

Lunch time presentation: Macartan Humphreys: Experimental evidence on housing discrimination in New York and what can be done about it

Afternoon: Design Clinic

  • The stages of research design and implementation
  • Introducing the research design form
  • Clinic: Make sure you are set up with R


Tuesday, 21 April 2015: Identification and Randomization

Morning: Methods presentations on randomization strategies

  • Mechanics of replicable randomization
  • Strategies for randomization: simple, clustered, factorial, intertemporal

Lunch time presentation: A refresher on key econometric concepts (means, standard deviation, standard error, the CLT, differences in means, and Hypothesis Testing)

Afternoon I : Design Workshop

  • Small group presentation of design form
  • Focus on randomization strategies for each design: Implement simulations
  • Focus on strategy for hypothesis testing for each design: Implement simulations


Wednesday, 22 April 2015: Threats to randomized experimental design

Morning: Methods presentations on threats to inference

  • Partial compliance: LATE and ITT
  • Spillovers & Attrition

Lunch time presentation: Maarten Voors: Violent Conflict and Behavior: a Field Experiment in Burundi (or other)

Afternoon: Design Clinic

  • Feedback and questions
  • Figuring out the power for each study
  • Work on research design


Thursday, 23 April 2015: Design Based inference and Design Presentations

Early Morning: Presentation on Design based Inference: Means, variance, adjustment, randomization inference.

Late Morning: Design Clinic I

Small group presentations of revised designs, addressing sample, randomization approach, potential threats.

Lunch time presentation: Dan Nielson: Information and Accountability in Uganda (or other)

Afternoon: Design Clinic II

  • Feedback and questions
  • Revise research designs, set up and troubleshoot simulations for tomorrow


Friday, 24 April 2015: Mock Implementation: Simulations and redesign

We will use the last day to go more deeply into designs in parallel sessions. Working in groups of 4 we will collectively simulate the type of data that we might expect to result from your designs, and use this to:

  • Generate Descriptive Statistics
  • Implement hypothesis tests: lets check that your hypotheses can indeed be run on your data. What do you find? We will run your analysis many times to figure out your statistical power.
  • Interpret expected and unexpected findings
  • Trouble shoot: what happens if you suffer from missing data, non-compliance, or worse?
  • Revise design