L7: Randomized Evaluations 1

Go to Home
Do the reading before this lecture
Slides for this lecture are available below. My notes.
Go to Next lecture

The material below is taken from the website J-PAL (Abdul Latif Jameel Poverty Action Lab) (highly recommended). Please check the website for more material.

Why Randomize

In the following lecture Dan Levy presents different impact evaluation methodologies, discuss the advantages of randomized evaluations, as well as which factors influence the choice of one impact evaluation method over another. Two case studies are presented to illustrate the benefits of random assignment, compared to other identification strategies

Slides; Case Study: Get Out the Vote in the US; Case Study: Vocational Training in Colombia.

How to Randomize

In this other lecture Joe Doyle llustrates how the research question determines the randomization strategy and shows how multiple research questions can be answered within one study. In addition, this lecture explores how a number of potential issues and pitfalls can be anticipated at the design stage. Finally, he uses two case studies to illustrate how the randomization design of an evaluation can be tailored based on the research question(s):

SlidesCase Study: Extra Teacher Program in Kenya; Case Study: Labor Displacement from Job Counseling Programs in France.

Power and Sample Size

In the following video Rachel Glennester introduces the concept of statistical power and walks through the factors which influence it. You can also find an exercise that explains the trade-offs in designing a well powered randomized trial, using the EGAP web application.


Please now answer Assignment 7 based on the reading for this lecture.

Deadline: Wednesday, September 30, at 15:15.

Suggested answers available here after the deadline.

Please connect to my Zoom (link in Absalon) on Wednesday, September 30, at 15:15


I will summarize the main ideas of this lecture and I will asnwer your questions.

After that you will have time to work online with your group. See below.

Group Work: Describe your behavioral intervention (go to Activity)

What might work to overcome the problem your group is working on. Remember to:

  • Focus on the specific bottleneck you have chosen before.
  • Describe your behavioral intervention in detail.
  • Keep it simple and avoid any unnecessary complication.
  • Remember to not change to many aspects at the same time (measurability): if you change more than one thing, consider having also more than one control group, to break down the effect.
  • Plan an interventions that may be scalable (the intervention must be effective and feasible).

Together with your group, you have to fill the corresponding cell of this Google Spreadsheet. Deadline: before next lecture.

Summary for this lecture:

  • A
  • B
  • C

Go to Home 
Go to Next lecture




Go to Home
My slides for this lecture
Go to Next lecture