1 / 19

Thinking about Measurement and Outcomes

Thinking about Measurement and Outcomes. Esther Duflo MIT and Poverty Action Lab. Today’s Objectives. We now have a question and a design for the evaluation. How do we prepare the data collection? What data should we collect? What sample size do we need to plan?.

iris-jensen
Télécharger la présentation

Thinking about Measurement and Outcomes

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Thinking about Measurement and Outcomes Esther Duflo MIT and Poverty Action Lab

  2. Today’s Objectives • We now have a question and a design for the evaluation. • How do we prepare the data collection? • What data should we collect? • What sample size do we need to plan?

  3. The Setting: Reservation in the Panchayati Raj • What are the main goals of the Panchayati Raj? • What are the main characteristics of the reservation policy? • What are the salient features of the Indian context?

  4. The Controversy About Reservations • Why were reservations deemed to be desirable in this context? • Why are some people doubting that reservations would be effective?

  5. The possible effects • Let’s start by drawing a list of everything we think reservations for women may affect.

  6. The Evaluation Question • Suppose budget was not an issue: should we go out, hire a large team of investigators, collect data on every single of these indicators in reserved and unreserved communities, and see what happens? • Pro: • Cons:

  7. Interpreting Multiple Outcomes • Suppose that you collect data on 20 different public goods • You find that for water wells, investment in the public good is significantly higher in Panchayats reserved for women. • You find that for irrigation, investment in the public good the investment in the public good is significantly hither in Panchayats that are not reserved. • You find that for 18 goods, the outcomes very similar, and not significantly different in both Panchayats • What can you conclude?

  8. Interpreting Multiple Outcomes • Suppose that you collect data on 20 different public goods • You find that for water wells, investment in the public good is significantly higher in Panchayats reserved for women. • You find that for irrigation, investment in the public good the investment in the public good is significantly hither in Panchayats that are not reserved. • You find that for 18 goods, the outcomes very similar, and not significantly different in both Panchayats • What can you conclude? The hypotheses to test must be defined before the beginning of the experiment, or we have no good way to assess their validity

  9. The Need for a Model • Randomization of the intervention does not free us from understanding the chains of causality leading from the intervention to the outcomes. • We need to get the implementation team the objectives of the intervention ex-ante • We need to draw the link Intervention intermediary variables final outcome

  10. Defining an Objective • Textbook distribution project in Kenya. • The program had no impact on the test score of the average child • However the best students benefited. • Is the program a success of a failure?

  11. Defining an Objective • Textbook distribution project in Kenya. • The program had no impact on the test score of the average child • However the best students benefited. • Is the program a success of a failure? Defining an objective is critical for evaluation. It is a also an important learning step for an organization

  12. Drawing the Chain of Causality • What are the intermediate variables through which the intervention is affecting the final outcome of interest? • How are they likely to be affected?

  13. Drawing the Chain of Causality • What are the intermediate variables through which the intervention is affecting the final outcome of interest? • How are they likely to be affected? Defining and Measuring Intermediate outcomes will enrich our understanding of the program, reinforce our conclusions, and make it easier to draw general lessons

  14. Why would reservations make a difference? • What are possible chains of outcomes in the case of reservation? • What are the critical steps in order to obtain the final results? • What variable should we try to obtain at every step of the way to discriminate between various models?

  15. One possible model Reservations Imperfect Democracy Some democracy More women Pradhan Pradhan’s preferences matter Women are empowered Women have different preferences Public good reflect Women’s preference Different health, education Outcomes? Different (and specific) Public goods

  16. Getting Ready to Collect the Data • What variables do we need to test the entire model? • We now realize that we have used almost all the variables we initially listed. • What tools would allow us to collect what part of the data we need at each stage? What is the best (richest) tool, and what is the cheapest tool?

  17. Working within a budget • Collecting data cost money, and we may have to make choices. • If we had to choose one variable in this chart to determine whether women made a difference, what would it be? • You have computed the minimum budget to detect the impact of water wells at the village level. What else can we find out with the same activities, without spending any extra money, and to what extent can we populate the flow chart?

  18. Working within a budget • With a larger budget, what would we want to do? • Would a data collection in several stages be advisable? What stages would you recommend?

  19. The objective and chains in your proposals • What is the main hypothesis in your proposal? • What is the key variable of interest (on which you will declare success or failure) • What are the intermediate outcomes that you are planning to collect data one?

More Related