Syndicate content

David McKenzie's blog

Is optimization just re-randomization redux? Thoughts on the recent ‘don’t randomize, optimize’ papers.

David McKenzie's picture
A couple of weeks ago, Berk blogged about a new paper by Bertsimas, Johnson and Kallus which argues that instead of randomization, it can be superior for power to choose the best of all possible allocations of subjects to treatment and control, where best is defined in terms of minimizing discrepancies in the mean and variance of the two groups.

Weekly links May 8: rainfall revisited, rethinking key results with long-term data, son preference, and more…

David McKenzie's picture

Weekly Links May 1: Trends in Impact Evaluation, JHR Symposium on Empirical Methods, Superstar Inventors and more...

David McKenzie's picture
  • The Growth Economics blog hits hard with “there’s more to life to manufacturing”, among other things, making the point that even the way we code industries and occupations is heavily biased towards manufacturing and misses most of the action taking part in services growth.

Are you teaching or taking a class in development economics in a developing country?

David McKenzie's picture
This is a joint post with Anna Luisa Paffhausen
 
We are currently conducting a study and survey on how development economics is taught in developing countries and would love your help getting the word out and/or participating.
 
Our survey is meant to be a stocktaking and study of whether and how developing economics is taught as part of an economics course in developing countries. We are focusing on undergraduate and masters level classes.
The aim is to use this to understand the following questions:
 

Weekly links April 24: When behavioral phenomena work and when they don’t, marketing skills for small businesses, spillovers, and much more…

David McKenzie's picture

Weekly links April 10: Online IE education x 3; monkeynomics, surveying under repression, and more…

David McKenzie's picture

Weekly links April 3: scaling up in perspective, surveying with mobile apps, and more…

David McKenzie's picture
  • Pew research on “what we learned about surveying with mobile apps”: “Immediate responses and feedback can be helpful and revealing” but “App response rates were lower than Web rates overall and for each of the 14 surveys we conducted”
  • Freakonomics asks “How do we know what really works in healthcare?” – a podcast about RCTs. I like this comment from a doctor about scaling up “I think Medicare’s comment was that it’s really hard to do. We’re not sure we could scale it. Well, we f***ing scaled open heart surgery. We scaled separating Siamese twins. We scaled transplanting hearts and lungs, curing complex cancers. We’re sequencing the human genome. You’re telling me we can’t have a nurse go out and check on your mom or grandmother in a highly organized, well-structured, well-trained intervention for which someone’s already doing it for hundreds and hundreds of patients every day?”

Can you measure flows over short periods? Aka why Justin Wolfers might (NOT) want to reconsider that parenting study

David McKenzie's picture
[EDIT: I POSTED TOO HASTILY HERE, SEE AN ADDENDUM BELOW WHERE I AGREE WITH JUSTIN AFTER ALL]
Today in the Upshot, Justin Wolfers heavily criticizes a recent study that has received lots of media attention claiming that child outcomes are barely correlated with the time that parents spend with their children. He writes:

Tips for Randomization in the Wild: Adding a Waitlist

David McKenzie's picture
This is a relatively small point, but one that has come up several times in conversations in the last few months, so I thought it worth noting here.
Context: you are randomly selecting people for some program such as a training program, transfer program, etc. in which you expect less than 100% take-up of the treatment from those assigned to treatment. You are relying on an oversubscription design, in which more people apply for the course/program than you have slots.

Pages