Esther Duflo speaking at Pop!Tech in 2009. Photo: kk + (flickr)

Ever wonder why some development projects succeed while others fail?

Esther Duflo and her colleagues at MIT’s Poverty Action Lab are working on the answer. Duflo is one of the newest recipients of the MacArthur Genius Grant because of her commitment to investigating what causes poverty to persist in some developing countries and what works to alleviate it.

She does this by setting up controlled field experiments in some of the poorest countries in Africa and South Asia. These experiments set out to prove how social and economic forces fuel the cycle of poverty in these areas. They also test how effective foreign aid projects are at lifting people out of poverty.

Duflo conducts her experiments using a method that mimics how drug companies conduct randomized medical trials. One group participates in a development project while the other does not. The differences between them are then measured to see if the project worked, and exactly how well.

Some of Duflo’s best known work is on HIV prevention in Kenya. Her research shows it’s more effective to teach girls specific ways to reduce their risk — like avoiding sexual relations with older men — than teaching basic medical facts about HIV and emphasizing abstinence as the best method of prevention. As she explains in her recent article for VoxEU.org, girls who were given risk-reduction information now use condoms more often, stay in school longer, and become pregnant less often.

“Economics is about the best way to allocate resources, and finding out what works is important to understanding how to allocate these resources,” Duflo told Philanthropy Action. Too few development strategies are vigorously tested. Proving what works can help.

(For more information on the Poverty Action Lab, check out Sarah Standish’s post "Researching Better Ways to End Poverty.")

December's Comment of the Month winner Jill Scantlan from Portland commented on Sarah Standish' post Researching Better Ways to End Poverty. She points out that community involvement is crucial to an aid program's success. She further argues that success should be measured both qualitatively and quantitatively. For her efforts, we will make a $25 donation to a project of her choice on Global Giving. Here's her comment:

This story brings up a lot of important issues concerning development projects. What is the most cost-effective way to improve literacy? Should we invest in providing lap tops for children or a mid-day meal program? How do we measure this?

I have spent a substantial amount of time observing and talking with NGOs in India who wrestle with these same issues. On the one hand, NGOs are usually bound to a government scheme, and on the other to an international aid agency. They have to prove that their projects are working and be able to measure that success in various forms of deliverables. In the end, some NGOs are successful and some make very little impact. What is the distinguishing factor?

What I have noticed is that the NGOs who use community participation in all (or most) of the stages of a project and who make the project fit the community (and not the other way around) are the most successful. In some cases, these approaches are not easily quantifiable.

J-PAL relies purely on quantitative measurements where a mixture of quantitative and qualitative would be more appropriate. A randomized sample, though statistically sound, can only tell you so much. Assessment and success goes beyond meeting targets and quotas. It is usually found in the attitudes and behavior of a community, which cannot be measured by numbers alone.

Keep writing in and share your though-provoking comments for a chance to win $25 towards the well-deserving charity of your choice!

* Lest anyone think $25 is not a lot, consider these figures from our affiliate Mercy Corps: $25 delivers clean, safe drinking water to 50 people in one of eastern Congo's sprawling displacement camps. $25 provides seeds to farmers in cyclone-devastated areas of Myanmar to plant five acres of rice. $25 gives traumatized children in Darfur 12 weeks of activities and psychological care to help them heal.

One of J-PAL's studies tested the effectiveness of computer-assisted learning in certain Indian schools. Photo: niyam bhushan (flickr)

A research group thinks the best way to determine whether aid programs work is to evaluate them using the scientific method.

J-PAL, short for for the Abdul Latif Jameel Poverty Action Lab, is a group of researchers - loosely affiliated with MIT - who help design and publish academic-quality studies of existing poverty alleviation programs in an effort to find out exactly what works and what doesn't. These researchers partner with non-governmental organizations (NGOs) to evaluate their programs. They also offer courses for other researchers to share their methods. (Class materials are available here for free.)

As J-PAL's website explains, the key to their approach lies in randomization, which is an important part of a well-designed study:

Suppose we would like to see whether one thing (e.g. "schooling") really improves a life outcome (e.g. "health"). The natural instinct is to compare the health of those who have schooling to those who don't. But this would be like comparing apples and oranges: People who have been to school are different in so many ways from those who haven't. Perhaps they have more advantaged social backgrounds, greater access to government services and so better access to schools. And only a few of these factors can be measured and accounted for in a standard statistical analysis. So this simple comparison — those with schooling to those without — may tell us little about the effect of schooling: It may instead be the effect of any of these numerous other differences like social background. If policy are set on the basis of such apples and oranges comparisons, quite a bit of disappointment may result.

J-PAL uses this approach to evaluate existing programs. For example, J-PAL researchers cooperated with an NGO called Pratham to study how much weekly computer use could boost Indian students' academic achievement. They introduced computer-based learning only into randomly selected schools that Pratham was already serving. Students at these schools received basic computer skills training and two hours per week of independent computer time with educational software. After a year, J-PAL found that these students' math test scores had risen, but that their other skills hadn't changed significantly — and all for slightly more money than another effective Pratham program J-PAL had also evaluated.

What makes J-PAL's work innovative is that such randomized studies haven't typically been used in evaluating poverty-alleviation programs, or even in the wider field of economics.

Such data is designed to help aid organizations and governmental bodies decide the most effective way to allocate their funds. For the extra cost of designing an extra study now, J-PAL believes, more money can be directed toward the most effective programs for better poverty-alleviation strategies in the future.