Lean experiments - Filling the evidence gap

ID: 

18659

Session: 

Long oral session 14: Issues in Global Health

Date: 

Thursday 14 September 2017 - 14:00 to 15:30

Location: 

All authors in correct order:

Wojkowska E1, Nakamura T1
1 Kopernik, Indonesia
Presenting author and contact person

Presenting author:

Ewa Wojkowska

Contact person:

Abstract text
The traditional delivery of development assistance makes it difficult for development-sector organisations to test new approaches to poverty reduction. As a sector we often commit to mid- or long-term programme designs without first collecting evidence and knowing whether the intervention is having the intended impact. This can lead to programmes and investments that are producing sub-optimal results. The good news is that the development sector is increasingly paying more attention to evidence collection in order to find the most effective approaches to poverty reduction and to guide investment decisions. Randomised-controlled trials (RCTs) remain the gold standard for previously unproven interventions.

However some gaps persist:
+ Due to the large scale, time and investment required, only a limited number of RCTs can be conducted.
+ RCTs are not appropriate for early stage interventions that may need many rounds of refinement before larger commitments are made.

There is therefore a huge opportunity to conduct rapid and lean evidence collection. Together with partners in the development sector, Kopernik (www.kopernik.info) rapidly tests interventions in real contexts to determine their potential to reduce poverty effectively, and collects and analyses data on their effectiveness. These learnings also feed into improving the intervention design. Such experiments and datasets are typically relatively small because central to this approach is to test solutions and measure impact in an efficient and rigorous (though not statistically significant) manner. While RCTs are still the gold standard for previously unproven interventions, leaner research methods are more appropriate to early stage interventions and new poverty-reduction approaches typically made possible by technological advances. This approach is complementary to other methods and builds a much-needed pipeline of promising solutions that deserve larger-scale testing and evidence collection.

Several examples of such experiments that Kopernik has conducted with partners will be presented.