. . . [T]his paper argues that ethnographic evaluation can make an essential contribution to impact evaluation by providing explanations as to why interventions have not been successful, as well as an evidence base for improving programme design and implementation generated by interaction with community members.
That is from the paper "Integrating Ethnographic Principles In Ngo Monitoring And Impact Evaluation" by Bell and Aggleton (Journal of International Development, 2012). The paper is not available online for free. I thank the authors for sending me a copy.
The authors conclude:
[Ethnographic evaluation] makes an essential contribution by providing robust explanations as to why interventions have not been very successful, as well as an evidence base for improving programme design and implementation. Key areas in which programme improvement cannot occur as a result of this deeper insight include a better understanding of (i) underlying principles and assumptions of programme approaches, (ii) the socio-cultural logics influencing young people’s decision-making, (iii) unexpected sexual health outcomes and (iv) local barriers to programme delivery and efficacy.
Of course, the key is to see this approach as a complement of experimental and cuasi-experimental ones (and the authors say that).