The moral frame of development research

by David Week on 22 July 2011


I’ve been reading a post by Tom Murphy entitled Aid’s God Complex and Bloggers Groupthink. Towards the end of the post, Tom says this:

Critics of research will decry the wonky conversations that do not address the issues at hand. Some will say that things need to get done and it is a waste of resources and time running an evaluation.

I think Tom’s missing the mark. I love research. I read research books and papers all the time: in psychology, in economics, in sociology, in anthropology. But what passes as research in development makes my skin crawl, and its not because of the reasons Tom gives. It’s because it seems to inhabit a moral frame which is the opposite to mine, and most of the people I work with.

I felt it was time I tried to put some structure around that visceral reaction, and provide a reason why most development research give me the heebie jeebies.

(By the way, Tom’s post triggered this one: but that doesn’t mean I think he holds the views I critique.)

Here goes:

The moral frame

What bothers me about most research in aid is that

Top down and distant

It is practiced “top down”, just like all that “bad aid” we talk about. A bunch of experts in an OECD country get a notion of how to make more “effective” aid, and then go and research it, returning to publish in OECD journals, talk at gatherings of the rich and powerful, and slap each other than back. When Hollywood stars behave like this, we reach for the vomit bucket, don’t we? Don’t we always say that one should ask the poor what they need? Why would this not so in the realm of knowledge, as in any other realm?

I remember being on the road to Ainaro, a few months after East Timorese independence, and stopping at a local market. I asked a guy leaning on our truck what he now needed most, expecting to hear things like clothes, shelter, credit, income… the usual suspects. He answered: “What we need most is up to date market information on coffee prices in town, so we don’t get cheated by the buyers.” People are not naive, stupid, or uninsightful. Above all, the life they are leading belongs to them, not to us. Why are they not included in research conversations? Why are they not the “client” for research?

Objectifying people

The form of the research treats its subjects as objects. It treats them as rats in a classic Skinner Box. Isolate one variable from all others (this is the purpose of randomization), and then by varying the inputs (this is the purpose of the control), see what kind of “behaviour” is elicited by the dear little lab rats. Do they take their immunization like good little children? Do they use their mosquito nets like Mommy wants them to?

Stimulus-Response. This is the tacit mental model of your average (and even not-so-average) development researcher. I think it morally vacuous. Do we not always say that people are the subjects and agents of their own development, not objects to be “developed” by others? Why would not apply this moral precept to the research community, as we would apply to any donor, agency, NGO or worker? Is this not just a moral precept, but a practical one, because the old precept—we can develop other people if we just find out “what works”—simply doesn’t work?

The coal face

The coal face of development is where “our” culture (I’m assuming the reader is from a donor country) meets the culture of the recipient. The coal face is in a thousand meetings and interactions and agreements made daily, between representatives of two cultures. These interactions occur at all levels of governance, from ministries to villages. In a post-colonial world, the interaction between those cultures has to be framed as an interaction between peers.

What I’m saying is that research agendas and methods cannot be decided in back rooms, away from that coal face. To do so is not good ethics. To do so is not effective practice. In particular:

  • Research agendas need to reflect what the “recipients” experience as the obstacles to their own efforts to developing their villages, economies and countries, and not what outside analysts see as obstacles to someone’s development.
  • The research model has to be framed as delivering practical advice to intelligent agents pursuing their own ends, not as advice to foreign program designers attempting to manipulate a population.


  • This comment comes from an email I sent David in regards to this post because the comments section were not previously working.  Now it is up and running, so I have agreed to share what I sent.I am always glad to cause another post to be written.  If I understand you correctly, your critique is largely with the current research methods (especially RCTs) and not with development research itself.  If that is the case, I offer no real disagreement.  I think I am definitely much more forgiving of RCTs than you, but concerns about being top-down and unethical are legitimate.  Thus far, I do not think they have been adequately addressed by randomistas.To me, development research has value because it helps to broaden or even challenge our understanding of an intervention or situation.  However, no single trial should be used to determine a complete policy.  It is unable to measure every aspect of what took place and is so specific that it seems nearly impossible to even consider scale.  Despite that, I do see value in the fact that it breaks up what was previously held to be true.Pardon the crude analogy, but I see research as one tool in the box.  Even if it is as important as hammer, that alone will not be enough to build a house.  There are many other components that are needed to realize the reality of eliminating poverty just like building a house.  Sadly, we have a lot of tools and no blueprint.  So we hammer away without nails or use screwdrivers to cut boards.As I said, I am always happy when something I write can cause others to jump into the discussion.  I dont want to sensationalize, but get meaningful discussion going.  Also, you helped to prove my point about aid blogging groupthink.  I really think that Shawn poorly applied that phrase.  There are many things which we do not completely agree.  That is what makes this space interesting and robust.  

    • Hi Tom. Thanks for that.

      I agree totally with your tools metaphor. As a philosophical pragmatist, it’s one of my favourite. Truth is what helps us cope successfully in our lives and in our projects. 

      To extend this metaphor, I also think you can ask of any tool: for whom has it made, and to what uses can it be put? A knife can be used by anyone, and has a many possible uses. A torque-wrench is designed for use by professionals (or expert amateurs), and is only useful in a narrow range of circumstances. This is why there are many more knives in the world than there are torque-wrenches.

      What bothers me about RCTs is just two things:

      • They are designed to be used by development designers and professionals, “on” the development population. Want to get those people to immunise? Bribe them with grain. That may “work”, but is not development.

      • The accompanying “gold standard” rhetoric: including bold statements that up until we randomistas showed up, no-one knew anything. This kind of epistemological narcissism is the worst thing you can have in development. It stops you from listening to people. In fact, if a field worker is found to be suffering from this dis-ease, they usually get sent home.

      I also agree that good research unsettles the taken-for-granted, established assumptions. I love that kind of research. I think very little of it is done in development, and that’s because most research is done by donors, who have a positive and powerful disincentive to discover to be told they’ve lost billions of dollars for nothing, and have to fundamentally change structure and attitudes.

      Again: RCTs don’t even come close to this kind of fundamental examination of the aid project.

Previous post:

Next post: