If there’s no research in your area of interest, you have three options:
- Do something else
- Proceed with caution, testing as you proceed
- Guess and hope. This is a terrible option!
How to know whether there is relevant research, and how to proceed.
First, check what is known which is relevant.
Even if there is no rigorous evaluation of precisely the kind of programme / policy that you are thinking of in your location, there may be:
- Rigorous evaluation of that kind of programme / policy somewhere else. That evaluation may identify the mechanism by which the programme works, and what conditions are necessary for that mechanism to work. That may be enough for you to form a good view of whether that programme will work in your context, ie., you may be able to see whether those conditions are met in your context.
- Research about the underlying mechanisms. For instance, many programmes rely on making it easy for people to do things that we want them to do or making it hard to do things that we do not want them to do. This is an example of masses of research about what influences human behaviour. If your programme relies on people behaving in the ways that we know they do, and responding to incentives and stimuli in the ways that research has shown that they do, it is likely to work.
Second, if there is that kind of research, proceed with caution.
- Test whether the conditions for the mechanism to work are in place in your context. In the example of ‘sugar daddies’, that would mean seeing whether
• ‘sugar daddies’ are a cause of HIV infection, and
• teenage girls are over-estimating or under-estimating the risks of sleeping with older men.
If those conditions are not in place, stop, because the programme is unlikely to work and may even create harm. - Run a small trial to see the programme’s effects in your context. Look out for unintended consequences including harms. (A good example of a pilot checking for unintended harms is here.) Make sure that your trial has a good way to identify the causes of any changes that you observe, i.e., has a counterfactual.
- If that is successful, then you can expand the programme (if there is need)
If there really is not that kind of research:
- Consider whether you really want to run this kind of programme. There might be some other type of programme which can deliver the kinds of outcomes that you need and which has already been studied.
- If you want to run a kind of programme that has not been tested, then start small, and test as you go along.
- Do a proper needs assessment
- Run a small pilot and test. For evaluating the programme using a randomised controlled trial, Test, Learn, Adapt published by the UK Government is a great guide.