On Models and Scaling Up

I was recently sent a proposal by the designers of a project who intended to demonstrate a particular approach to undertaking development work in a geographical district in a developing country, and if it was successful they then intended to ‘scale up’ the model to other districts . This was, they said, in order to overcome the piecemeal approach of just working at village level, which led to uneven development. The models embraced both the technical and the social – technical in terms of engineering solutions, but social in the way that they intended to work with different groups to encourage them to commit to the engineering solutions. The idea of modelling assumed that the same outcomes were possible with standardised approaches to both objects and people.

One of the difficulties that this presents is of assessing the effectiveness of the models in their own right as distinct from the organisation’s staff taking up these models with other, local people. The premise seems to be that if the models ‘work’ then anybody can take them up elsewhere with the same effect. This, of course, is the basis of scientific thinking as it implies to the natural world. A method is generalisable if anyone can apply it with the same results. However, if  effectiveness is in good part due to both the quality of  thinking about method (models) but also the calibre of the people who are working and the quality of the relationships  they are forming  with others to help them work, then there is no separating out the contextual from the generalisable. Success will arise from a whole host of local and national factors, while the idea of ‘scaling up’ implies that it is the generalisable factors which are the most important.  What is emphasised, then, is abstracting from the context and the privileging of the general over the particular.

However, in a different district, if local staff  are required to take up the model and are unsupported, they will no longer be working with the same initial starting conditions. By attempting to scale up it may be that this project will simply bring about uneven development at a district level instead of uneven development at a village level. In the absence of a similar intervention in a different district by the same organisations, or a strong state which can impose ways of working on different districts and can match this with resources, it may be difficult to achieve the same effect elsewhere, and to separate out the effect of the intervening organisation and its staff from what is happening.

In this modelling approach it is possible to see the vestiges of the ideas contained in a  post on Hannah Arendt that we wrote previously where there is a concept of a true ground,  a first cause which can be discovered, and in this case ‘applied’ elsewhere.

If we were to take up William Easterly’s arguments in The White Man’s Burden,  then we would abandon ideas of replicability at scale as a utopian myth. ‘Success’  in one district will be contingent on a whole host of factors including the level of education of the population, the prevalence of the required skills and know-how, the local readiness for such an approach, cultural norms, political struggles, in other words what is known in sociology as the habitus. What you want to achieve, what local people want to achieve, and what is possible may all be irreconcilable. We do have sustainable services in the developed world, but one explanation of how they came about would argue that the ‘readiness’ of society bump-started sustainable services such as health and education rather than the other way round, or at the very least they co-created each other. In Western societies the combination of a strong state raising taxes, the level of education amongst the population,  the development of thinking about social services all made broad scale developments possible, which in turn furthered thinking about social services which does not cease to this day. The ‘model’ of services, which itself is constantly changing, was only one factor amid a host of factors which contributed to the development of services over a long period of time.

This is not to make an excuse for doing nothing, nor is it an argument that there is nothing generalisable from experience which is useful and applicable elsewhere. It is, however, an argument for heeding what James C Scott had to say in Seeing Like a State, which we have explored in other posts. Large-scale, abstract plans which envisage ‘whole’ change are fraught with difficulty, and there is no necessary basis for thinking that because you have proved x you can also prove x to the power 10. In complex situations different initial starting conditions can cause all kinds of different outcomes to come about, which may or may not achieve what this particular organisation intends.

Chris Mowles

Related Thailand Models Articles

Leave a Reply

Your email address will not be published. Required fields are marked *