All the world’s a model (at least that’s how we see it) and what that means for the NHS
Have you asked yourself what Science does? Science seeks to explain things. It looks at the world and then produces theories of what is happening. A theory is a model of what is happening. We use that model to predict new events and the model remains valid until it is contradicted by observation and we have to find a better model that is able to predict the new events as well as cope with the old one. This is the view of Science that I learnt at School and can is encapsulated within Popper’s view of Science based upon empirical falsifiability published in The Logic of Scientific Discovery in 1934.
In the principal alternative view of Science Kuhn argues that this is not actually what Scientists do, that what Scientists do is actually more complex than that, and that change is more gradual with progress occurring in an incremental fashion until a crisis point is reached and a Scientific revolution occurs as outlined in Kuhn’s 1962 book The Structure of Scientific Revolutions.(TSoSR)
Both of these views of Science share the view that the models that Science produces are dynamic and change in the light of new evidence. In a recent article(link is external), John Horgan asserts that when he interviewed Kuhn for the 50th anniversary of TSoSR, Kuhn denied that science is constantly approaching the truth. At the end of Structure he asserted that science, like life on earth, does not evolve toward anything but only away from something. This has significant implications for those Scientists seeking to suggest that Science is in a search for ultimate truth which will remove the need for other attempts to find ultimate truth eg religion. But that is a subject for another blog.
However, whilst the Kuhnian recognition that Science does not progress smoothly and is at its heart as much a social phenomenon as a logical one, my experience of the way in which we use models is even more complex.
Newtonian mechanics remained supreme for centuries until the end of the nineteenth century, when a number of phenomena falsified the theory in classic Popperian fashion, eg the photoelectric effect first described by Heinrich Hertz in 1887. In line with Kuhn’s view a number of discrete lines of enquiry and propositions from the likes of Einstein and Planck coalesced into quantum theory a term which first seems to have appeared in 1924 and came to describe a field of study often regarded as the most important in the twentieth century. However, and here’s the rub, the cases where the quantum theory produces a different result from the classical Newtonian view are not the everyday and most people carry on with a Newtonian view of the world because the model is good enough for their purposes.
I saw this most clearly in my scientific training where I progressed through various models of atomic and molecular structure. Oddly, I remember first learning that the important thing about atoms was that they were indivisible; odd because the next thing I learnt was about their components; protons electrons and neutrons. As I started to specialise as a Chemist, I focused on electrons, and I was exposed to more and more complicated models that sought to explain how electrons behaved and became less and less about what electrons actually were. Eventually in my undergraduate days, I was treated to quantum mechanical models, but these were only really developed for the first two elements, which seemed to limit their value for chemistry. However, the real eye opener came when I undertook my first postgraduate project about intermolecular forces between hydrogen chloride molecules. For the purposes of this study, the molecule was reduced to a straight line with a positive charge at one end and a negative charge at the other, which was massively less sophisticated than other models I had seen to that point. This was done for two reasons: first, we couldn’t actually solve the problem using any of the more sophisticated models, and secondly, it was good enough to explain and predict how molecules would behave, which was the aim of the study.
This it seems to me is how we select scientific models in the real world. The quality criteria is not “What is best?” but “What is good enough?” and the key constraint is feasibility.
Thus if we need to explain how to get to the end of the road, we do not invoke a model which includes the fact that the Earth is round. A flat earth model works just fine, and that is all we use, even though we understand that the Earth is not actually flat. The purpose of the model is to explain and predict what happens when you walk down the road, not define a truth about planetary geometry. Thus, a new “better” model may never completely replace the old model where that is simpler and can still explain and predict the problem under consideration.
When needs change, so the model needed to explain and predict may also change in response. Thus, in the 19th century when until the railways required a more sophisticated model, time was kept locally and varied say between Bristol and London. Once trains connected these cities a more sophisticated model was needed and the idea of a standard time across the country was developed.
In the less controlled messier problems situations in which the social sciences have to deal, far beyond the neat falsifiable problems that Popper considered, very often the model used is too simple and fails the “good enough” criteria. Consider the NHS. Until very recently, and possibly still in some quarters, it was viewed a monolithic organisation which could be managed using a command and control structure. In reality, most of its activities were carried out by private contractors who were never part of the system (aka General Practitioners) and even its larger organisational structures (typically the hospitals) had several parallel command control structures dealing with clinical, financial and managerial issues at the same time. However, what matters is not that the NHS is not a homogeneous organisation, it is that a command and control structure based on that model does not adequately describe and predict the behaviour of the NHS.