Belief is a beautiful armour

Earlier in the year, I went to a meeting that was opened by Sir Mark Walport with a talk about the difference between evidence and values. I think I have watched this thought process grow within the office of the Government Chief Scientific Advisor since Professor Sir John Beddington first pointed out that the clue to the limitations of his role were in the title – he is an advisor – and mostly politicians make judgements based on their values. Mark has developed this theme, and his talk gave examples of not just where evidence was used to make policy, but also where “values” gave rise to policies and (most damagingly) where the reasoning was confused.

One of the better quotes of the meeting was by a senior civil servant who said, “There is no point presenting a dossier of evidence that a policy will not work to a minister when he has just been elected to make it happen!” There was also a slight whiff of intellectual arrogance in the questions that those driven by evidence were somehow better that those driven by belief – whereas I think Mark was saying they were different and should be recognised as such.

This all got me thinking about evidence. I was once trained as a scientist and the basic principles have never left me. As I understand it, science starts with wanting to understand something. At first, you measure the thing you want to understand. After a while you have enough data to construct a hypothesis about what is happening. That hypothesis can then be used to predict the result of a measurement you haven't yet done – and the veracity of that prediction tells you if your hypothesis is correct or not. Often there is no current experiment that can be carried out to verify the hypothesis and so the scientist enters a period where they “believe” their theory but cannot yet “prove” it. Sometimes the experiment is difficult to control to be reproducible (a current cause that many are addressing) and the scientist has to “edit” their results (Millikan is said to have discarded some results from his oil drop experiment because they did not fit his evolving hypothesis). Often several scientists are competing to explain a phenomenon and have different hypotheses – and (in extreme examples) spend equal time trying to prove their own ideas and rubbish their competitors! I saw enough disagreements in my early career to know that data is not wholly reproducible and different scientists develop different models to explain similar but not identical data. Each believes their theory and disbelieves their rivals (this has led to some spectacular disagreements at conferences!). The goal of all science, though, it to devise and perform the experiment that proves their hypothesis is correct. This leaves me with the belief that science cannot unambiguously claim to be right all the time, but will end up with the truth at some point.

The world of politics is different, in that it starts with ideology and it is difficult to prove the hypotheses on which many policies are based because of the complexity of the systems they attempt to describe and play out over long timescales. This means that if the ”experiment” does not go as predicted, it is usually easy to find “parameters” which could not be controlled to sufficient precision to enable the result to unambiguously say the hypothesis was false. Or, the time over which the experiment to give a result must be carried out is longer than a political cycle and someone changes the parameters because they have a different hypothesis. The holy grail of “evidence based policy” is often perverted into (and I am grateful for a senior civil servant using this phrase over 10 years ago) “policy based evidence”. This is the use of selective data that can be interpreted to show the hypothesis was correct and the experiment showed the predicted result. (Listen to most political announcements on such issues as climate change, our position in Europe or even the occasional war and they demonstrate various degrees of this).

What brought this all back to mind was the current debate about science and innovation funding. The science “lobby” is asserting that without strong investment in basic science through universities there will be little economic growth. The “austerity” faction is basically saying “we don’t believe you, prove it”, but has a problem in that no evidence can prove their case. This has led to some very nice analysis and a few less justified claims on the science side and frequent repetition of the requirement to prove the claims from the other. Last time around (in political cycles), the science lobby did a good job but only held the funding line rather than winning the argument. This time it is proving more difficult because they are up against the unfettered belief that austerity is good for all of us.

Having worked in industry for the first 26 years or my career, I have observed (and have many qualitative examples) that the development of new technologies (both product and process) depends on the existence of the relevant scientific understanding to be successful (and robust!), but that the full attainment of impact requires something more. The missing element is the linkage between what is possible and what is profitable (or fundable for non-commercial goals). This usually comes from people in industry who are also scientists and who learned their trade in universities, but who decided to apply their skills at the downstream end of the intellectual supply chain. They work as part of the multidisciplinary team that makes up most businesses and carry with them the technical needs of the market – and work with those scientists still in universities to jointly assemble the right package of understanding and implementation.

My time at the Technology Strategy Board taught me many things about business and its funding but, most of all, it reinforced and multiplied the observation that innovation requires both science and commerce to be successful, and that this, in turn, requires strong interaction between those scientists in universities and those in business. Many in universities are already predisposed to this interaction (although their principal metrics don’t necessarily drive this behaviour), but it seems to have become increasingly difficult for many in companies – even the large ones who might be expected to have the capacity and experience to encourage it – to find the time to seek out and interact with those in universities. The TSB was set up by combining many of the pre-existing government mechanisms for encouraging businesses to interact with universities and put them in the hands of people from business who knew the area. I would claim we had moderate success given the time we were allowed to do that, but was increasingly disappointed by the interference from government who used the flow of money to focus funding where they thought it would benefit them rather than lead to strong commercial success. Like the science base argument, it is a debate between a thin evidence base (it takes time for companies to go from grants to financial success and we had only been going since 2007) and a belief that businesses should be self-sufficient. Add in a drive to minimise government spending and you get the increasingly short-term nature of the required success metrics inhibiting the approach that was initially successful.

I also detect an element in the science base argument that more science is all you need and they should control all the money (there was always a suspicion that the funding for the TSB was taken from that allocated to the research councils and that it is a zero sum game!).

So, perhaps I should take my own medicine – what do I think/believe?

I believe that consistent national investment in the science base is an essential prerequisite for strong economic growth. There is enough evidence to support that this is true but it is not “proven” – but there is no contrary evidence. However, it is not enough on its own.

I believe that consistent national investment in the interface between the science base and businesses is also required. The evidence base that this is true is a lot smaller because government have not been consistent in their approach and therefore the “experiment” in not reproducible.

I believe that “innovation” is a different activity to that carried out in the science base, and confusing them undermines the overall effectiveness of the processes that enable science to contribute to economic growth.

Both processes take time to deliver “impact” and focussing on short term metrics is one way non-believers can undermine the evidence.