Wednesday 7 October 2020 4:46 am

Following the science? This government lacks a basic grasp of the scientific method

Paul Ormerod is an economist at Volterra Partners LLP, a Visiting Professor in the Department of Computer Science at UCL, and author of Against the Grain: Insights of an Economic Contrarian, published by the IEA in conjunction with City A.M.

Verification and validation. It is hard to imagine a more nerdy phrase.

But it is, in essence, how science makes progress. It is what we have to do to check whether a scientific claim or theory is correct.  

And it has been seriously neglected during the Covid-19 crisis.

Read more: Technical glitch in UK’s coronavirus testing system fixed, minister says

Just over a century ago, for example, Albert Einstein revolutionised the world of physics with his general theory of relativity. The verification part of the process was to check that his high-powered maths did what he said it did.

Einstein claimed his theory was superior to the very longstanding one of Isaac Newton. This had to be validated by confronting it with real world evidence before it was believed. Some of the tests were pretty esoteric, involving things like fluctuations in the orbit of Mercury.  But Einstein’s theory passed with flying colours. It was both verified and validated.

Last week, Matt Hancock, the secretary of state for health, claimed that without the new lockdown measures there could be “hundreds of thousands of deaths”.

A quick but effective piece of validation shows the scientific model on which this is based is most unlikely to be true.

The evidence now suggests that, at the peak of the epidemic in late March/early April, there were around 100,000 new cases each day. Total deaths in these months were around 40,000.

So to get “hundreds of thousands” of deaths — which means at least 200,000 — simple arithmetic shows that there would have to be a minimum of half a million new cases every day.

We can add to this calculation the combination of better knowledge of how to treat the virus and the continued shielding efforts of the most vulnerable, which have caused the death rate to fall sharply.  

If Hancock’s claim is correct, there would have to be a million new cases a day at a second peak.

Even Chris Whitty and Patrick Vallance in their horror slideshow last week only felt able to claim 50,000 cases each day.

But the gloomy pair appear to have ignored the basic principles of verification and validation way back in February and March.

The lockdown was triggered by the predictions of Neil Ferguson and his team at Imperial College. Unless it was introduced, they projected, there would be 510,000 deaths.  

Almost as soon as it was published, a range of scientists raised questions about verification of Ferguson’s model. Did the computer code do what was claimed? Ferguson himself tweeted “I wrote the thousands of lines of undocumented code 13+ years ago to model flu pandemics”.  For many weeks after it was not available for scrutiny.

It was rapidly pointed out that the model did not allow for any change in behaviour.  In the face of a pandemic, it assumed that people would carry on behaving exactly as they had always done — hardly consistent with the evidence of how people really do behave in these circumstances. Yet this was the basis for the lockdown restrictions we have seen over the past six months.

Verification and validation: boring words indeed. But their neglect by the government’s advisers has saddled us with the enormous social and economic costs of lockdown.

Read more: Coronavirus: Tory MP slams local lockdowns as ‘devastating’

Main image credit: Getty

City A.M.'s opinion pages are a place for thought-provoking views and debate. These views are not necessarily shared by City A.M.

Share: