On Saturday 31 October 2020 at 6 pm, Prime Minister Boris Johnson announced a second English nationwide lockdown, to commence on Thursday 5 November. Before he revealed the widely leaked details, he invited Chris Witty, England’s Chief Medical Officer, and Sir Patrick Vallance, the UK Government’s Chief Scientific Advisor to explain the data on which the Government had based its decision. Shortly afterwards, I began to see posts on my Facebook feed, claiming that the data were false and provided spurious justification for a decision that was based on political expediency rather than on impartial scientific reasoning.
By chance, I was at the time in the middle of reading ‘The Believing Brain’ by Michael Shermer, a book about how we humans form beliefs about the world. So, I found myself reflecting on my ideas about data and the role that they play in making high-quality decisions.
In my professional life, I have always relied on data to lead my decisions. Early in my career, when I was managing director of a company making spectacle frames, I based a very successful long-range business plan on a detailed analysis of unit manufacturing costs. Many years later, as a project management consultant, my colleagues and I built our business around data collected from diverse organisations and then analysed to identify superior management practice. Shortly before I retired, I spoke passionately at project management conferences about the merits of ‘evidence-based management’.
When I retired from full-time work, I served for a time as a member of the Steering Committee of ‘Project X’, a joint UK-based government, academia and business research network. Some of the research, together with what I knew of research elsewhere, convinced me of the value of big data coupled with machine learning to decrease industry’s costs of prediction and increase its accuracy of forecasting.
But what my Facebook newsfeed and Michael Shermer’s book both reminded me is that data always require interpretation.
Yes, data indeed lie at the heart of the scientific method. Data obtained through experimentation either validate or disprove particular hypotheses. But as the hard, clarity of physical science shades into the more ambiguous sphere of social science, we enter the realm of what Shermer in his book calls “integrative science”: data used in conjunction with both ‘theory’ and ‘narrative’. It is this combination of three elements, he convincingly argues, that creates ‘believability.’
Which brings us back to those Facebook posts I received that were, in effect, crying ‘foul’ about the data presented in Downing Street on 31 October. Their authors didn’t trust the Government’s narrative, and so they called into question the data that had supposedly led to the decision to impose a second lockdown.
The trouble is that ‘narrative’ is, in effect, ‘story’ and while stories are central to the human psyche, they are so powerful a device that we should handle them with care. (See ‘The Fascination and Danger of Stories’.) When people use them to create ‘believability’ or to justify a particular decision, the narrative often masks the theory that underlies it, whilst the data provide evidence that supports its storyline.
Confirmation bias inevitably inclines us to agree with narratives and data that support our prior beliefs and theories. We are pre-disposed to accept “integrative science” if it comes from a person or group with whom we identify. On the other hand, if we have already decided that we are not at one with the people proposing a particular narrative, then we are unlikely to accept either their data or their description. As I wrote in the stories article quoted in the previous paragraph, “once we have linked a particular story to our identity, we are no longer open to evidence that casts doubt on its truth. We don’t use our formidable powers of intelligence and rationality to decide whether or not we are right to identify so strongly with it. On the contrary, we scan the evidence selectively for anything that supports our pre-existing point of view. And since most evidence, scientific or otherwise, contains at least a small degree of ambiguity, it usually seems to us to confirm what we already believed to be true.”
It would seem that whether data are good, bad or ugly ultimately depends on your viewpoint!
I would say that any verifiable and relevant data is better than no data at all, i.e guess work.
Great to hear from you, Max. I hope you are keeping well.
Creating a narrative around limited data but good knowledge of the organisational environment is in my view a fundamental tool for Project and Programme Managers. Essential to create a direction of travel then watch data closely and act to achieve relevant goals along the way