Tuesday, June 30, 2020

Models: Some Tips

Environmental & Science Education
Edward Hessler 

Computer modeling of COVID-19 has captured considerable public attention and is reported on regularly.

A June 24 report in the British scientific journal Nature discusses five ways models can serve rather than confuse us. The authors refer to their essay as a manifesto. In other words "strong stuff."

They make use of the Brit term "mind" as in the automated train warning phrase "Mind the gap"--to pay attention, take caution, note the space when crossing from train or to platform or reverse. For full details see the essay.

--Mind the assumptions, especially the associated uncertainties.

--Mind the hubris. It is not easy to develop models that are useful and capture reality but it is easy to think that we have. The trade-off is one between the breadth and complexity of the model. 

--Mind the framing. Models are influenced by the choice of tools the model builders use as well as their disciplinary interests. The authors note that "Existing guidelines for infectious-disease modelling reflect these concerns (stakeholder involvement, multiple views, transparency, analysis of uncertainty) but have not been widely adopted. Simplified, plain-language versions of the model can be crucial. When a model is no longer a black box, those using it must react to assess individual parameters and the relationships between them."

--Mind the consequences. We like numbers and tend to trust them sometimes more than judgements. Numbers have a tendency to stop us thinking further about other relevant issues.

--Mind the unknowns. Models can hide our ignorance. The authors cite a trusted American expert and hero. "Experts should have the courage to respond that 'there is no number-answer to your question', as US government epidemiologist Anthony Fauci did when probed by a politician."

And finally a wise observation, a piece of advice. "Mathematical models are a great way to explore questions. They are also a dangerous way to assert answers. Asking models for certainty or consensus is more a sign of the difficulties in making controversial decisions than it is a solution...." (My emphasis).

No comments:

Post a Comment