Thoughts on Science and Predicting the Future

Science Blogs: The Science of Predicting the Future

Six Italian scientists were sentenced to six years in prison last week for not giving the public sufficient warning of the L'Aquila earthquake. The CNN video is included below. The scientists are appealing the decision, but the fact that this case occurred is sparking debate about science and public policy.

Ethan Siegel at The Atlantic offers some thoughts on the challenges we face. We frequently make our estimates about the future based on repetitive experiences from the past:

This is a fabulous example of a pre-scientific prediction! I've taken information from very, very similar situations that I've experienced before, I know — looking back — how those previous situations turned out, and so I can infer how this current situation is likely to turn out. This is something we do all the time in our lives, and something we've done frequently throughout history. The phrase Red Sky at Night, Sailor's Delight didn't come about because we understood the science behind the next day's weather and the properties of the atmosphere the night before, it came about because when we observed phenomenon A (the red sky at night), it was very often followed by phenomenon B (good sailing weather the next day).

We use this all the time in our lives ...

Science is different:

Sometimes, this type of pre-scientific prediction is the best we can do. If we can make this into a truly scientific prediction, we stand to do much better, but it's a much more difficult task. A truly scientific prediction requires the following three things:

1. that the scientific theory that governs your phenomenon is completely understood,
2. the conditions that will affect the possible outcome(s) are known and understood in their entirety, and
3. that you have enough computing power to figure out what the outcome is going to be.

In addition, because measurements are imperfect (and sometimes physical laws aren't 100% predictive), you are also going to have a quantifiable uncertainty associated with your scientific prediction.

So, depending on the topic, scientists' ability to offer meaningful guidance to the public will vary. But if we are to translate science into meaningful action, we also need at least three other elements present in society:

We can speak intelligently about what the outcome will be in terms of probabilities and uncertainties, but this also requires a few things that are far from given:

1. Scientists who can communicate these results clearly and effectively,
2. A media / government that can understand that information, make reasonable and effective policies based on that information, and communicate these results to the populace, and
3. A populace that’s scientifically literate enough to understand what’s communicated to them and act in accordance with those recommendations.

This ought to be one of the main goals of science, as it's one of the most important services that science can perform for a society.

Toward the end of his essay, he writes:

You do the world a disservice when you scapegoat scientists for a disaster they could not predict and an incompetent government official they could not control. We now live in a world where we jail scientists for failing to clean up the government's miscommunication about a disaster they could not predict, while we simultaneously accuse them of fear-mongering for the impending disasters that good science does predict.

If you want to know what's going to happen in the future with any sort of accuracy, you need science. It's the only thing that's ever worked, and the more we do it, the better we get at it. This means we need to make the world safe for scientists to do science, we need to treat the science being done with the respect it deserves, and we need to improve and encourage communication between scientists and the public.

Anthony Giddens wrote about these challenges in Runaway World: How Globalization is Reshaping Our Lives. Increasingly, our challenges are far more complex than our time-honored heuristic decision-making models can handle. In the face of an impending disaster (Giddens uses the Mad Cow epidemic as an example), if scientists don't warn strongly enough, they could be vilified like the recently convicted Italian scientists were. If they warn too strongly and nothing happens, or if their strong warnings lead to actions that prevent a disaster, then scientists can be labeled scaremongers ... the bad thing they predicted didn't happen.

There is no easy solution to the dilemma. I think one other element should be added to the discussion. It probably falls under Siegel's comments about scientists communicating clearly. Many of the problems we face require buy-in from the masses. As soon as scientists become too tightly connected with a particular political agenda, they risk having scientific information collapse into a partisan football to be kicked around rather than science becoming a guiding light to shape policy. Scientists should be very slow to connect scientific information to political agendas and resist being taken captive for political ends.

What do you think about the Italian court decision? What challenges do you see for science as it relates to public policy?