Sections

Services

Information

UK United Kingdom

Science is imperfect – you can be certain of that

THE STATE OF SCIENCE: Former Chief Scientist for Australia Professor Penny Sackett explores how we deal with uncertainty in science. Listen to the podcast below for more. Professor Penny Sackett – Uncertainty…

Imperfection doesn’t stop scientists from seeking answers. Cea

THE STATE OF SCIENCE: Former Chief Scientist for Australia Professor Penny Sackett explores how we deal with uncertainty in science. Listen to the podcast below for more.

Professor Penny Sackett – Uncertainty and Science, Adelaide Festival of Ideas

Uncertainty is a fact of life. We live with it every day, but that doesn’t stop us from having productive and joyful lives.

As scientists, we know our measurements are imperfect at some level, but that doesn’t stop us from pursuing science.

Our decisions are affected by uncertain scientific knowledge, uncertainty caused by ignorance and the human emotion of uncertainty. So it is important to consider the process of science, what scientists mean when they talk about uncertainty, and the impact all of this has on our decisions.

The perpetual cycle of science

Science is the interplay of three basic steps conducted in a never-ending cycle:

1) observation and measurement of nature,
2) organisation and synthesis to form models of nature (whether built of wood or computer algorithms),
3) development of predictive theories, or principles, that explain how these models interact.

The process is cyclical because the predictions of theories are tested with new observations. The more tests the theory “passes”, the better the theory. But there is always uncertainty due to a limited number of measurements, incomplete models, or assumptions that may be inappropriate.

The scientific process works precisely because the cycle of science generates more measurements, modifies models, and tests assumptions. When a scientist publishes a result, it takes the form of a number with an “uncertainty”. The result is not the number alone, but the number and the range of relative certainty taken together.

World champions of uncertainty

Even Nobel-prize winning results have uncertainty. Australian Brian Schmidt was recently awarded the Nobel Prize in physics for transforming our understanding of the essential forces at play in the universe.

Beginning 1994, he and his colleagues began a program of analysing supernovae explosions in the distant cosmos in order to determine what has happened since the Big Bang. They found the universe isn’t gradually slowing down, as you would expect if gravity were pulling the pieces back together again, but instead flying apart at an ever-faster rate.

This result was so surprising, Professor Schmidt himself says that at first he thought he and his colleagues had made a mistake. They checked their data and went over all of their assumptions, testing the ones they could. Finally, after satisfying a referee that their methods were reasonable and sound, they published their results.

Unless you were a cosmologist, you might not have been able to tell how exciting their result was, because this is how, as careful scientists, they described it:

“For a spatially flat universe composed of normal matter and a cosmological constant, we find that Ωmatter is 0.4 (with a range of 0 to 0.9) and ΩΛ is 0.6 (with a range of 0.1 to 1). An Ω = Ωmatter = 1 universe is excluded with greater than 80% confidence.”

In other words, they were 80% sure there is something other than just normal matter and gravity operating in the universe, at least if some commonly-held assumptions are true.

That wasn’t the end: it was the beginning.

Another team published similar results at the same time. Other scientists offered different testable explanations for the supernovae results. Still others thought of completely different ways to measure the expansion rate of the universe. And the supernovae work continued, with more and better measurements to test the underlying assumptions.

All of this took years, and the result withstood scientific scrutiny. There is still a range of uncertainty, but it is much smaller than when Professor Schmidt’s team published the first result.

Uncertain times

Why is it so important to talk about science and uncertainty just now?

Because today, in this era of rapid, if not always responsible, communication, we are in a position to be more aware than ever about the enormous importance, and the intrinsic limitations, of science. But our ability to process and come to grips with this deluge of information has not caught up. We can feel overwhelmed, paralysed, and, well, uncertain.

As humans, we want to be certain. We demand the impossible: “fail-safe” transportation, “flood-proof” infrastructure, and medical guarantees. Social scientists tell us humans often freeze in the face of uncertainty. A pause before action in the face of the uncertain can be a good thing if it causes us to consider evidence that matters more carefully.

But not all evidence, not all uncertainty, is equally important for a particular decision.

What is important is that we weigh the risk of inaction against the risk of a less-than-perfect action. Because inaction is actually a choice, and thus an action. Choosing not to act is a choice to relinquish some influence over the future, yielding it instead to others, or to chance.

If we don’t act, we lose the ability to reduce the very uncertainty that troubles us. We lose the ability to learn, even from mistakes. In a changing world, “good enough for now” with new learning is often better than striving for perfection at some later date.

Values and evidence

In the face of uncertain evidence, how do we choose?

Social science also tells us that humans commonly rely on values, as opposed to evidence, to make decisions. Invoking values, or principles, in addition to evidence, is not necessarily bad. Wise and humane action requires both objective evidence and subjective values.

Solid scientific evidence tells us what is likely to happen under certain circumstances – not with absolute certainty, but with a reasonably well-understood degree of uncertainty.

Science can help us assess the consequences of our choices.

But science cannot tell us what we should do. It can’t tell us that we should stop smoking, that we should reduce our production of carbon dioxide by a certain amount or that we should use a particular medical treatment.

What we choose to do will be affected by our values, particularly in the increasingly complex world in which we live. Choices we make in one arena affect other areas of our lives, and the lives of others.

That is why both scientific evidence and values matter.

Acting on uncertainty

So what can we do? At least two things.

First, let’s not confuse evidence with values. Or, in the case of the unfortunate and dangerous climate change “debate” in Australia, let’s not confuse scientific evidence with possible policy choices and actions.

Second, let’s discuss openly, sincerely and civilly, both the assumptions on which the evidence rests and the values that we use to judge our actions.

Scientific evidence is discussed in peer-reviewed scientific literature, at least among scientists. But we need also to discuss our values, and examine how certain we are of those values. When we say “this is what we should do,” we are making an assumption that we share the same values. This assumption can be tested through open, sincere and civil dialogue.

Not only with those who share our values, but also with those who might challenge them – not with the aim of preaching, but of learning.

This article is based on the Keynote Address by Professor Penny D Sackett at the Adelaide Festival of Ideas on October 7.

This is the fifth part of The State of Science. To read the other instalments, follow the links below:

Sign in to Favourite

Join the conversation

4 Comments sorted by

  1. Bruce Caithness

    Retiree

    Karl Popper would have totally rejected Penny Sackett's description of science as being based initially on observation. Penny is not alone as Richard Dawkins also consistently repeats the empiricist prejudice that our knowledge derives from sensations. It is sensationally amiss. Guesses come first. Inferences are the root of all our mental processes. The quest for good explanations is the basic regulating principle of science. An explanation is said to be corroborated if it withstands its tests. All that is proven is that it has not yet failed testing.

    A brilliant essay on science is that of John Barnhart: Karl Popper: philosopher of critical realism

    http://findarticles.com/p/articles/mi_m1374/is_n4_v56/ai_18501025/?tag=content;col1

    report
    1. Penny Sackett

      Professor at Australian National University

      In reply to Bruce Caithness

      Thanks so much for your comments Mr Caithness, and for taking the time to read and reflect on my contribution.

      I would agree that science attempts to explain: explain what we observe in nature and what we might observe in future.

      I did not say (in fact was careful not to say) that the perpetual cycle of science begins with observation. I listed three steps in a cyclical process, each of which depends on the other. (The audio link describes this a bit more). The first step in the list was observation…

      Read more
    2. Bruce Caithness

      Retiree

      In reply to Penny Sackett

      Thank you Professor Sackett for responding to my criticism. On listening to the podcast I recognise that your paper was more subtle than my initial reading of the text. It was also heartening to hear your reflections on the distinction between scientific findings and values and the challenge this entails.

      report