By SeanPublished On: January 26, 2019Last Updated: August 14, 2022
Let’s face it, not all science is created equal (in terms of funding, anyway). The current reporting landscape, for example, has placed the spotlight on topics like biotechnology and gene engineering. Most fields of science are lucky to have a media mention once every few years, while others get press conferences and sensational news stories on a seemingly weekly basis. Scientists in these fields must be doing something right, so what can we learn about their approach to research? Strong inference may provide the answer.
Table of Contents
It’s Not ‘Just Luck’
Taking a quick look at the list of Nobel Prizes in Chemistry, it’s striking that in recent years, most were awarded to advances in biochemistry-related fields. Compare this to the entire 20th-century prizes, where just a handful was for biology or biochemistry-related research.
You might reason that it’s simply luck; after all, every field in science progresses at its own pace. It is the nature of scientific discovery to have both fruitful and slower-moving periods. Perhaps current biochemists are riding the crest of a wave, the field’s ‘Golden Age’ of scientific discovery.
The particle physics bubble in the 1950s is another good example of this. It was a time when experimentation was turning up new discoveries faster than theory could keep up with. With advancement comes research interest, and of course, all-important funding. But this has to start somewhere, and the scientists, in turn, must continue to deliver.
Of course, luck does play a role in individual discoveries. But does this hold when an entire field is publishing like clockwork in top multidisciplinary journals? Or when these researchers in question are wiping the floor year-on-year, in the halls of Stockholm come December? Surely there must be something more, perhaps a paradigm in their ways of thinking. What could be the factors behind delivering tangible and consistent success?
The Theory of Strong Inference
By the systematic generation and disproof of hypotheses, the truth (or the closest representation of it) can be inductively inferred in the most efficient manner. This is the basis of the scientific method and the starting point for ‘strong inference’.
In 1964, physicist John R. Platt published in Science a controversial paper titled Strong Inference – Certain systematic methods of scientific thinking may produce much more rapid results than others. Platt argues that an exceedingly brilliant mind is not a requirement for delivering consistent results. Instead, he recommends a systematic and consistent approach to discovery:
Devise alternative hypotheses
Devise a crucial experiment (or several of them), with alternative possible outcomes, each of which will, as nearly as possible, exclude one or more of the hypotheses
Carry out the experiment so as to get a clean result
Recycle steps 1-3, making sub-hypotheses or sequential hypotheses to refine the possibilities that remain
Platt uses the biochemistry labs of Francis Crick as a prime example of strong inference in action; well, his paper WAS written in 1964. Written at the top of their blackboard would be an untested result, fresh from the labs or a recent paper. Beneath this result would be a number of different explanations (hypotheses) for it.
Experiments or controls were then planned by the research group with the aim of ruling out the possibility of each hypothesis. Throughout the day, members of the lab would drop by to correct or improve upon these experiments.
The Benefits of Strong Inference
Firstly, applying logical thought to all possibilities that arise from a result prevents favoring one hypothesis above the others. Traditionally, scientists are trained to think up a hypothesis and then look for evidence to support it. But there is a fatal flaw to this approach; the more we work on a single hypothesis, the more we become attached to it. After all, we are simply human. We want to see our work bear fruit, and we want to see our research turn into results.
This makes us liable for forcing theories onto facts, or facts onto theories. We perform even more experiments that seem to ‘support’ our favorite hypothesis, strengthening our conviction but leading us further away from the truth. Alternate working hypotheses remove this danger.
Nature doesn’t give away its secrets simply because we work hard to uncover them; as the saying goes, true experimentation lies in letting the chips fall where they may. In the next step, each hypothesis should be subject to systematic experimentation. Generally, the ‘weakest’ hypothesis is the one most vulnerable to experimental testing and hence can be ruled out early on.
After a series of experiments to test each hypothesis, the one left standing must be the closest version of the truth available. By repeating the process, we get ever closer to understanding the insights of nature.
Developing a ‘Sharp’ Approach
The fact of the matter is that the above methods are in use by scientists today after all, the disproving of hypotheses is the aim of any good experiment. But by following a systematic approach, the accumulation of information allows for new theories to be strongly inferred in the most efficient manner possible.
How can this translate into your own research success? First comes the proposal of different hypotheses that justify a result or theory. There should not be only one, but multiple working hypotheses. As mentioned, it is important to avoid falling into the trap of becoming attached to a ‘favorite’ hypothesis. Hence, each valid hypothesis must hold equal plausibility.
Next is the careful design of experiments that can put the hypotheses to the test. What does the hypothesis predict, and how can we quickly and effectively disprove it? This is where clean or ‘sharp’ experimentation comes in; we must only invest our time in experiments that give clear, insightful results. Only with explicit results are we able to systematically and confidently rule out one hypothesis after another.
Another important but easily overlooked area is the habit of ‘analytical bookkeeping’. Most of the time, our lab notebooks are used only for record-keeping purposes. However, Platt describes in his book The Excitement of Science that it should be used for more, citing Enrico Fermi’s famous ‘notebook method’. Apart from recording details of methods and observations, it is important to also include related questions, hypotheses and predictions. This allows for self-critical analysis, a platform for self-improvement.
Strong Inference Today
It is important to remember the real purpose of science, which is to explain the workings of nature. In our quest to understand the world around us, however, we are prone to overcomplicating experiments and losing focus.
In the end, a systematic approach to experimental design yields valuable ‘sharp’ results. The process does not call for a brilliant or exceedingly creative mind. Strong inference is something that we can all learn to utilize. And indeed, it is both a shame and a surprise that institutions do not teach this methodical approach to every graduate student.
Despite being over 50 years old, the strong inference technique is still very much relevant today. As our circle of knowledge expands, nature’s secrets also evolve to become more complicated, and more difficult to uncover. More so now than ever, alternative hypotheses and sharp exclusions provide powerful additions to any scientist’s toolkit.
By systematically and regularly applying these principles, we might be able to improve the rates at which our research generates success.
Platt, J. R. (1964). Strong Inference. Science, 146(3642), 347-353.
Platt, J. R. (1962). The Excitement of Science. Houghton Mifflin.
About the Author
Sean is a consultant for clients in the pharmaceutical industry and is an associate lecturer at La Trobe University, where unfortunate undergrads are subject to his ramblings on chemistry and pharmacology.