1.5: The Scientific Method

The scientific method is a means of uncovering and explaining physical phenomena. It relies on observation and logical reasoning to achieve insight into the actions and relations of the phenomena, and a means of relating it to similar items. The method starts with observations and measurements. Based on these and background knowledge such as mathematics, physics, the behavior of similar phenomena, and so on, a tentative explanation, or hypothesis, is derived. The hypothesis is then tested using experimental or field data. The precise nature of the tests depend on the area of study, but the idea will always be the same: namely to see if the predictions of the hypothesis are borne out under alternate applicable conditions or data sets. A proper hypothesis must be testable and falsifiable. That is, the hypothesis must be able to be proven wrong by subsequent observation. If it is not, then the hypothesis is better classified as philosophy. For example, Newtonian gravitation could be proven false by letting go of a suspended hammer and watching it remain motionless or fall upwards, rather than fall down toward the Earth. Similarly, Evolution could be proven false by the discovery of “fossil rabbits in the Precambrian”, to quote famous biologist J. B. S. Haldane (Haldane was being somewhat snippy, but in general, he meant anything that would be clearly out of the expected time-line, in this case a small mammal predating even the most simple creatures with backbones).

A hypothesis is tested by repeated cycles of prediction, observation and measurement, and also subject to peer review, that is, the critique of others in the field of study who may point out inconsistencies in the explanations or logic, errors in experimental design, and so on. This cycle continues until the weight of data and scientific opinion elevates the hypothesis to a theory. One might say that a theory is a hypothesis that has withstood the rigors of the scientific method. This cycle was well expressed by the Marquise du Châtelet¹. She explained hypotheses as “probable propositions”. While it would only take one observation to falsify a hypothesis, several would be required to vindicate it: “each non-contradictory result would add to the probability of the hypothesis and ultimately…we would arrive at a point where its ‘certitude’ and even its ‘truth’, was so probable that we could not refuse our assent”.

It is important to note that the scientific usage of the word theory is entirely different from its popular usage, which is perhaps closer to “hunch” or “seat-of-the-pants guess”. Also, a scientific theory is not true in the same sense as a fact. Facts come in three main varieties: direct observations, indirect observations and those that may be logically deduced. A direct observation is something that you have measured yourself, such as noting the time it takes for a ball to reach the ground when released from a given height. An indirect observation is something that may be inferred from other known quantities or proper historical data, such as “George Washington was the first president of the United States of America”. An example of the third variety would be “If \(x\) is an even integer larger than four and smaller than eight, then \(x\) must be six”. At first glance, it may seem that facts are highest on the pecking order, but scientific theories are much more useful than facts because isolated facts have very little predictive capacity. It is the predictive and explanatory ability of theories that make them so powerful.

Consider the following. Suppose you hold out a stone at arm’s length and let go. It drops to the ground. That’s a fact. You saw it happen. Unfortunately, by itself, it doesn’t tell you very much. Suppose you repeat this several times and each time the stone drops in precisely the same way as it did initially. This is beginning to get useful because you’re noticing a pattern, and patterns can be predictive. Now, suppose you pick up stones of differing sizes, say 100 grams, 200 grams, half a kilogram and a kilogram, and drop each of them in turn. You observe that they each hit the ground in the same amount of time. Further, you drop them from different heights and you notice that the higher up they are, the longer it takes for them to hit the ground, but they all take the same amount of time to reach the ground.

You might now formulate a hypothesis: namely that the mass of a stone doesn’t have an effect on how fast it falls from a given height and that height and fall time are directly related. Your hypothesis is predictive. Although you used only four sizes of stones and a few heights, your broadened hypothesis should apply to any stone dropped from any height. So now you (and a bunch friends) starting picking up random pairs of stones and drop them from random heights, and sure enough, you see the same effect again and again. If you do this enough and it is continually verified without exception, you might even make a “law of falling stones”, particularly if you were able to quantify the times and heights through careful measurement and reduce the relation to a nice formula. It is useful because you can now predict what will happen with any stone dropped from any height. But this law is rather limited. It only applies to stones because you may have noticed that stones drop much faster than pieces of cork. While you might then proceed to make a “law of falling cork”, that would unnecessarily complicate things. Instead, you could take a step back and try to figure out why stones and cork both fall, but not at the same rate. Eventually, you might discover that the difference has to do with air friction and you can now create a law governing falling bodies in a frictionless environment. That’s even more useful than the original “law of falling stones”.

But even this new and improved “law of falling bodies” doesn’t offer a lot of insight into what is really going on in the larger scheme of things. Through repeated observations and experiments this could be extended to cover not just falling bodies on the Earth, but the interactions between any bodies, including falling stones and cork on the moon, or the interaction between the Earth and the Sun, the Sun and the other planets, the Sun and other stars, and so on. What you’ll have arrived at is a full-blown theory of gravitation (Newtonian gravitation). Now that is an extremely useful tool. It helps us design airplanes, get satellites into orbit, even get people to the moon and back safely.

Thus a theory is a “best estimate so far”, a model to explain some observable aspect of the universe. It is understood that as our view of what is observable widens and our knowledge extends, so too a given theory may be refined. The Newtonian gravitation model was sufficient to describe the movements of the planets around the sun and is still used to plan the flight of spacecraft. In the early 1900’s, however, Einstein’s Theory of Relativity refined Newtonian Gravitation to include more extreme (relativistic) effects. It is important to note that we did not “throw out” Newton’s equations; rather, we now have a better understanding of the set of conditions under which they apply. While this trek towards more and more refinement is not truth in and of itself, to paraphrase the late Harvard paleontologist, Stephen Jay Gould, to ignore the weight of scientific data behind an established theory would be perverse.

References

¹Du Châtelet was that most rare of 18th century women: a mathematician and physicist. She translated Newton’s Principia Mathematica into French and was a companion of Voltaire. Unfortunately, after an affair with the poet Jean François de Saint-Lambert in her early 40’s, she became pregnant and died six days after giving birth.