2.5: Test Your Hypotheses

Models form the cornerstone of human thinking, but a model, any model—conceptual, physical, or mathematical—is only as good as the facts that support it. To determine whether a model accurately describes or predicts a natural event, scientists carry out what may be the most important part of their work: creating testable hypotheses.

A hypothesis is a carefully worded, conditional, and testable explanation of how nature works. You’ll often hear people call them educated guesses, but there’s little guesswork involved. Scientists typically frame hypotheses within the context of known explanations. Because of previous observations and experiments, they often have a pretty good idea of what’s happening—or at least they think they do—and so hypotheses represent the next step in a progression of observations and experiments designed to figure something out. They aren’t random and they aren’t long lists that include every possible explanation. They are focused, narrowly interpreted, tentative statements that can be tested using carefully planned observations and carefully designed experiments. For example, you might state the hypothesis, “Whales jump out of the water to remove barnacles from their skin.” You could test this by observing whales.

One often overlooked and misunderstood characteristic of hypotheses is how they are worded. Scientists pose hypotheses in a manner that allows them to be rejected by observations and experiments. It may seem a curious way to go about something, but science generally attempts to disprove possibilities—eliminate hypotheses—rather than prove them. Scientists call this method falsification, the process of disproving a hypothesis. While some argue that falsification can be limiting in certain fields, especially branches of modern particle and theoretical physics (e.g., Carroll 2014), most scientists view it as a cornerstone of the scientific method. Our test of the whale-jumping hypothesis would be rejected (i.e., falsified) if we observed whales without barnacles jumping or if the act of jumping had no effect on whale-attached barnacles. See how that works?

The process of eliminating competing hypotheses was developed by Austrian-British science philosopher Karl Popper (1902–1994; Popper 2005). It belongs to a broader approach called deductive reasoning. Deduction operates as a kind of top-down logic: If A is true, and B belongs to A, then B is also true. For example, if all mammals breathe air, and a dolphin is a mammal, then dolphins breathe air.

While most scientists use deductive reasoning, the opposite form, inductive reasoning, also plays a part in science. Inductive reasoning operates as a bottom-up approach to figuring something out. Specific observations lead to general principles. For example, if dolphins use tools, and sea lions use tools, and both dolphins and sea lions are mammals, then, inductively thinking, all mammals use tools. Induction helps in the process of generating appropriate scientific questions and hypotheses from observations or preliminary experiments. Many scientists consider this the most interesting part of their research because it draws upon the creative mind. Science requires a person to imagine all the possibilities. For that reason, scientists tend to be highly creative people, despite their characterization as boring nerds.

A third form of reasoning, abductive reasoning, plays a larger role outside of science, where explanations or conclusions are needed to fix something, make a diagnosis, or render a judgment. Abductive reasoning is best explained as the quickest way to the most logical and usually simplest explanation, even though facts and observations may be incomplete. Car mechanics, doctors, and jurors often use abductive reasoning to complete a task because they are presented with limited data. Scientists often use abductive reasoning to generate hypotheses. It’s considered the more creative approach to scientific inquiry.

Hypothesis making through inductive and abductive reasoning and hypothesis falsification through deductive reasoning make up what’s known as the hypothetico-deductive approach, also known as the hypothetico-deductive model or simply the H-D method. It relies on the process of generating a hypothesis that can be falsified by a set of observations or experiments and then carrying out those observations or experiments to reject or support the hypothesis.

The H-D method remains a mainstay of science because it aspires to objectivity. Scientists go to great lengths to be impartial in their work from beginning to end. They don’t always get it right, but the self-correcting nature of science and the willingness of scientists to admit their errors make the H-D method the best thing going. To quote American philosopher and education reformer John Dewey (1859–1952):

Science represents the safeguard of the race against . . . natural propensities and the evils that flow from them. It consists of the special appliances and methods which the race has slowly worked out in order to conduct reflection under conditions whereby its procedures and results are tested. . . . Without initiation into the scientific spirit one is not in possession of the best tools which humanity has so far devised. (Dewey 1916)