Formalizing Ideas into Hypotheses

A brief discussion of what scientific ideas are and how they are organized is in order at this point, as we will be beginning discussion of how to turn research questions into actual hypotheses. When it comes to scientific ideas, there is a hierarchy of laws, theories, and hypotheses. These categories are not the sort of thing that come from a central authority, and there can be a great deal of overlap between them, but they generally come from common use. That is, if a scientific principle is so general and universal that it can be said to hold in all (or nearly all) situations, we generally call it a law. Scientific laws are few and far between, but good examples are the Newtonian laws of motion. The next step down in the hierarchy is a theory, which is a principle that is supported by a good deal of data, but it is not yet so universal that it can be said to hold in all situations. Finally, the hypothesis is a proposed principle or set of principles that may have some support but does not yet find enough support to be called a theory. Again, these categories are somewhat nebulous, but they can be a useful way to talk about scientific principles. For example, if you were to find a type of motion that violated a Newtonian law of motion, that would indicate either that you had a major scientific finding or flawed data.

When we think about scientific principles, whether they be law, theory, or hypothesis, we must always consider two fundamental concepts: generality and parsimony. Generality refers to the range of phenomena that can be described by the principle, while parsimony refers to the simplicity of the principle. You may have heard the saying "The simplest explanation is the most likely one." This is a rewording of a medieval friar's statement: “The principle that entities should not be multiplied needlessly; the simplest of two competing theories is to be preferred.”

This saying, known commonly as Occam's Razor, after William of Occam, is saying that simple explanations are generally better than more complex ones. Often, the formulation of a scientific principle requires a tradeoff between this preference for simplicity, or parsimony, and generality. We often must make a choice between a principle that describes a great deal of behavior but is very complex, or a principle that is simpler but does not describe as much. To a great extent, the practice of statistics is built upon this very principle, but we'll get into that later in the course.

One of the most critical aspects of a scientific principle is that it must be testable, or even falsifiable. The term that describes a statement that is by its nature impossible to disprove is referred to as a tautology, or a tautological statement. If one were to state that the reason why objects in motion tend to stay in motion (Newton's law of inertia) is because an almighty deity makes it happen, this is a problematic statement for scientific inquiry. This is not to say that this is untrue, but in order to investigate whether that statement is true, one would need to be able to measure that almighty deity. This is, of course, impossible. When you formulate a hypothesis or question, you must be careful to only develop ideas that are falsifiable.