Two Possibilities

Suppose, for example, that Professor Grumpinkston has a hypothesis that students seated closer to the front his classroom learn more and earn higher grades. He can test this hypothesis by comparing grades with seating position. The test indicates that closer seated students either earn higher grades, or they do not.

• Data and Hypothesis Agree: Suppose that real world data agree with a hypothesis. That is, the implications of a theory are supported by the facts. For Professor Grumpinkston's hypothesis, students seated near the front earn higher grades. What can be concluded?

  A successful test provides support for the hypothesis, but not absolute proof. For Professor Grumpinkston's particular class, during that particular semester, grades are higher for closer seated students. This, however, is only a single test. One test does not absolutely prove the hypothesis or identify the underlying scientific principle.

  To get reasonable proof, a hypothesis needs to be tested many, many, many times under many, many, many different conditions. But even then, a hypothesis is never 100 percent, absolutely proven. Science can NOT prove any hypothesis absolutely. Professor Grumpkinston needs to test his hypothesis over and over again, semester after semester.

  Each test that supports a hypothesis moves it closer to a verified principle. If the hypothesis is verified enough, it is ultimately incorporated into the theory for a new, expanded theory.



• Data and Hypothesis Do not Agree: Now suppose that real world data do not agree with a hypothesis? That is, the implications of a theory are not supported by the facts. Suppose students seated closer to the front of Professor Grumpinkston's class do not receive higher grades.

  While a hypothesis can never be absolutely proven to be absolutely true, it can be disproved, it can be rejected as false. Rejecting a hypothesis, however, can be extremely useful and informative. When disproven, a hypothesis is NOT true. It does not match the data. Professor Grumpinkston's hypothesis is wrong. This information is worth knowing. Something other than seating position is causing higher grades.

  If a hypothesis is refuted, then something is wrong. Perhaps the theory generating the hypothesis is wrong. Perhaps one or more of the previously verified principles or unverifiable axioms in the theory are wrong. Perhaps the specific test is bad. Perhaps the data are flawed.

  Whatever the reason for refutation, a "failed" test prompts further inquiry. Professor Grumpinkston needs further study of why students earn different grades.



• Further Testing Needed: The one inescapable conclusion from the hypothesis-verification process is the need for further testing. Whether a hypothesis is accepted or rejected, the recommendation is to test again. But this conclusion should not be surprising. Science is an ongoing process of discovery.

Controlling Other Factors