How Do You Test a Hypothesis?

A hypothesis is an unproven theory, an assumption about a phenomenon or a calculated guess about the way something works. Many hypothesis's are formed without any goals, but a true hypothesis is merely the beginning of a scientific method that often appears in the world of rationalism.

In order to have a hypothesis, you only need to form a question, then to formulate an answer based on what results you feel would occur if your hypothesis was tested. The hard part, is what comes after you've formed your hypothesis....

....testing your hypothesis...

In order to test your hypothesis, you will need to choose a method of testing. The most common are controlled experiments or an observational study. You can use both testing methods, though you'll want to complete one test before starting another, in order to receive more accurate results. Let's go over some examples.

Hypothetical Question: Does sugar cause hyperactivity in children?

Sample Hypothesis: Sugar causes hyperactivity in children.

In a controlled experiment, you will want to form control groups.Two control groups are often used the most, though for this sample hypothesis, we should use three control groups.

In control group Alpha, Beta and Delta - we will have ten children between the ages of 4 and 10, for a total of 30 children in all groups together.

In control group Alpha - we will feed the children a sugar filled diet.

In control group Beta - we will feed the children a sugar free diet.

In control group Delta - we will feed the children a sugar free diet that resembles a sugar filled diet. (this would be known as a placebo test)

None of the subjects of this test will be made aware of what the experiment is, or whether or not their diet contains sugar or not. Each participant will be randomly selected.

You will need to set a time limit for your controlled experiment. You can set it for 3 to 4 hours, or even 3 to 4 days.

Allow each group to consume the chosen diets, then when the experiment is over, write down the results of your test. How many children in group Alpha, were hyperactive? How many were not hyperactive? How many children in group Beta were hyperactive? How many were calm? How many children in group Delta were calm or hyperactive? These will be your results.

At the end of the test, you should have a good idea of whether or not sugar causes hyperactivity, based on your results.

If only the kids in group Alpha were hyperactive and the rest were mostly calm, then you've proved your hypothesis to be correct. If most of the kids in all of the groups were hyperactive, then you have proven your hypothesis incorrect, because most of the children were hyperactive regardless of their sugar intake. If you find that only the kids in group Alpha and group Delta, were hyperactive, then you'll need to do more testing. As this could form a new hypothetical question - is hyperactivity caused by the belief that sugar is in the diet?...

In an observational study, the experimenter cannot be in complete control of the experiment. This is often either because there is no way to randomly test the hypothesis, or because a true controlled test would be subject to major ethical complications.

In our sample hypothesis, we want to prove or disprove the hypothesis that hyperactivity is caused by consuming sugar. Although sugary diets are regularly consumed by many homes all over the world, the parents of the children might not find it ethically acceptable to allow an experimenter to feed them a diet of unknown sugar levels, or an only sugar diet. If this were to happen, there would be no way to get enough participants to test the hypothesis in a controlled experiment.

In an observational study, instead of randomly selecting 30 children to consume a certain diet, we would choose children who are already consuming sugar filled diets and children who are consuming sugar free diets. The participants would have to be aware that they were participating in the test and would have to allow you to observe the children for a set period of time.

Once you have found the participants of your study, you will test your hypothesis by taking observation to various factors in the participants lifestyles.

What is their regular diet?

How was their behavior one hour after eating?

How was their behavior three hours after eating?

What activities did the child take part in during the day?

Were the children subject to discipline at all during the day?

Was the child hyperactive? (when compared to normally "active" or "calm")

Observational studies are often considered 'uncontrolled experiments', do to the number of loop holes that could affect the study. The major one in our hypothesis, is that the investigator could be biased in some way. Another issue that might crop up, are the parents who decide to change their child's diet during the study (maybe they give them a little less sugar that day), which would lower the accuracy of your results.

Due to problems similar to these, most experimenters will observe more participants, in the hope that the more they study, the more accurate the results will be. In most cases, these hypothesis's cannot truly be proven or unproven, which leaves them subject to many problems.

The sad part is, most medical and legal hypothesis's can only be tested through observational studies, which means they can never truly be correct or incorrect.

In order to get the most accurate results, you would be wise to test your hypothesis in every way possible. If results vary in all situations, then your hypothesis remains a hypothesis until someone can prove or disprove it. If you find your hypothesis to be correct through your experiments, then you will have proven your hypothesis. If you find your results give opposite results than what you expected, then your hypothesis has been proven incorrect.

Once your hypothesis has been sufficiently tested, it now metaphors into a theory.