How to Use the Scientific Method

Observe. New knowledge is formed from curiosity. The observation process, sometimes called “question identification,” is quite simple. You observe something that you cannot explain with existing knowledge, or observe some phenomena that have been explained by existing knowledge but can still be explained in another way. At this point, the main question is how to explain what caused them to happen.

Research existing knowledge about your question. Let’s say you observe that the car won’t start. Your question is: why doesn’t the car explode? Maybe you have some knowledge about cars and will use that to try to find the cause. You can also refer to the user manual or find information online about this. If you are a scientist trying to understand some strange phenomena, you can consult scientific journals, which publish research that has been done by other scientists. You should read as much as you can about your question because it is likely that the answer is already there or you will find information that will help form your own hypothesis.

Hypothesis building. A hypothesis is a potential explanation for an observed phenomenon. However, it is more than just a judgment because it is based on a thorough examination of existing knowledge on the subject. It is essentially an educational judgment. The hypothesis should posit a cause-effect relationship. For example: “My car didn’t start because it ran out of gas”. It should provide a possible cause for the result and should be something you can test and predict. You can fill up the gas to test the “out of gas” hypothesis and you can predict if your hypothesis is correct, the car will start when the gas is added to the tank. The result stated as a fact will help it to be more like the real hypothesis. For those who are still unsure, use “if” and “then” statements: If I try to start the car and it doesn’t start , it runs out of gas.

List your materials. Make sure all the tools you need to do this project are listed. If anyone else wants to work on your idea, they’ll need to know EVERY material was used.

List your process. Record exactly each step taken to test the hypothesis. Again, this is an important step so that someone else can repeat your experiment.

Hypothesis test. Design an experiment through which either the hypothesis will or may not be confirmed. Experiments should be designed to attempt to separate the proposed phenomenon and cause. In other words, it should be “controlled”. Going back to the simple question about the car, we can test the hypothesis by putting gas in the tank but if we add more gas and replace the battery, we can’t tell for sure if the gas is out or the battery is the problem. For a more complex question, there can be hundreds of possible causes, and it may be difficult or impossible to separate them into individual experiments.

• Perfect record keeping. The experiment must be reproducible. That is, someone else has to do the same thing you did and get the same results. Therefore, it is very important to accurately record everything that is done in your test. At the same time, it is extremely important to store all the data. Today, several archival systems store raw data collected during scientific research. When needing to learn about your experiments, other scientists can consult these archives or contact you for data. Providing all the details of the experiment is crucial.

Analyze the results and draw conclusions. Hypothesis testing is simply a way of collecting data that will help you either confirm or not test a hypothesis. If the car explodes when adding gas, your analysis is quite simple: the hypothesis is confirmed. However, with more sophisticated tests, it may not be possible to determine whether a hypothesis is confirmed without spending considerable time looking at the data collected in the hypothesis test. In addition, even if the data corroborates or fails to confirm the hypothesis, you must always beware of the possibility that other things, collectively known as “exogenous” or “hidden” variables, may affect to the result. Suppose the car starts up when adding more gas, but at the same time, the weather changes and changes from rainy to sunny. Can you be sure that the gasoline and not the change in humidity caused the car to start? It is also possible that you have an inconclusive test. It is possible that the car runs for a few seconds after refueling and then shuts down again.

Report on research results. In general, scientists report research findings in scientific journals or present them in conferences. They report not only the results but also the methodology and any problems or questions that arise during the hypothesis testing process. Reporting research results makes it possible for others to use them.