How to do an Experiment
Students often do experiments in science classes, but are not sure what to do if they need to do an experiment on their own for a science project.
How NOT to do an experiment
1. Know the difference between a demonstration and an experiment. The person doing a demonstration knows what will happen. When you do an experiment, you have a question and are looking for an answer. You don’t know what will happen.
Many people, including some science teacher, do demonstrations and call them experiments. When you heat glass and show that it can be bent, this isn’t an experiment.
“Experiments”in Chemistry classes are part demonstration and part experiment. The students don’t know what will happen, but the teacher does. The teacher is asking students to do this activity so they will learn both the process and the answer. That’s a good teaching method. But a real experiment would begin with a question the student is asking. The student would need to plan a way to find an answer to their question. They wouldn’t be given the procedure by the teacher.
Some students, perhaps someone you know, have done “science projects” by building a “volcano.” It isn’t a volcano, of course; it’s a model. It certainly isn’t an experiment.
2. If you already know what the result will be it’s not an experiment. I have actually had students who put one plant on the window sill and one in a dark closet. The plant in the closet died. The student already knew that plants needed light. Others have “proved” that plant die without water. What did they expect?
3. To have an experiment, you must test the hypothesis with a reasonable number of samples. One person who smoked and got lung cancer wasn’t enough to prove anything.
I have had students who supposedly bought two plants that looked about the same. After giving the plants fertilizer from two companies, or giving them different amounts of water, one plant grew taller than the other.
First: With two plants, or two children treated exactly the same, one usually grows taller than the other. It’s called natural variation. The student didn’t prove anything. They should have used at least 20 plants in each group. Using hundreds or thousands of plants would give even more reliable information.
Second: I tend to believe the student bought the two plants the day before the project was due and never put any fertilizer on either of them.
Third: Even if done carefully, the changes would depend on the plant. Some plants need more fertilizer, others do not. Some plants grow better with extra water. Others grow best in dry conditions.
How SHOULD you do an Experiment?
1. Find a good question.
Find a question you are interested in.
Find a question that has an answer that can be found.
Find a question that can be answered with material you have or can afford and in the time you have available.
Find a question that cannot be easily answered by checking the Internet or going to the library.
2. Do some research. Knowing more about the subject will help you make informed guesses (hypotheses) about the answer.
3. Choose the best hypothesis. It should be one you think is likely to be true. It should also be one that you can test.
4. Plan the experiment, the method for proving that your hypothesis is true or false.
a. List the materials you will need.
b. List the steps of the experiment in order.
c. Describe the data (information) you will collect and how it will be measured.
5. Carry out the experiment, noting observations and collecting data.
6. Analyze the data. The answer might be obvious. If you have a lot of data, try to organize it in tables or charts. You might try showing it on a graph.
7. After careful thinking, write a conclusion. Do not say you have proved something if all you really know is that it seems likely.
Avoid Common Mistakes
1. Is it really cause and effect? One common mistake people make is to believe that it event 1 happens and then event 2, that the first event caused the second one.
Sometimes we see information like “People who make high grades in math and science were more likely to have orange juice for breakfast.” Either the researcher or people hearing the results often jump to the conclusion that drinking orange juice will help students make higher grades.
We don’t know that. It might be a coincidence. There might be no relationship. There might be some other way to explain the relationship. It could be that students whose parents have a better education make higher grades and that better educated parents are more likely to serve orange juice for breakfast. Or perhaps intelligent students make higher grades (seems obvious) and that intelligent students know orange juice is good for them and they care about their health.
2. It is important to question your conclusions. Ask if there might be other ways to interpret your results. Is there any way to test your hypothesis further.
3. And a pet peeve of mine: Watch your spelling. I have had hundreds of students who misspell the word, “conclusion.” They tend to spell it “Conclution.” I pointed out the error and they did it again. I pronounced their word, “Con- clut- ee -on” and they looked puzzled and did it again. Even when I took five or ten points off their grades, they keep doing it.
Spelling is important in science too. If your teacher corrects your spelling on commonly used words, it would be a smart idea to learn to spell them correctly.
The next page contains two Examples of Good Experiments, 2 Experiments . Both are simple enough to do in Middle School and complex enough to be good projects for a High School Science Project.