Key concepts for the scientific method include specific vocabulary that students may have heard in other contexts, like data, control, variables, and hypothesis. Some of these words have a slightly different meaning in a scientific setting, particularly the word "theory". When you discuss what a theory is, be certain to differentiate between it and a hypothesis. In science, a hypothesis is an educated guess that will be tested in the experiment. A theory has already been tested repeatedly and has experimental results that verify it.
Choose an experiment for the class, or if there is a school or grade-level science fair, let students pick an individual project to participate in. Once a topic is chosen, the students should do research to define a hypothesis, and pick their control subject and variables. An example of this might be growing plants to test which fertilizer works better. Have the students brainstorm what might happen, and remind them that right now, they are making a hypothesis. Assemble the materials and talk about how the results will be recorded.
Once the experiment is set up, students must collect data. This is where the experimental setup is crucial. A good experiment has a control and variables. In the example experiment, suppose that a student is testing fertilizer brands. He must use the same plant, the same amount of sunlight, give it the same room to grow, and the same potting soil for each variable and the control. The control plant will receive no fertilizer, and the other plants will receive a certain fertilizer type, and their growth will be measured over a time period.
Students will have data to analyze and compare. Which plants grew the tallest? Why? Did one fertilizer work better? Have the students give descriptive results as well as numeric ones. Did one plant grow more flowers? Or more leaves? This teaches them to be observant of all the variables. Also, at this point, they should understand that often it takes more than one experiment to achieve reliable results. Ask them why they think this might be.