The first step to designing an experiment is deciding what the experiment will test or prove. Have a clearly stated hypothesis -- a basic idea to be proven or refuted. For example: Dogs that eat hamburger will gain more weight than dogs eating only dog food.
Locate a controlled setting for your experiment and determine what equipment will be required to test your hypothesis. To test the example hypothesis about dogs eating hamburger gaining more weight, you would need a controlled environment, like a kennel in addition to a scale to measure weight gain, dog food and hamburger.
Take the subjects or objects being tested and divide them into at least two groups: a control group, and an experimental group. The control group will be treated normally and compared to the group being treated differently. Take measurements to show how things were at the start of the experiment so that changes can be observed and recorded. These starting conditions are called a baseline. The two groups must be similar at the start of the experiment to exclude extraneous variables (things that may interfere with the experimental results).
In the example using dogs, the group of dogs that eats only dog food is the control group, while the experimental group eats hamburger. To cut down on extraneous variables, they should all be a similar type of dog at roughly the same age.
Variable environmental factors can affect experimental results, so carefully note laboratory conditions and changes. List all equipment used, steps taken and in which order. Experiments are worthless if they can't be repeated, so this data will be important to support conclusions drawn from the results of the experiment. The more data recorded, the more strength the conclusions will have. In the example hypothesis, data recorded would include the number and type of dogs used, weight gained by each dog over a set amount of time, how much food was administered, etc.
Compare the experimental results to the control. Use the collected data to draw conclusions supporting or refuting the original hypothesis. In the example experiment, if the data indicates that the dogs who ate hamburger gained more weight than the control group, the experiment supports the hypothesis. The conclusion can be drawn that dogs who eat hamburger do gain more weight than those which eat only dog food. If the conclusion is not what you expected, examine the data to see what went wrong and design a new experiment correcting any mistakes.