What Are the Common Elements of a Well-Designed Experiment?

Design

• A well-designed experiment is simple, easily replicated and maintains strict control over all variables. Conduct thorough background research to construct a hypothesis. A hypothesis is a well-informed guess about the relationship of cause and effect, such as "does vitamin C affect colds?" Design the experiment objectively, without preconceived expectations about the results. Choose unbiased, representative samples for experimental subjects and control groups. Precisely collect and record all data sets.

Variables

• Design an experiment to prove or disprove a causal relationship between the independent variable, vitamin C, and the dependent variable, the condition of the patient. The independent variable is an input factor controlled by the experimenter. The factor, or factors, have different levels and values. For example, the vitamin C may be administered in varying doses. Precise, consistent measurement of the independent variable data, such as the patient's health, is necessary for a successful experiment.

Control

• A common element of well-designed experiments is strict control of the main variables. Rule out the effects of extraneous variables, other than the independent and dependent ones. Good experiments employ blind control groups to compare results. Give a control group a placebo instead of vitamin C and assess the results. Don't tell the control members who receives the placebo. Conduct a double blind experiment by not telling the administrators who gets the placebo.

Other Elements

• Repeat the experiment several times with different subjects to demonstrate the statistical significance of the results and rule out randomness or chance. Keep the controlled variables constant throughout the experiment. Carefully analyze the date to correctly interpret the results and draw a conclusion from the experiment. Tabulate and control all variables to avoid confounding and misinterpreting their effects. Isolate and identify the key variable as the causative factor. Don't conduct an experiment using too small of a sample.