Clinical trials are intervention studies using human subjects in which a treatment (intervention) group is compared to a control group. The control group is treated the same as the treatment group except that subjects receive a placebo instead of an intervention.

Subjects in an intervention study should be randomly assigned to treatment/control groups, and it is best if neither the subject nor investigator knows which subjects are receiving treatment and control to avoid bias. A study in which neither investigator nor subject knows who is receiving treatment is called a double blind study. Sometimes it is impossible for the investigators to be kept in the dark about who is receiving treatment, so single blind studies, in which only the subject is kept in the dark, are employed.

A "crossover design" strengthens an intervention study, and negates the possibility that differences between treatment and control are due to differences between the individuals in each group. In a crossover design study, subjects receive both treatment and control conditions at different periods. A "washout" period between conditions is employed to make sure the first intervention does not influence the second intervention. It is best that one group receives treatment first and the other control first so that order of intervention is not a confounding factor.