A variable the researcher holds constant in an experiment is a controlled factor or constant. All variables remain constant in a controlled experiment except for the one tested and its effects. Neglecting to control all other variables can lead to experimental bias and invalidate the results of the experiment.
Besides controlled factors, two other factors are present in a controlled experiment. A factor that is allowed to change is called an independent variable, which consists of a treatment that is applied to the experimental subject. The effects of this treatment, called dependent variables, are observed and recorded by the researcher.
Because the validity of an experiment is directly affected by its construction and execution, attention to experimental design is extremely important. Good experimental design does not favor one outcome over another. Besides controlling all factors other than the independent and dependent variable, researchers guard against the placebo effect. The placebo effect occurs when participants anticipate the results of the experiment and react in a prescribed way even though no treatment is applied. The placebo effect can be countered by including a placebo group to which no treatment is applied for analysis.
To guard against bias in group assignment, well-designed experiments assign members to a group randomly. In a double blind experiment, neither the group member nor the researcher knows to which group the treatment is applied.