Controlling Variables

a. Independent vs. Dependent Variables

An experiment involves the manipulation of one or more variables by the experimenter in order to determine the effect of this manipulation on another variable. In the following research hypothesis, if albino rats are subjected to microwave radiation then their food consumption will decrease. In this hypothesis, which is reducible to if A, then B, the presence or absence of radiation is designated as the independent variable - the variable under the control of the experimenter. An independent variable is under the control of the experimenter to see how changing it causes a change in the dependent variable; the dependent variable depends on the independent variable. The terms independent variable and treatment are often used interchangeably. The dependent variable, the variable that is measured, is the amount of food consumed by the rats. The dependent variable reflects any effects associated with manipulation of the independent variable.

b. Nuisance or Extraneous Variables

In addition to independent variables, all experiments include one or more nuisance variables, which are undesired sources or variation in an experiment that may affect the dependent variable. Using the research hypothesis above, nuisance variables may include the sex of the rats, variation in the weight before the experiment, presence of infectious diseases in one or more of the cages where the rats are housed, temperature variation among cages and differences in previous feeding experiences of the rats. Unless these are controlled, these nuisance variables can bias the outcome of the experiment.

There are several approaches used to control nuisance variables. One approach is to hold the nuisance variable constant for all subjects. For example, use only male rats of the same weight. A second approach is to assign all subjects randomly to the experimental situations. In this case unsuspected sources of variation or bias are distributed over the entire experiment and do not affect just one or a limited number of treatments. A third approach is to include the variable as on of the factors in the experimental design. A fourth approach, to be discussed later, is to remove the effects of a nuisance variance statistically.