An extraneous variable is an independent variable that may add noise to an experiment or bias the results, disrupting the effect that the chosen independent variable has on the dependent variable. In scientific experiments, the things that variables are highly controlled, but in the softer sciences, such as experimental psychology, it is more difficult to separate those characteristics that are required to be constant from those that must vary with each experiment. The goal in designing experiments is to reduce, as much as possible, the effects of extraneous variables.
There are three types of extraneous variables: subject, experimental, and situational. The subject variable is any variable that is characteristic of the person to be studied: age, health, mood, and physical features, for example. Experimental variables are those introduced by the bias of the psychologist or other observer conducting the psychological tests. Environmental variables come from environmental factors such as any distractions to the test subject from his or her surroundings: poor lighting, noise pollution, or other disruptions to testing. If an extraneous variable is potentially influential, it can be held constant for all subjects, or a spectrum of high to low intensity of a potentially harmful extraneous variable can be found through careful selection of test participants.
An example of a psychological experiment that might be compromised by an extraneous variable is sentence completion. A student whose intelligence quotient (IQ) is known is asked to complete a sentence fragment. The number of words he or she uses to complete the sentence is then recorded for each individual. The psychologist is looking to see if IQ, the independent variable, influences the number of words the student uses to complete the sentence, the dependent variable. An extraneous variable that biases an individual’s results might be whether the student ate breakfast that day, or how cold it was in the test facility, or how often the proctor frowned during the test.
When a researcher selects an independent variable in an experiment, behaviors of unexpected extraneous variables will sometimes be dependent on the intensity of the independent variable. An example of this situation is what is called a confounding variable. An example of such a variable might be the point of view of the test subject. For example, if subjects consist of both males and females, both nature and nurture will influence their perspectives on the test. Some might be comfortable with the experimental behaviors required, while others are not, or some may have pre-knowledge of the test and the expected results, while others do not.