Extraneous and Confounding Variables and Systematic vs Non-Systematic Error
Extraneous Variables are undesirable variables that influence the relationship between the variables that an experimenter is examining. Another way to think of this, is that these are variables the influence the outcome of an experiment, though they are not the variables that are actually of interest. These variables are undesirable because they add error to an experiment. A major goal in research design is to decrease or control the influence of extraneous variables as much as possible.
For example, letís say that an educational psychologist has developed a new learning strategy and is interested in examining the effectiveness of this strategy. The experimenter randomly assigns students to two groups. All of the students study text materials on a biology topic for thirty minutes. One group uses the new strategy and the other uses a strategy of their choice. Then all students complete a test over the materials. One obvious confounding variable in this case would be pre-knowledge of the biology topic that was studied. This variable will most likely influence student scores, regardless of which strategy they use. Because of this extraneous variable (and surely others) there will be some spread within each of the groups. It would be better, of course, if all students came in with the exact same pre-knowledge. However, the experimenter has taken an important step to greatly increase the chances that, at least, the extraneous variable will add error variance equivalently between the two groups. That is, the experimenter randomly assigned students to the two groups.
Random assignment is a powerful tool though it does nothing to decrease the amount of error that occurs as a result of extraneous variables, in only equalizes it between groups. In fact, even if the experimenter gave a pre-knowledge test ahead of time and then assigned students to groups, so that the groups were as equal as possible on pre-knowledge scores, this still would not change the fact that students would differ one from the other in terms of pre-knowledge and this would add "error variance" in the experiment.
The thing that makes random assignment so powerful is that greatly decreases systematic error Ė error that varies with the independent variable. Extraneous variables that vary with the levels of the independent variable are the most dangerous type in terms of challenging the validity of experimental results. These types of extraneous variables have a special name, confounding variables. For example, instead of randomly assigning students, the instructor may test the new strategy in the gifted classroom and test the control strategy in a regular class. Clearly, ability would most likely vary with the levels of the independent variable. In this case pre-knowledge would become a confounding extraneous variable. (Animated illustration of extraneous and confounding variables and systematic vs. non-systematic error variance.)
One of the most common types of confounding occurs when an experimenter does not or can not randomly assign participants to groups, and some type of individual difference (e.g., ability, extroversion, shyness, height, weight) acts as a confounding variable. For example, any experiment that involves a comparison of men and women is inherently plagued with confounding variables, the most commonly cited of which is that the social environment for males and females is very different. This does not mean that there is no meaning or value in gender comparison studies, or other studies in which random assignment is not employed, it simply means that we need to be more cautious in interpreting the results.