Confounding Variables

When an associate between two measured variables (e.g., a predictor and an outcome) is affected by one or more unmeasured attributes, then the relation is said to be confounded. Any constructs, conditions, group memberships, or conditional effects that exert a meaningful influence on the relation between variables may be considered as confounding variables. It is important to note that by its very definition, confounding variables are often unmeasured or even unknown, which makes them a direct threat to the validity of any inferences made from the observed relations.

For example, an individual may seek psychological treatment for his or her irrational fear of spiders (arachnophobia). In therapy, the psychologist could draw the inference that the trigger for the fear response is the presence, image, or ideation ...