Confounding Factor

A confounding factor in an analysis is a variable that is related to one or more of the variables measured or manipulated in a study, but not easily identified as such and/or not controlled in the design, so its pure (i.e., independent of other variables in the study) influence on the outcome cannot be determined. A confounding factor may obscure or hide an actual relation between variables or falsely demonstrate an apparent relation between the analysis variables where no real relation exists. If confounding factors are not identified, highly erroneous conclusions may be drawn from the results.

For example, there is a relation between the number of firefighters putting out a fire and the total losses sustained in that fire, which might lead to an erroneous conclusion that firefighters cause the losses if the confounding variable of “the size of the fire” (closely correlated with the number of firefighters involved) were not controlled for in the data analysis.