The variable that actually influences the [[Dependent Variable]] (instead of the assumed [[Independent Variable]]). I.e., it affects both the treatment (explanation) and the outcome (response). To avoid this influence, convert a Confounding Variable into a [[Control Variable]]. Confounders are sometimes referred to as **Lurking Variables**, especially when they are [latent](Latent%20Variable.md) confounders. Example: If you would try to predict grammar knowledge among kids based on the size of their feet, two of the *confounding* variables will probably be age and sex. And there are probably a few more *lurking* variables...