A lurking variable is an extraneous variable that is not included in statistical analysis. They either hide in an existing relaitonship between variables that are taken for the analysis or they create a phantom correlation where none is supposed to exist. This eventually leads to biased or misleading results in the study.
Effects arise from how they provide another interpretation for the relationship that exists between the independent and dependent variables while remaining hidden. The most common example would be the causal variable when we find a spurious correlation present.
- Falsely show a strong relationship between two variables
- Hide the relationship between two existing variables
- Causes a bias in the results of the study
One way you can identify them is through regression analysis. You can start by plotting the residuals, and then observe a trend whether linear or nonlinear, this will provide evidence to prove whether a particular variable is affecting the response variable.
Secondly, by Knowing the factors that could affect the relationship between the variables in the study (but which haven’t been included in the study) you can uncover potential lurking variables.
Difference between a Confounding and Lurking Variable is that a confounding variable is included in the study and establishes the foundation for the relationship between two other variables. A Lurking Variable is not included in this study and thus no foundations are established for any relationships it causes or removes.