Endogeneity
Also called: endogenous variable, simultaneity bias, reverse causality
Endogeneity means a predictor is correlated with the error term, which breaks the causal interpretation of a regression coefficient. It arises three ways: an omitted confounder, reverse causality, or measurement error. Police numbers and crime rates illustrate the second — cities hire more police because crime is high, so a naive regression finds police cause crime. Identification strategies exist entirely to defeat it.
Want more than a definition? Learn it in Regression Range →