The myth of interpretability of econometric models

As expected, our variable is "significant", but also, probably more interesting, X_2, has no impact on the residuals