Irrelevant Variable

A regressor whose true coefficient is zero. Including it keeps OLS unbiased but raises the variance of the estimates.

When to use

Lecture 9 contrasts it with omitting a relevant variable: adding an irrelevant x3x_3 inflates variances (through Rj2R_j^2 when it is correlated with other regressors) but does not bias the other coefficients.

See Lec 9.

Appears in

No references yet.