Please describe the trade-off of including a new independent variable into the regression model, especially for the unbiasedness and variance of the OLS estimator.
Tradeoff of including a new independent variable is that it may
lead to overfitting of the model.
(1) With overfitting the variances would be
larger, thereby standard errors would be large
(2) The OLS estimators would be inefficient due to
larger variaces.
(3) Confidence Interval and hypothesis testing results
may not be realiable (due to larger variances) as
calculated t-value and f-value would be lower and we will not
reject the null hypothesis more often than not.
(4) The OLS estimators for the model (with
additional independent variable) would be unbiased
as long as the new independent variable is not correlated with
error term. That is,
Other problems with overfitting:
(5) there will be loss of degrees of freedom
(6) It may create multicollinearity problem
Get Answers For Free
Most questions answered within 1 hours.