Bias Flashcards
What are three things that bias standard errors?
- Heteroskedasticity (error not constant) – standard errors too small
- Multicollinearity – inflates standard errors
- Inclusion of covariates not correlated with outcome
If errors are iid, but very large variance, are coefficients biased?
No
If errors are iid, but very large variance, are standard errors biased?
No
If errors are iid, but very large variance, are standard errors precise?
No
If Covariance between X1 and U is positive, is coefficent biased?
Yes, because unbiasedness of Beta1 depends on E(U | X1, X2) = 0. If Cov(X1, U) != 0, it implies E(U | X1, X2) != 0 and therefore Beta1 is biased.
If you Omit explanatory variable uncorrelated with other independent variables (but correlated with Y), is coefficient biased?
no
If you Omit explanatory variable uncorrelated with other independent variables (but correlated with Y), are standard errors biased?
no
If you omit explanatory variable uncorrelated with other independent variables (but correlated with Y), wha happens to precision of standard errors?
less precise
If X1 is correlated with error term and Y, is coefficient biased?
Yes, because independent variable is endogenous.
Error terms not normally distributed, are coefficients biased?
No
Error terms not normally distributed, are coefficients precise?
No
Error terms not normally distributed, can you conduct significant tests?
Significance tests incorrect, but as sample gets bigger this issue will become less severe (by CLT)
If Var(u | X1, X2) depends on X1 or X2, are standard errors precise?
SE will be too small; cannot resolve by increasing sample size; need Weighted Least Squares or robust standard errors
If Covariance (X1, X2) is positive, are coefficients biased?
No, as long as X1 and X2 are not perfectly collinear
If Y is continuous but not normally distributed, are coefficients biased?
no