If education and IQ are related, state whether any least squares assumptions will
be violated when estimating the simple regression of ln(wage) on educ. In this
case, will the estimates contain bias? Why or why not? And if so, in what
direction? Please explain your conclusions using the definition of your estimator
and the definition of unbiasedness.
So if education and IQ are actually related then IQ has an effect on education and since education is related to ln(Wage) then so is IQ. This means the estimates will contain bias. This will make the estimators larger since they both effect ln(wage) positively. Does this sound right? Doesn鈥檛 sound very statistical.|||More precisely, you have a problem of omittied variables. As IQ should be included in the regression and is not, the estimators would be biased and inconsistent, and the variance will be biased, so inference would be in general incorrect|||yes there will be bias, away from zero, as education has an indirect effect of signalling the ability.
your argument needs to say that IQ affects wage directly (given same education, smarter person is a better worker).
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment