Experience $($EXP$)$ is positively associated with earnings $($EARN$),$ and years of schooling $(S)$ is positively associated with earnings. Also, EXP and $S$ are negatively correlated in younger workers. If a researcher estimates the regression :

$log(EARNINGS)={}_1+{}_2$ EXP $+u,$

on younger workers, the estimated coefficient on $b_2$ will be

biased downwards and underrepresent the effect of EXP on earnings.

biased upwards and overestimate the effect of EXP on earnings.

None of these answers are correct.

unbiased.

There is insufficient information to determine.