Suppose that there is no difference in average health status for individuals with different levels of education when they are not insured, but that those with higher levels of education are much more likely to be insured. Consider the simple difference in outcomes (SDO) of health status for insured versus non-insured individuals. In this case, the fact higher education levels are positively related to having insurance, means that there will be selection bias from differences in education?
1) yes, selection bias from differences in education
2) no, no selection bias from education or other factors
3) no, no selection bias from education, but there may be bias from other variables