If you regress using only observations 3 and 5, these are the data:
Observation # | bwght | motheduc |
3 | 3374 | 12 |
5 | 2835 | 12 |
Since thiere's no difference in the independent variable between the observations, the coefficient will be 0. So the equation will be something like:
bwght = constant + b.mothereduc, where b is likely to be zero
Observation motheduc bwght 2,778 3,600 3,374 1,219 2,835 faminc 37.5 15.5 42.5 22.5 22.5 3 bwght...