Answer to Question 1
Large samples are more sensitive to small differences between groups, which is an advantage; on the other hand, large samples cost more (time, money, and other resources), which is a disadvantage.
Small samples involve fewer resources, which is an advantage; on the other hand, small samples make it harder to detect real, but small, differences between groups.
Answer to Question 2
The SAT is imperfectly correlated with grades in college, a standard measure of success. Given that the correlation between SAT scores and first year grade point average is about .40, the amount of variability in grades accounted for by the SAT score is about 16 percent. This means that the vast majority of variability is related to unknown variables.